{"constraints": {"max_calls": 2, "max_credits": 6.177, "max_latency_ms": 178144, "minimum_quality": 0.8096}, "family": "migration_planning", "id": "synthetic-migration_planning-000163", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.757, "estimated_latency_ms": 80917, "estimated_pass_rate": 0.6544, "estimated_quality": 0.8208, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.177, "max_latency_ms": 178144, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6544, "quality": 0.8208, "quality_target_met": true, "utility": 60.851047}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.177,\"max_latency_ms\":178144,\"minimum_quality\":0.8096},\"task\":{\"ambiguity\":0.585,\"category\":\"coding\",\"complexity\":0.5629,\"consequence\":0.7075,\"evidence_need\":0.5118,\"family\":\"migration_planning\",\"input_tokens\":4880,\"reversibility\":0.2861,\"risk\":0.5804,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.137,\"id\":\"worker_a\",\"p95_latency_ms\":32827,\"pass_rate\":0.6073,\"provider_group\":\"provider_z\",\"quality\":0.8008,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.413,\"id\":\"worker_b\",\"p95_latency_ms\":83335,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9575,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.62,\"id\":\"worker_c\",\"p95_latency_ms\":48090,\"pass_rate\":0.7746,\"provider_group\":\"provider_y\",\"quality\":0.8588,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.757,\"estimated_latency_ms\":80917,\"estimated_pass_rate\":0.6544,\"estimated_quality\":0.8208,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.177,\"max_latency_ms\":178144,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.585, "category": "coding", "complexity": 0.5629, "consequence": 0.7075, "evidence_need": 0.5118, "family": "migration_planning", "input_tokens": 4880, "request_id": "synthetic-migration_planning-000163", "reversibility": 0.2861, "risk": 0.5804, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 1.137, "id": "worker_a", "p95_latency_ms": 32827, "pass_rate": 0.6073, "provider_group": "provider_z", "quality": 0.8008, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.413, "id": "worker_b", "p95_latency_ms": 83335, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9575, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.62, "id": "worker_c", "p95_latency_ms": 48090, "pass_rate": 0.7746, "provider_group": "provider_y", "quality": 0.8588, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 14.15, "max_latency_ms": 212309, "minimum_quality": 0.7206}, "family": "migration_planning", "id": "synthetic-migration_planning-000406", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.808, "estimated_latency_ms": 104798, "estimated_pass_rate": 0.5808, "estimated_quality": 0.7955, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.15, "max_latency_ms": 212309, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5808, "quality": 0.7955, "quality_target_met": true, "utility": 58.637657}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":14.15,\"max_latency_ms\":212309,\"minimum_quality\":0.7206},\"task\":{\"ambiguity\":0.5611,\"category\":\"coding\",\"complexity\":0.7612,\"consequence\":0.4228,\"evidence_need\":0.4167,\"family\":\"migration_planning\",\"input_tokens\":5760,\"reversibility\":0.3426,\"risk\":0.55,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":15.066,\"id\":\"worker_a\",\"p95_latency_ms\":89891,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9627,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.72,\"id\":\"worker_b\",\"p95_latency_ms\":72665,\"pass_rate\":0.8625,\"provider_group\":\"provider_y\",\"quality\":0.8468,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.088,\"id\":\"worker_c\",\"p95_latency_ms\":32133,\"pass_rate\":0.4759,\"provider_group\":\"provider_x\",\"quality\":0.7817,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.808,\"estimated_latency_ms\":104798,\"estimated_pass_rate\":0.5808,\"estimated_quality\":0.7955,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.15,\"max_latency_ms\":212309,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5611, "category": "coding", "complexity": 0.7612, "consequence": 0.4228, "evidence_need": 0.4167, "family": "migration_planning", "input_tokens": 5760, "request_id": "synthetic-migration_planning-000406", "reversibility": 0.3426, "risk": 0.55, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 15.066, "id": "worker_a", "p95_latency_ms": 89891, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9627, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.72, "id": "worker_b", "p95_latency_ms": 72665, "pass_rate": 0.8625, "provider_group": "provider_y", "quality": 0.8468, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.088, "id": "worker_c", "p95_latency_ms": 32133, "pass_rate": 0.4759, "provider_group": "provider_x", "quality": 0.7817, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.51, "max_latency_ms": 34638, "minimum_quality": 0.7588}, "family": "migration_planning", "id": "synthetic-migration_planning-000535", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.062, "estimated_latency_ms": 31877, "estimated_pass_rate": 0.4608, "estimated_quality": 0.6086, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.51, "max_latency_ms": 34638, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4608, "quality": 0.6086, "quality_target_met": false, "utility": 32.958717}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.51,\"max_latency_ms\":34638,\"minimum_quality\":0.7588},\"task\":{\"ambiguity\":0.4415,\"category\":\"coding\",\"complexity\":0.7067,\"consequence\":0.4357,\"evidence_need\":0.7365,\"family\":\"migration_planning\",\"input_tokens\":7280,\"reversibility\":0.543,\"risk\":0.3732,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.062,\"id\":\"worker_a\",\"p95_latency_ms\":31877,\"pass_rate\":0.6045,\"provider_group\":\"provider_y\",\"quality\":0.7129,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.987,\"id\":\"worker_b\",\"p95_latency_ms\":66094,\"pass_rate\":0.7565,\"provider_group\":\"provider_x\",\"quality\":0.8558,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.307,\"id\":\"worker_c\",\"p95_latency_ms\":86035,\"pass_rate\":0.9682,\"provider_group\":\"provider_z\",\"quality\":0.973,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.062,\"estimated_latency_ms\":31877,\"estimated_pass_rate\":0.4608,\"estimated_quality\":0.6086,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.51,\"max_latency_ms\":34638,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4415, "category": "coding", "complexity": 0.7067, "consequence": 0.4357, "evidence_need": 0.7365, "family": "migration_planning", "input_tokens": 7280, "request_id": "synthetic-migration_planning-000535", "reversibility": 0.543, "risk": 0.3732, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 1.062, "id": "worker_a", "p95_latency_ms": 31877, "pass_rate": 0.6045, "provider_group": "provider_y", "quality": 0.7129, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.987, "id": "worker_b", "p95_latency_ms": 66094, "pass_rate": 0.7565, "provider_group": "provider_x", "quality": 0.8558, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.307, "id": "worker_c", "p95_latency_ms": 86035, "pass_rate": 0.9682, "provider_group": "provider_z", "quality": 0.973, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.954, "max_latency_ms": 49469, "minimum_quality": 0.6835}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000027", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.935, "estimated_latency_ms": 27803, "estimated_pass_rate": 0.4141, "estimated_quality": 0.5575, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.954, "max_latency_ms": 49469, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4141, "quality": 0.5575, "quality_target_met": false, "utility": 20.917211}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.954,\"max_latency_ms\":49469,\"minimum_quality\":0.6835},\"task\":{\"ambiguity\":0.4849,\"category\":\"review\",\"complexity\":0.7681,\"consequence\":0.5621,\"evidence_need\":0.8211,\"family\":\"policy_compliance\",\"input_tokens\":3480,\"reversibility\":0.5683,\"risk\":0.8293,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.935,\"id\":\"worker_a\",\"p95_latency_ms\":27803,\"pass_rate\":0.605,\"provider_group\":\"provider_y\",\"quality\":0.7087,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.254,\"id\":\"worker_b\",\"p95_latency_ms\":77956,\"pass_rate\":0.989,\"provider_group\":\"provider_z\",\"quality\":0.9862,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.95,\"id\":\"worker_c\",\"p95_latency_ms\":73370,\"pass_rate\":0.8037,\"provider_group\":\"provider_x\",\"quality\":0.8238,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.935,\"estimated_latency_ms\":27803,\"estimated_pass_rate\":0.4141,\"estimated_quality\":0.5575,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.954,\"max_latency_ms\":49469,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4849, "category": "review", "complexity": 0.7681, "consequence": 0.5621, "evidence_need": 0.8211, "family": "policy_compliance", "input_tokens": 3480, "request_id": "synthetic-policy_compliance-000027", "reversibility": 0.5683, "risk": 0.8293, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 0.935, "id": "worker_a", "p95_latency_ms": 27803, "pass_rate": 0.605, "provider_group": "provider_y", "quality": 0.7087, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.254, "id": "worker_b", "p95_latency_ms": 77956, "pass_rate": 0.989, "provider_group": "provider_z", "quality": 0.9862, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.95, "id": "worker_c", "p95_latency_ms": 73370, "pass_rate": 0.8037, "provider_group": "provider_x", "quality": 0.8238, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.681, "max_latency_ms": 46912, "minimum_quality": 0.7356}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000132", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.984, "estimated_latency_ms": 29618, "estimated_pass_rate": 0.4081, "estimated_quality": 0.6147, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.681, "max_latency_ms": 46912, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4081, "quality": 0.6147, "quality_target_met": false, "utility": 37.009624}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.681,\"max_latency_ms\":46912,\"minimum_quality\":0.7356},\"task\":{\"ambiguity\":0.4239,\"category\":\"review\",\"complexity\":0.7342,\"consequence\":0.6355,\"evidence_need\":0.5595,\"family\":\"policy_compliance\",\"input_tokens\":7280,\"reversibility\":0.5946,\"risk\":0.7107,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":12.179,\"id\":\"worker_a\",\"p95_latency_ms\":88170,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.889,\"id\":\"worker_b\",\"p95_latency_ms\":64467,\"pass_rate\":0.7995,\"provider_group\":\"provider_y\",\"quality\":0.8579,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.984,\"id\":\"worker_c\",\"p95_latency_ms\":29618,\"pass_rate\":0.581,\"provider_group\":\"provider_z\",\"quality\":0.7471,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.984,\"estimated_latency_ms\":29618,\"estimated_pass_rate\":0.4081,\"estimated_quality\":0.6147,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.681,\"max_latency_ms\":46912,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4239, "category": "review", "complexity": 0.7342, "consequence": 0.6355, "evidence_need": 0.5595, "family": "policy_compliance", "input_tokens": 7280, "request_id": "synthetic-policy_compliance-000132", "reversibility": 0.5946, "risk": 0.7107, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 12.179, "id": "worker_a", "p95_latency_ms": 88170, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.889, "id": "worker_b", "p95_latency_ms": 64467, "pass_rate": 0.7995, "provider_group": "provider_y", "quality": 0.8579, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.984, "id": "worker_c", "p95_latency_ms": 29618, "pass_rate": 0.581, "provider_group": "provider_z", "quality": 0.7471, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.086, "max_latency_ms": 162733, "minimum_quality": 0.7826}, "family": "migration_planning", "id": "synthetic-migration_planning-000649", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.381, "estimated_latency_ms": 132848, "estimated_pass_rate": 0.727, "estimated_quality": 0.9016, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.086, "max_latency_ms": 162733, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.727, "quality": 0.9016, "quality_target_met": true, "utility": 72.474177}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.086,\"max_latency_ms\":162733,\"minimum_quality\":0.7826},\"task\":{\"ambiguity\":0.6247,\"category\":\"coding\",\"complexity\":0.7568,\"consequence\":0.4345,\"evidence_need\":0.6261,\"family\":\"migration_planning\",\"input_tokens\":3760,\"reversibility\":0.3802,\"risk\":0.6114,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":0.998,\"id\":\"worker_a\",\"p95_latency_ms\":25283,\"pass_rate\":0.5504,\"provider_group\":\"provider_x\",\"quality\":0.774,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.328,\"id\":\"worker_b\",\"p95_latency_ms\":61442,\"pass_rate\":0.9178,\"provider_group\":\"provider_z\",\"quality\":0.9632,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.055,\"id\":\"worker_c\",\"p95_latency_ms\":46123,\"pass_rate\":0.7823,\"provider_group\":\"provider_y\",\"quality\":0.865,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.381,\"estimated_latency_ms\":132848,\"estimated_pass_rate\":0.727,\"estimated_quality\":0.9016,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.086,\"max_latency_ms\":162733,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6247, "category": "coding", "complexity": 0.7568, "consequence": 0.4345, "evidence_need": 0.6261, "family": "migration_planning", "input_tokens": 3760, "request_id": "synthetic-migration_planning-000649", "reversibility": 0.3802, "risk": 0.6114, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 0.998, "id": "worker_a", "p95_latency_ms": 25283, "pass_rate": 0.5504, "provider_group": "provider_x", "quality": 0.774, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.328, "id": "worker_b", "p95_latency_ms": 61442, "pass_rate": 0.9178, "provider_group": "provider_z", "quality": 0.9632, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.055, "id": "worker_c", "p95_latency_ms": 46123, "pass_rate": 0.7823, "provider_group": "provider_y", "quality": 0.865, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.673, "max_latency_ms": 134776, "minimum_quality": 0.8741}, "family": "migration_planning", "id": "synthetic-migration_planning-000835", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.94, "estimated_latency_ms": 134503, "estimated_pass_rate": 0.8178, "estimated_quality": 0.9015, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.673, "max_latency_ms": 134776, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8178, "quality": 0.9015, "quality_target_met": true, "utility": 65.231902}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.673,\"max_latency_ms\":134776,\"minimum_quality\":0.8741},\"task\":{\"ambiguity\":0.3498,\"category\":\"coding\",\"complexity\":0.9336,\"consequence\":0.4084,\"evidence_need\":0.5413,\"family\":\"migration_planning\",\"input_tokens\":560,\"reversibility\":0.4759,\"risk\":0.4931,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":0.907,\"id\":\"worker_a\",\"p95_latency_ms\":30752,\"pass_rate\":0.5108,\"provider_group\":\"provider_y\",\"quality\":0.7884,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.226,\"id\":\"worker_b\",\"p95_latency_ms\":69307,\"pass_rate\":0.8416,\"provider_group\":\"provider_x\",\"quality\":0.8311,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.714,\"id\":\"worker_c\",\"p95_latency_ms\":65196,\"pass_rate\":0.8851,\"provider_group\":\"provider_z\",\"quality\":0.9783,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.94,\"estimated_latency_ms\":134503,\"estimated_pass_rate\":0.8178,\"estimated_quality\":0.9015,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.673,\"max_latency_ms\":134776,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3498, "category": "coding", "complexity": 0.9336, "consequence": 0.4084, "evidence_need": 0.5413, "family": "migration_planning", "input_tokens": 560, "request_id": "synthetic-migration_planning-000835", "reversibility": 0.4759, "risk": 0.4931, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 0.907, "id": "worker_a", "p95_latency_ms": 30752, "pass_rate": 0.5108, "provider_group": "provider_y", "quality": 0.7884, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.226, "id": "worker_b", "p95_latency_ms": 69307, "pass_rate": 0.8416, "provider_group": "provider_x", "quality": 0.8311, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.714, "id": "worker_c", "p95_latency_ms": 65196, "pass_rate": 0.8851, "provider_group": "provider_z", "quality": 0.9783, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.795, "max_latency_ms": 44938, "minimum_quality": 0.5977}, "family": "migration_planning", "id": "synthetic-migration_planning-000127", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.148, "estimated_latency_ms": 32952, "estimated_pass_rate": 0.4002, "estimated_quality": 0.7005, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.795, "max_latency_ms": 44938, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4002, "quality": 0.7005, "quality_target_met": true, "utility": 46.46072}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.795,\"max_latency_ms\":44938,\"minimum_quality\":0.5977},\"task\":{\"ambiguity\":0.4495,\"category\":\"coding\",\"complexity\":0.5778,\"consequence\":0.5675,\"evidence_need\":0.584,\"family\":\"migration_planning\",\"input_tokens\":7720,\"reversibility\":0.5825,\"risk\":0.5716,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.901,\"id\":\"worker_a\",\"p95_latency_ms\":82661,\"pass_rate\":0.8434,\"provider_group\":\"provider_x\",\"quality\":0.8984,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.678,\"id\":\"worker_b\",\"p95_latency_ms\":56003,\"pass_rate\":0.9331,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.148,\"id\":\"worker_c\",\"p95_latency_ms\":32952,\"pass_rate\":0.546,\"provider_group\":\"provider_y\",\"quality\":0.7963,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.148,\"estimated_latency_ms\":32952,\"estimated_pass_rate\":0.4002,\"estimated_quality\":0.7005,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.795,\"max_latency_ms\":44938,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4495, "category": "coding", "complexity": 0.5778, "consequence": 0.5675, "evidence_need": 0.584, "family": "migration_planning", "input_tokens": 7720, "request_id": "synthetic-migration_planning-000127", "reversibility": 0.5825, "risk": 0.5716, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 3.901, "id": "worker_a", "p95_latency_ms": 82661, "pass_rate": 0.8434, "provider_group": "provider_x", "quality": 0.8984, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.678, "id": "worker_b", "p95_latency_ms": 56003, "pass_rate": 0.9331, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.148, "id": "worker_c", "p95_latency_ms": 32952, "pass_rate": 0.546, "provider_group": "provider_y", "quality": 0.7963, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.954, "max_latency_ms": 32558, "minimum_quality": 0.691}, "family": "migration_planning", "id": "synthetic-migration_planning-000676", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.786, "estimated_latency_ms": 20342, "estimated_pass_rate": 0.3198, "estimated_quality": 0.5723, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.954, "max_latency_ms": 32558, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3198, "quality": 0.5723, "quality_target_met": false, "utility": 32.22527}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.954,\"max_latency_ms\":32558,\"minimum_quality\":0.691},\"task\":{\"ambiguity\":0.3472,\"category\":\"coding\",\"complexity\":0.902,\"consequence\":0.548,\"evidence_need\":0.5592,\"family\":\"migration_planning\",\"input_tokens\":5640,\"reversibility\":0.337,\"risk\":0.4756,\"summary\":\"Design compatibility testing for a artifact storage schema migration.\"},\"workers\":[{\"credits_per_call\":13.616,\"id\":\"worker_a\",\"p95_latency_ms\":59477,\"pass_rate\":0.9579,\"provider_group\":\"provider_z\",\"quality\":0.9468,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.107,\"id\":\"worker_b\",\"p95_latency_ms\":57543,\"pass_rate\":0.8164,\"provider_group\":\"provider_y\",\"quality\":0.9034,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.786,\"id\":\"worker_c\",\"p95_latency_ms\":20342,\"pass_rate\":0.4869,\"provider_group\":\"provider_x\",\"quality\":0.713,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.786,\"estimated_latency_ms\":20342,\"estimated_pass_rate\":0.3198,\"estimated_quality\":0.5723,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.954,\"max_latency_ms\":32558,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3472, "category": "coding", "complexity": 0.902, "consequence": 0.548, "evidence_need": 0.5592, "family": "migration_planning", "input_tokens": 5640, "request_id": "synthetic-migration_planning-000676", "reversibility": 0.337, "risk": 0.4756, "summary": "Design compatibility testing for a artifact storage schema migration."}, "workers": [{"credits_per_call": 13.616, "id": "worker_a", "p95_latency_ms": 59477, "pass_rate": 0.9579, "provider_group": "provider_z", "quality": 0.9468, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.107, "id": "worker_b", "p95_latency_ms": 57543, "pass_rate": 0.8164, "provider_group": "provider_y", "quality": 0.9034, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.786, "id": "worker_c", "p95_latency_ms": 20342, "pass_rate": 0.4869, "provider_group": "provider_x", "quality": 0.713, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.769, "max_latency_ms": 177667, "minimum_quality": 0.8243}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000206", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.836, "estimated_latency_ms": 132193, "estimated_pass_rate": 0.9264, "estimated_quality": 0.9597, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.769, "max_latency_ms": 177667, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.9264, "quality": 0.9597, "quality_target_met": true, "utility": 71.560528}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.769,\"max_latency_ms\":177667,\"minimum_quality\":0.8243},\"task\":{\"ambiguity\":0.5306,\"category\":\"analysis\",\"complexity\":0.4533,\"consequence\":0.2579,\"evidence_need\":0.4125,\"family\":\"performance_tuning\",\"input_tokens\":4080,\"reversibility\":0.7223,\"risk\":0.5304,\"summary\":\"Review a proposed optimization for schema evolution for hidden regressions.\"},\"workers\":[{\"credits_per_call\":2.919,\"id\":\"worker_a\",\"p95_latency_ms\":73214,\"pass_rate\":0.9527,\"provider_group\":\"provider_z\",\"quality\":0.9209,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.176,\"id\":\"worker_b\",\"p95_latency_ms\":22543,\"pass_rate\":0.4304,\"provider_group\":\"provider_y\",\"quality\":0.6995,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.917,\"id\":\"worker_c\",\"p95_latency_ms\":58979,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.836,\"estimated_latency_ms\":132193,\"estimated_pass_rate\":0.9264,\"estimated_quality\":0.9597,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.769,\"max_latency_ms\":177667,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5306, "category": "analysis", "complexity": 0.4533, "consequence": 0.2579, "evidence_need": 0.4125, "family": "performance_tuning", "input_tokens": 4080, "request_id": "synthetic-performance_tuning-000206", "reversibility": 0.7223, "risk": 0.5304, "summary": "Review a proposed optimization for schema evolution for hidden regressions."}, "workers": [{"credits_per_call": 2.919, "id": "worker_a", "p95_latency_ms": 73214, "pass_rate": 0.9527, "provider_group": "provider_z", "quality": 0.9209, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.176, "id": "worker_b", "p95_latency_ms": 22543, "pass_rate": 0.4304, "provider_group": "provider_y", "quality": 0.6995, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.917, "id": "worker_c", "p95_latency_ms": 58979, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.685, "max_latency_ms": 49941, "minimum_quality": 0.4946}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000543", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.797, "estimated_latency_ms": 26635, "estimated_pass_rate": 0.38, "estimated_quality": 0.6641, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.685, "max_latency_ms": 49941, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.38, "quality": 0.6641, "quality_target_met": true, "utility": 33.708967}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.685,\"max_latency_ms\":49941,\"minimum_quality\":0.4946},\"task\":{\"ambiguity\":0.3762,\"category\":\"review\",\"complexity\":0.6753,\"consequence\":0.6869,\"evidence_need\":0.7842,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.4893,\"risk\":0.7078,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.797,\"id\":\"worker_a\",\"p95_latency_ms\":26635,\"pass_rate\":0.5424,\"provider_group\":\"provider_x\",\"quality\":0.7824,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.907,\"id\":\"worker_b\",\"p95_latency_ms\":63187,\"pass_rate\":0.8547,\"provider_group\":\"provider_z\",\"quality\":0.7994,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.125,\"id\":\"worker_c\",\"p95_latency_ms\":65025,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.797,\"estimated_latency_ms\":26635,\"estimated_pass_rate\":0.38,\"estimated_quality\":0.6641,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.685,\"max_latency_ms\":49941,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3762, "category": "review", "complexity": 0.6753, "consequence": 0.6869, "evidence_need": 0.7842, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-000543", "reversibility": 0.4893, "risk": 0.7078, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 0.797, "id": "worker_a", "p95_latency_ms": 26635, "pass_rate": 0.5424, "provider_group": "provider_x", "quality": 0.7824, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.907, "id": "worker_b", "p95_latency_ms": 63187, "pass_rate": 0.8547, "provider_group": "provider_z", "quality": 0.7994, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.125, "id": "worker_c", "p95_latency_ms": 65025, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.078, "max_latency_ms": 143902, "minimum_quality": 0.7874}, "family": "migration_planning", "id": "synthetic-migration_planning-000502", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.465, "estimated_latency_ms": 86008, "estimated_pass_rate": 0.7416, "estimated_quality": 0.7904, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.078, "max_latency_ms": 143902, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.7416, "quality": 0.7904, "quality_target_met": true, "utility": 61.961316}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.078,\"max_latency_ms\":143902,\"minimum_quality\":0.7874},\"task\":{\"ambiguity\":0.3387,\"category\":\"coding\",\"complexity\":0.7715,\"consequence\":0.8229,\"evidence_need\":0.6595,\"family\":\"migration_planning\",\"input_tokens\":1320,\"reversibility\":0.6005,\"risk\":0.625,\"summary\":\"Design compatibility testing for a webhook delivery schema migration.\"},\"workers\":[{\"credits_per_call\":3.252,\"id\":\"worker_a\",\"p95_latency_ms\":64433,\"pass_rate\":0.8409,\"provider_group\":\"provider_z\",\"quality\":0.8592,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.91,\"id\":\"worker_b\",\"p95_latency_ms\":61089,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.934,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.213,\"id\":\"worker_c\",\"p95_latency_ms\":21575,\"pass_rate\":0.5576,\"provider_group\":\"provider_x\",\"quality\":0.6755,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.465,\"estimated_latency_ms\":86008,\"estimated_pass_rate\":0.7416,\"estimated_quality\":0.7904,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.078,\"max_latency_ms\":143902,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3387, "category": "coding", "complexity": 0.7715, "consequence": 0.8229, "evidence_need": 0.6595, "family": "migration_planning", "input_tokens": 1320, "request_id": "synthetic-migration_planning-000502", "reversibility": 0.6005, "risk": 0.625, "summary": "Design compatibility testing for a webhook delivery schema migration."}, "workers": [{"credits_per_call": 3.252, "id": "worker_a", "p95_latency_ms": 64433, "pass_rate": 0.8409, "provider_group": "provider_z", "quality": 0.8592, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.91, "id": "worker_b", "p95_latency_ms": 61089, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.934, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.213, "id": "worker_c", "p95_latency_ms": 21575, "pass_rate": 0.5576, "provider_group": "provider_x", "quality": 0.6755, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.798, "max_latency_ms": 39540, "minimum_quality": 0.6632}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000542", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.935, "estimated_latency_ms": 25337, "estimated_pass_rate": 0.2327, "estimated_quality": 0.6138, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.798, "max_latency_ms": 39540, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2327, "quality": 0.6138, "quality_target_met": false, "utility": 36.110156}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.798,\"max_latency_ms\":39540,\"minimum_quality\":0.6632},\"task\":{\"ambiguity\":0.5448,\"category\":\"analysis\",\"complexity\":0.727,\"consequence\":0.664,\"evidence_need\":0.4252,\"family\":\"performance_tuning\",\"input_tokens\":720,\"reversibility\":0.6097,\"risk\":0.4113,\"summary\":\"Choose between throughput and tail-latency changes for prompt caching.\"},\"workers\":[{\"credits_per_call\":10.603,\"id\":\"worker_a\",\"p95_latency_ms\":52141,\"pass_rate\":0.9602,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.102,\"id\":\"worker_b\",\"p95_latency_ms\":56222,\"pass_rate\":0.8724,\"provider_group\":\"provider_x\",\"quality\":0.8697,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.935,\"id\":\"worker_c\",\"p95_latency_ms\":25337,\"pass_rate\":0.3896,\"provider_group\":\"provider_z\",\"quality\":0.7315,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.935,\"estimated_latency_ms\":25337,\"estimated_pass_rate\":0.2327,\"estimated_quality\":0.6138,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.798,\"max_latency_ms\":39540,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5448, "category": "analysis", "complexity": 0.727, "consequence": 0.664, "evidence_need": 0.4252, "family": "performance_tuning", "input_tokens": 720, "request_id": "synthetic-performance_tuning-000542", "reversibility": 0.6097, "risk": 0.4113, "summary": "Choose between throughput and tail-latency changes for prompt caching."}, "workers": [{"credits_per_call": 10.603, "id": "worker_a", "p95_latency_ms": 52141, "pass_rate": 0.9602, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.102, "id": "worker_b", "p95_latency_ms": 56222, "pass_rate": 0.8724, "provider_group": "provider_x", "quality": 0.8697, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.935, "id": "worker_c", "p95_latency_ms": 25337, "pass_rate": 0.3896, "provider_group": "provider_z", "quality": 0.7315, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.586, "max_latency_ms": 31680, "minimum_quality": 0.4899}, "family": "migration_planning", "id": "synthetic-migration_planning-000175", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.821, "estimated_latency_ms": 30002, "estimated_pass_rate": 0.3368, "estimated_quality": 0.5951, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.586, "max_latency_ms": 31680, "policy": "lean", "reason_codes": ["independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3368, "quality": 0.5951, "quality_target_met": true, "utility": 15.902645}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.586,\"max_latency_ms\":31680,\"minimum_quality\":0.4899},\"task\":{\"ambiguity\":0.3818,\"category\":\"coding\",\"complexity\":0.5697,\"consequence\":0.7181,\"evidence_need\":0.8286,\"family\":\"migration_planning\",\"input_tokens\":5240,\"reversibility\":0.2935,\"risk\":0.7197,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.775,\"id\":\"worker_a\",\"p95_latency_ms\":79979,\"pass_rate\":0.8852,\"provider_group\":\"provider_z\",\"quality\":0.9167,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.821,\"id\":\"worker_b\",\"p95_latency_ms\":30002,\"pass_rate\":0.4886,\"provider_group\":\"provider_x\",\"quality\":0.6956,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.903,\"id\":\"worker_c\",\"p95_latency_ms\":55618,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9813,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.821,\"estimated_latency_ms\":30002,\"estimated_pass_rate\":0.3368,\"estimated_quality\":0.5951,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.586,\"max_latency_ms\":31680,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3818, "category": "coding", "complexity": 0.5697, "consequence": 0.7181, "evidence_need": 0.8286, "family": "migration_planning", "input_tokens": 5240, "request_id": "synthetic-migration_planning-000175", "reversibility": 0.2935, "risk": 0.7197, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 2.775, "id": "worker_a", "p95_latency_ms": 79979, "pass_rate": 0.8852, "provider_group": "provider_z", "quality": 0.9167, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.821, "id": "worker_b", "p95_latency_ms": 30002, "pass_rate": 0.4886, "provider_group": "provider_x", "quality": 0.6956, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.903, "id": "worker_c", "p95_latency_ms": 55618, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9813, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.447, "max_latency_ms": 59018, "minimum_quality": 0.6611}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000774", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.185, "estimated_latency_ms": 32841, "estimated_pass_rate": 0.3233, "estimated_quality": 0.5647, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.447, "max_latency_ms": 59018, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3233, "quality": 0.5647, "quality_target_met": false, "utility": 30.161892}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.447,\"max_latency_ms\":59018,\"minimum_quality\":0.6611},\"task\":{\"ambiguity\":0.4932,\"category\":\"review\",\"complexity\":0.6484,\"consequence\":0.5659,\"evidence_need\":0.7864,\"family\":\"policy_compliance\",\"input_tokens\":2840,\"reversibility\":0.525,\"risk\":0.6641,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":3.644,\"id\":\"worker_a\",\"p95_latency_ms\":64197,\"pass_rate\":0.7401,\"provider_group\":\"provider_x\",\"quality\":0.9066,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.66,\"id\":\"worker_b\",\"p95_latency_ms\":83171,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9793,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.185,\"id\":\"worker_c\",\"p95_latency_ms\":32841,\"pass_rate\":0.4879,\"provider_group\":\"provider_y\",\"quality\":0.6829,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.185,\"estimated_latency_ms\":32841,\"estimated_pass_rate\":0.3233,\"estimated_quality\":0.5647,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.447,\"max_latency_ms\":59018,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4932, "category": "review", "complexity": 0.6484, "consequence": 0.5659, "evidence_need": 0.7864, "family": "policy_compliance", "input_tokens": 2840, "request_id": "synthetic-policy_compliance-000774", "reversibility": 0.525, "risk": 0.6641, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 3.644, "id": "worker_a", "p95_latency_ms": 64197, "pass_rate": 0.7401, "provider_group": "provider_x", "quality": 0.9066, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.66, "id": "worker_b", "p95_latency_ms": 83171, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9793, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.185, "id": "worker_c", "p95_latency_ms": 32841, "pass_rate": 0.4879, "provider_group": "provider_y", "quality": 0.6829, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.08, "max_latency_ms": 195846, "minimum_quality": 0.8088}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001250", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.33, "estimated_latency_ms": 88678, "estimated_pass_rate": 0.792, "estimated_quality": 0.8137, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.08, "max_latency_ms": 195846, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.792, "quality": 0.8137, "quality_target_met": true, "utility": 61.287987}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.08,\"max_latency_ms\":195846,\"minimum_quality\":0.8088},\"task\":{\"ambiguity\":0.6671,\"category\":\"analysis\",\"complexity\":0.4545,\"consequence\":0.4547,\"evidence_need\":0.569,\"family\":\"performance_tuning\",\"input_tokens\":1400,\"reversibility\":0.8981,\"risk\":0.4911,\"summary\":\"Diagnose latency variance in prompt caching and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":12.182,\"id\":\"worker_a\",\"p95_latency_ms\":50197,\"pass_rate\":0.9535,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.078,\"id\":\"worker_b\",\"p95_latency_ms\":24148,\"pass_rate\":0.5286,\"provider_group\":\"provider_z\",\"quality\":0.7021,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.252,\"id\":\"worker_c\",\"p95_latency_ms\":64530,\"pass_rate\":0.8919,\"provider_group\":\"provider_x\",\"quality\":0.8583,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.33,\"estimated_latency_ms\":88678,\"estimated_pass_rate\":0.792,\"estimated_quality\":0.8137,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.08,\"max_latency_ms\":195846,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6671, "category": "analysis", "complexity": 0.4545, "consequence": 0.4547, "evidence_need": 0.569, "family": "performance_tuning", "input_tokens": 1400, "request_id": "synthetic-performance_tuning-001250", "reversibility": 0.8981, "risk": 0.4911, "summary": "Diagnose latency variance in prompt caching and rank the next experiments."}, "workers": [{"credits_per_call": 12.182, "id": "worker_a", "p95_latency_ms": 50197, "pass_rate": 0.9535, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.078, "id": "worker_b", "p95_latency_ms": 24148, "pass_rate": 0.5286, "provider_group": "provider_z", "quality": 0.7021, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.252, "id": "worker_c", "p95_latency_ms": 64530, "pass_rate": 0.8919, "provider_group": "provider_x", "quality": 0.8583, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.363, "max_latency_ms": 184719, "minimum_quality": 0.6996}, "family": "migration_planning", "id": "synthetic-migration_planning-001204", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.345, "estimated_latency_ms": 100566, "estimated_pass_rate": 0.5617, "estimated_quality": 0.7989, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.363, "max_latency_ms": 184719, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5617, "quality": 0.7989, "quality_target_met": true, "utility": 59.586595}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.363,\"max_latency_ms\":184719,\"minimum_quality\":0.6996},\"task\":{\"ambiguity\":0.6882,\"category\":\"coding\",\"complexity\":0.9227,\"consequence\":0.7536,\"evidence_need\":0.6789,\"family\":\"migration_planning\",\"input_tokens\":2040,\"reversibility\":0.4525,\"risk\":0.607,\"summary\":\"Review a staged migration for schema evolution and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.075,\"id\":\"worker_a\",\"p95_latency_ms\":24859,\"pass_rate\":0.4988,\"provider_group\":\"provider_y\",\"quality\":0.8031,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.27,\"id\":\"worker_b\",\"p95_latency_ms\":75707,\"pass_rate\":0.7338,\"provider_group\":\"provider_x\",\"quality\":0.9161,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.371,\"id\":\"worker_c\",\"p95_latency_ms\":58588,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9665,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.345,\"estimated_latency_ms\":100566,\"estimated_pass_rate\":0.5617,\"estimated_quality\":0.7989,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.363,\"max_latency_ms\":184719,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6882, "category": "coding", "complexity": 0.9227, "consequence": 0.7536, "evidence_need": 0.6789, "family": "migration_planning", "input_tokens": 2040, "request_id": "synthetic-migration_planning-001204", "reversibility": 0.4525, "risk": 0.607, "summary": "Review a staged migration for schema evolution and identify irreversible steps."}, "workers": [{"credits_per_call": 1.075, "id": "worker_a", "p95_latency_ms": 24859, "pass_rate": 0.4988, "provider_group": "provider_y", "quality": 0.8031, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.27, "id": "worker_b", "p95_latency_ms": 75707, "pass_rate": 0.7338, "provider_group": "provider_x", "quality": 0.9161, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.371, "id": "worker_c", "p95_latency_ms": 58588, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9665, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.124, "max_latency_ms": 250362, "minimum_quality": 0.8711}, "family": "migration_planning", "id": "synthetic-migration_planning-000724", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.454, "estimated_latency_ms": 195192, "estimated_pass_rate": 0.7375, "estimated_quality": 0.901, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.124, "max_latency_ms": 250362, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7375, "quality": 0.901, "quality_target_met": true, "utility": 72.749664}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.124,\"max_latency_ms\":250362,\"minimum_quality\":0.8711},\"task\":{\"ambiguity\":0.361,\"category\":\"coding\",\"complexity\":0.8216,\"consequence\":0.7664,\"evidence_need\":0.622,\"family\":\"migration_planning\",\"input_tokens\":2520,\"reversibility\":0.5041,\"risk\":0.5568,\"summary\":\"Review a staged migration for private credentials and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.509,\"id\":\"worker_a\",\"p95_latency_ms\":86751,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.005,\"id\":\"worker_b\",\"p95_latency_ms\":32292,\"pass_rate\":0.541,\"provider_group\":\"provider_z\",\"quality\":0.7012,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.94,\"id\":\"worker_c\",\"p95_latency_ms\":76149,\"pass_rate\":0.753,\"provider_group\":\"provider_y\",\"quality\":0.7949,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.454,\"estimated_latency_ms\":195192,\"estimated_pass_rate\":0.7375,\"estimated_quality\":0.901,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.124,\"max_latency_ms\":250362,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.361, "category": "coding", "complexity": 0.8216, "consequence": 0.7664, "evidence_need": 0.622, "family": "migration_planning", "input_tokens": 2520, "request_id": "synthetic-migration_planning-000724", "reversibility": 0.5041, "risk": 0.5568, "summary": "Review a staged migration for private credentials and identify irreversible steps."}, "workers": [{"credits_per_call": 11.509, "id": "worker_a", "p95_latency_ms": 86751, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.005, "id": "worker_b", "p95_latency_ms": 32292, "pass_rate": 0.541, "provider_group": "provider_z", "quality": 0.7012, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.94, "id": "worker_c", "p95_latency_ms": 76149, "pass_rate": 0.753, "provider_group": "provider_y", "quality": 0.7949, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.341, "max_latency_ms": 198359, "minimum_quality": 0.9523}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000978", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.92, "estimated_latency_ms": 162217, "estimated_pass_rate": 0.7542, "estimated_quality": 0.88, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.341, "max_latency_ms": 198359, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7542, "quality": 0.88, "quality_target_met": false, "utility": 72.277492}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.341,\"max_latency_ms\":198359,\"minimum_quality\":0.9523},\"task\":{\"ambiguity\":0.5209,\"category\":\"review\",\"complexity\":0.7931,\"consequence\":0.8571,\"evidence_need\":0.8028,\"family\":\"policy_compliance\",\"input_tokens\":5240,\"reversibility\":0.3151,\"risk\":0.6812,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.576,\"id\":\"worker_a\",\"p95_latency_ms\":88809,\"pass_rate\":0.9938,\"provider_group\":\"provider_z\",\"quality\":0.9173,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.479,\"id\":\"worker_b\",\"p95_latency_ms\":54822,\"pass_rate\":0.8222,\"provider_group\":\"provider_x\",\"quality\":0.8163,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.865,\"id\":\"worker_c\",\"p95_latency_ms\":18586,\"pass_rate\":0.5343,\"provider_group\":\"provider_y\",\"quality\":0.7247,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.92,\"estimated_latency_ms\":162217,\"estimated_pass_rate\":0.7542,\"estimated_quality\":0.88,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.341,\"max_latency_ms\":198359,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5209, "category": "review", "complexity": 0.7931, "consequence": 0.8571, "evidence_need": 0.8028, "family": "policy_compliance", "input_tokens": 5240, "request_id": "synthetic-policy_compliance-000978", "reversibility": 0.3151, "risk": 0.6812, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 13.576, "id": "worker_a", "p95_latency_ms": 88809, "pass_rate": 0.9938, "provider_group": "provider_z", "quality": 0.9173, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.479, "id": "worker_b", "p95_latency_ms": 54822, "pass_rate": 0.8222, "provider_group": "provider_x", "quality": 0.8163, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.865, "id": "worker_c", "p95_latency_ms": 18586, "pass_rate": 0.5343, "provider_group": "provider_y", "quality": 0.7247, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.465, "max_latency_ms": 49339, "minimum_quality": 0.5176}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000488", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.074, "estimated_latency_ms": 28410, "estimated_pass_rate": 0.3414, "estimated_quality": 0.5946, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.465, "max_latency_ms": 49339, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3414, "quality": 0.5946, "quality_target_met": true, "utility": 39.076346}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.465,\"max_latency_ms\":49339,\"minimum_quality\":0.5176},\"task\":{\"ambiguity\":0.7091,\"category\":\"analysis\",\"complexity\":0.7859,\"consequence\":0.5496,\"evidence_need\":0.5034,\"family\":\"performance_tuning\",\"input_tokens\":1680,\"reversibility\":0.6553,\"risk\":0.4764,\"summary\":\"Diagnose latency variance in access control and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.932,\"id\":\"worker_a\",\"p95_latency_ms\":54016,\"pass_rate\":0.875,\"provider_group\":\"provider_z\",\"quality\":0.8916,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.804,\"id\":\"worker_b\",\"p95_latency_ms\":50636,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.074,\"id\":\"worker_c\",\"p95_latency_ms\":28410,\"pass_rate\":0.5227,\"provider_group\":\"provider_y\",\"quality\":0.7386,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.074,\"estimated_latency_ms\":28410,\"estimated_pass_rate\":0.3414,\"estimated_quality\":0.5946,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.465,\"max_latency_ms\":49339,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7091, "category": "analysis", "complexity": 0.7859, "consequence": 0.5496, "evidence_need": 0.5034, "family": "performance_tuning", "input_tokens": 1680, "request_id": "synthetic-performance_tuning-000488", "reversibility": 0.6553, "risk": 0.4764, "summary": "Diagnose latency variance in access control and rank the next experiments."}, "workers": [{"credits_per_call": 3.932, "id": "worker_a", "p95_latency_ms": 54016, "pass_rate": 0.875, "provider_group": "provider_z", "quality": 0.8916, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.804, "id": "worker_b", "p95_latency_ms": 50636, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.074, "id": "worker_c", "p95_latency_ms": 28410, "pass_rate": 0.5227, "provider_group": "provider_y", "quality": 0.7386, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.733, "max_latency_ms": 201730, "minimum_quality": 0.8429}, "family": "migration_planning", "id": "synthetic-migration_planning-000844", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.41, "estimated_latency_ms": 166662, "estimated_pass_rate": 0.7381, "estimated_quality": 0.9152, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.733, "max_latency_ms": 201730, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7381, "quality": 0.9152, "quality_target_met": true, "utility": 73.091966}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.733,\"max_latency_ms\":201730,\"minimum_quality\":0.8429},\"task\":{\"ambiguity\":0.4349,\"category\":\"coding\",\"complexity\":0.8259,\"consequence\":0.8437,\"evidence_need\":0.4914,\"family\":\"migration_planning\",\"input_tokens\":3560,\"reversibility\":0.4255,\"risk\":0.4648,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.089,\"id\":\"worker_a\",\"p95_latency_ms\":73284,\"pass_rate\":0.7666,\"provider_group\":\"provider_y\",\"quality\":0.9027,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.132,\"id\":\"worker_b\",\"p95_latency_ms\":28706,\"pass_rate\":0.5111,\"provider_group\":\"provider_z\",\"quality\":0.7514,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.189,\"id\":\"worker_c\",\"p95_latency_ms\":64672,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.41,\"estimated_latency_ms\":166662,\"estimated_pass_rate\":0.7381,\"estimated_quality\":0.9152,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.733,\"max_latency_ms\":201730,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4349, "category": "coding", "complexity": 0.8259, "consequence": 0.8437, "evidence_need": 0.4914, "family": "migration_planning", "input_tokens": 3560, "request_id": "synthetic-migration_planning-000844", "reversibility": 0.4255, "risk": 0.4648, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.089, "id": "worker_a", "p95_latency_ms": 73284, "pass_rate": 0.7666, "provider_group": "provider_y", "quality": 0.9027, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.132, "id": "worker_b", "p95_latency_ms": 28706, "pass_rate": 0.5111, "provider_group": "provider_z", "quality": 0.7514, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.189, "id": "worker_c", "p95_latency_ms": 64672, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.166, "max_latency_ms": 155207, "minimum_quality": 0.8694}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000804", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.946, "estimated_latency_ms": 65000, "estimated_pass_rate": 0.6195, "estimated_quality": 0.8041, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.166, "max_latency_ms": 155207, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6195, "quality": 0.8041, "quality_target_met": false, "utility": 60.764303}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.166,\"max_latency_ms\":155207,\"minimum_quality\":0.8694},\"task\":{\"ambiguity\":0.7709,\"category\":\"review\",\"complexity\":0.47,\"consequence\":0.8263,\"evidence_need\":0.5916,\"family\":\"policy_compliance\",\"input_tokens\":2520,\"reversibility\":0.6265,\"risk\":0.6803,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":14.111,\"id\":\"worker_a\",\"p95_latency_ms\":65575,\"pass_rate\":0.8914,\"provider_group\":\"provider_y\",\"quality\":0.9428,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.05,\"id\":\"worker_b\",\"p95_latency_ms\":44488,\"pass_rate\":0.8397,\"provider_group\":\"provider_x\",\"quality\":0.8816,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.896,\"id\":\"worker_c\",\"p95_latency_ms\":20512,\"pass_rate\":0.4907,\"provider_group\":\"provider_z\",\"quality\":0.7168,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.946,\"estimated_latency_ms\":65000,\"estimated_pass_rate\":0.6195,\"estimated_quality\":0.8041,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.166,\"max_latency_ms\":155207,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7709, "category": "review", "complexity": 0.47, "consequence": 0.8263, "evidence_need": 0.5916, "family": "policy_compliance", "input_tokens": 2520, "request_id": "synthetic-policy_compliance-000804", "reversibility": 0.6265, "risk": 0.6803, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 14.111, "id": "worker_a", "p95_latency_ms": 65575, "pass_rate": 0.8914, "provider_group": "provider_y", "quality": 0.9428, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.05, "id": "worker_b", "p95_latency_ms": 44488, "pass_rate": 0.8397, "provider_group": "provider_x", "quality": 0.8816, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.896, "id": "worker_c", "p95_latency_ms": 20512, "pass_rate": 0.4907, "provider_group": "provider_z", "quality": 0.7168, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.248, "max_latency_ms": 176507, "minimum_quality": 0.8477}, "family": "migration_planning", "id": "synthetic-migration_planning-000415", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.989, "estimated_latency_ms": 142492, "estimated_pass_rate": 0.702, "estimated_quality": 0.915, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.248, "max_latency_ms": 176507, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.702, "quality": 0.915, "quality_target_met": true, "utility": 73.643923}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.248,\"max_latency_ms\":176507,\"minimum_quality\":0.8477},\"task\":{\"ambiguity\":0.6595,\"category\":\"coding\",\"complexity\":0.9318,\"consequence\":0.4057,\"evidence_need\":0.5557,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.5442,\"risk\":0.5463,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.493,\"id\":\"worker_a\",\"p95_latency_ms\":54918,\"pass_rate\":0.7253,\"provider_group\":\"provider_z\",\"quality\":0.9039,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.364,\"id\":\"worker_b\",\"p95_latency_ms\":61465,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.132,\"id\":\"worker_c\",\"p95_latency_ms\":26109,\"pass_rate\":0.4568,\"provider_group\":\"provider_x\",\"quality\":0.7801,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.989,\"estimated_latency_ms\":142492,\"estimated_pass_rate\":0.702,\"estimated_quality\":0.915,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.248,\"max_latency_ms\":176507,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6595, "category": "coding", "complexity": 0.9318, "consequence": 0.4057, "evidence_need": 0.5557, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-000415", "reversibility": 0.5442, "risk": 0.5463, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 3.493, "id": "worker_a", "p95_latency_ms": 54918, "pass_rate": 0.7253, "provider_group": "provider_z", "quality": 0.9039, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.364, "id": "worker_b", "p95_latency_ms": 61465, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.132, "id": "worker_c", "p95_latency_ms": 26109, "pass_rate": 0.4568, "provider_group": "provider_x", "quality": 0.7801, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.523, "max_latency_ms": 165144, "minimum_quality": 0.7417}, "family": "migration_planning", "id": "synthetic-migration_planning-001084", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.459, "estimated_latency_ms": 91084, "estimated_pass_rate": 0.8279, "estimated_quality": 0.8896, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.523, "max_latency_ms": 165144, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8279, "quality": 0.8896, "quality_target_met": true, "utility": 63.86707}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.523,\"max_latency_ms\":165144,\"minimum_quality\":0.7417},\"task\":{\"ambiguity\":0.318,\"category\":\"coding\",\"complexity\":0.8342,\"consequence\":0.526,\"evidence_need\":0.6608,\"family\":\"migration_planning\",\"input_tokens\":2160,\"reversibility\":0.4162,\"risk\":0.6477,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.282,\"id\":\"worker_a\",\"p95_latency_ms\":68809,\"pass_rate\":0.9484,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.177,\"id\":\"worker_b\",\"p95_latency_ms\":22275,\"pass_rate\":0.5369,\"provider_group\":\"provider_x\",\"quality\":0.7877,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.028,\"id\":\"worker_c\",\"p95_latency_ms\":81923,\"pass_rate\":0.8304,\"provider_group\":\"provider_z\",\"quality\":0.8453,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.459,\"estimated_latency_ms\":91084,\"estimated_pass_rate\":0.8279,\"estimated_quality\":0.8896,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.523,\"max_latency_ms\":165144,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.318, "category": "coding", "complexity": 0.8342, "consequence": 0.526, "evidence_need": 0.6608, "family": "migration_planning", "input_tokens": 2160, "request_id": "synthetic-migration_planning-001084", "reversibility": 0.4162, "risk": 0.6477, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 14.282, "id": "worker_a", "p95_latency_ms": 68809, "pass_rate": 0.9484, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.177, "id": "worker_b", "p95_latency_ms": 22275, "pass_rate": 0.5369, "provider_group": "provider_x", "quality": 0.7877, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.028, "id": "worker_c", "p95_latency_ms": 81923, "pass_rate": 0.8304, "provider_group": "provider_z", "quality": 0.8453, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.594, "max_latency_ms": 33375, "minimum_quality": 0.5747}, "family": "migration_planning", "id": "synthetic-migration_planning-000853", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.991, "estimated_latency_ms": 20653, "estimated_pass_rate": 0.2608, "estimated_quality": 0.5856, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.594, "max_latency_ms": 33375, "policy": "lean", "reason_codes": ["complex_task", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2608, "quality": 0.5856, "quality_target_met": true, "utility": 35.168981}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.594,\"max_latency_ms\":33375,\"minimum_quality\":0.5747},\"task\":{\"ambiguity\":0.5417,\"category\":\"coding\",\"complexity\":0.9076,\"consequence\":0.6506,\"evidence_need\":0.4021,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.5354,\"risk\":0.7331,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":12.472,\"id\":\"worker_a\",\"p95_latency_ms\":69027,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9649,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.798,\"id\":\"worker_b\",\"p95_latency_ms\":72291,\"pass_rate\":0.7934,\"provider_group\":\"provider_y\",\"quality\":0.8137,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.991,\"id\":\"worker_c\",\"p95_latency_ms\":20653,\"pass_rate\":0.4641,\"provider_group\":\"provider_x\",\"quality\":0.7589,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.991,\"estimated_latency_ms\":20653,\"estimated_pass_rate\":0.2608,\"estimated_quality\":0.5856,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.594,\"max_latency_ms\":33375,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5417, "category": "coding", "complexity": 0.9076, "consequence": 0.6506, "evidence_need": 0.4021, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-000853", "reversibility": 0.5354, "risk": 0.7331, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 12.472, "id": "worker_a", "p95_latency_ms": 69027, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9649, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.798, "id": "worker_b", "p95_latency_ms": 72291, "pass_rate": 0.7934, "provider_group": "provider_y", "quality": 0.8137, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.991, "id": "worker_c", "p95_latency_ms": 20653, "pass_rate": 0.4641, "provider_group": "provider_x", "quality": 0.7589, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.833, "max_latency_ms": 125714, "minimum_quality": 0.8328}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000948", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.569, "estimated_latency_ms": 84100, "estimated_pass_rate": 0.7241, "estimated_quality": 0.8463, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.833, "max_latency_ms": 125714, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7241, "quality": 0.8463, "quality_target_met": true, "utility": 62.838149}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.833,\"max_latency_ms\":125714,\"minimum_quality\":0.8328},\"task\":{\"ambiguity\":0.5956,\"category\":\"review\",\"complexity\":0.6701,\"consequence\":0.8105,\"evidence_need\":0.7548,\"family\":\"policy_compliance\",\"input_tokens\":4400,\"reversibility\":0.3096,\"risk\":0.5277,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.169,\"id\":\"worker_a\",\"p95_latency_ms\":19611,\"pass_rate\":0.5885,\"provider_group\":\"provider_y\",\"quality\":0.6918,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.4,\"id\":\"worker_b\",\"p95_latency_ms\":64489,\"pass_rate\":0.817,\"provider_group\":\"provider_x\",\"quality\":0.9216,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.803,\"id\":\"worker_c\",\"p95_latency_ms\":75151,\"pass_rate\":0.9567,\"provider_group\":\"provider_z\",\"quality\":0.9586,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.569,\"estimated_latency_ms\":84100,\"estimated_pass_rate\":0.7241,\"estimated_quality\":0.8463,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.833,\"max_latency_ms\":125714,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5956, "category": "review", "complexity": 0.6701, "consequence": 0.8105, "evidence_need": 0.7548, "family": "policy_compliance", "input_tokens": 4400, "request_id": "synthetic-policy_compliance-000948", "reversibility": 0.3096, "risk": 0.5277, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 1.169, "id": "worker_a", "p95_latency_ms": 19611, "pass_rate": 0.5885, "provider_group": "provider_y", "quality": 0.6918, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.4, "id": "worker_b", "p95_latency_ms": 64489, "pass_rate": 0.817, "provider_group": "provider_x", "quality": 0.9216, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.803, "id": "worker_c", "p95_latency_ms": 75151, "pass_rate": 0.9567, "provider_group": "provider_z", "quality": 0.9586, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.907, "max_latency_ms": 223000, "minimum_quality": 0.9175}, "family": "migration_planning", "id": "synthetic-migration_planning-000802", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.059, "estimated_latency_ms": 168901, "estimated_pass_rate": 0.7656, "estimated_quality": 0.927, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.907, "max_latency_ms": 223000, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7656, "quality": 0.927, "quality_target_met": true, "utility": 74.750823}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.907,\"max_latency_ms\":223000,\"minimum_quality\":0.9175},\"task\":{\"ambiguity\":0.4955,\"category\":\"coding\",\"complexity\":0.6783,\"consequence\":0.6486,\"evidence_need\":0.4922,\"family\":\"migration_planning\",\"input_tokens\":1160,\"reversibility\":0.4238,\"risk\":0.6936,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.791,\"id\":\"worker_a\",\"p95_latency_ms\":20801,\"pass_rate\":0.5111,\"provider_group\":\"provider_y\",\"quality\":0.7952,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.108,\"id\":\"worker_b\",\"p95_latency_ms\":76395,\"pass_rate\":0.8744,\"provider_group\":\"provider_z\",\"quality\":0.8407,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.16,\"id\":\"worker_c\",\"p95_latency_ms\":71705,\"pass_rate\":0.9885,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.059,\"estimated_latency_ms\":168901,\"estimated_pass_rate\":0.7656,\"estimated_quality\":0.927,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.907,\"max_latency_ms\":223000,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4955, "category": "coding", "complexity": 0.6783, "consequence": 0.6486, "evidence_need": 0.4922, "family": "migration_planning", "input_tokens": 1160, "request_id": "synthetic-migration_planning-000802", "reversibility": 0.4238, "risk": 0.6936, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 0.791, "id": "worker_a", "p95_latency_ms": 20801, "pass_rate": 0.5111, "provider_group": "provider_y", "quality": 0.7952, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.108, "id": "worker_b", "p95_latency_ms": 76395, "pass_rate": 0.8744, "provider_group": "provider_z", "quality": 0.8407, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.16, "id": "worker_c", "p95_latency_ms": 71705, "pass_rate": 0.9885, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.834, "max_latency_ms": 223659, "minimum_quality": 0.7698}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001071", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.312, "estimated_latency_ms": 156519, "estimated_pass_rate": 0.8639, "estimated_quality": 0.9234, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.834, "max_latency_ms": 223659, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8639, "quality": 0.9234, "quality_target_met": true, "utility": 69.931938}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.834,\"max_latency_ms\":223659,\"minimum_quality\":0.7698},\"task\":{\"ambiguity\":0.3819,\"category\":\"review\",\"complexity\":0.8496,\"consequence\":0.7033,\"evidence_need\":0.6928,\"family\":\"policy_compliance\",\"input_tokens\":4400,\"reversibility\":0.4128,\"risk\":0.5208,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":3.791,\"id\":\"worker_a\",\"p95_latency_ms\":71436,\"pass_rate\":0.8496,\"provider_group\":\"provider_z\",\"quality\":0.8761,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.521,\"id\":\"worker_b\",\"p95_latency_ms\":85083,\"pass_rate\":0.942,\"provider_group\":\"provider_y\",\"quality\":0.9847,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.947,\"id\":\"worker_c\",\"p95_latency_ms\":24342,\"pass_rate\":0.5596,\"provider_group\":\"provider_x\",\"quality\":0.6763,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.312,\"estimated_latency_ms\":156519,\"estimated_pass_rate\":0.8639,\"estimated_quality\":0.9234,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.834,\"max_latency_ms\":223659,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3819, "category": "review", "complexity": 0.8496, "consequence": 0.7033, "evidence_need": 0.6928, "family": "policy_compliance", "input_tokens": 4400, "request_id": "synthetic-policy_compliance-001071", "reversibility": 0.4128, "risk": 0.5208, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 3.791, "id": "worker_a", "p95_latency_ms": 71436, "pass_rate": 0.8496, "provider_group": "provider_z", "quality": 0.8761, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.521, "id": "worker_b", "p95_latency_ms": 85083, "pass_rate": 0.942, "provider_group": "provider_y", "quality": 0.9847, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.947, "id": "worker_c", "p95_latency_ms": 24342, "pass_rate": 0.5596, "provider_group": "provider_x", "quality": 0.6763, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.7, "max_latency_ms": 25580, "minimum_quality": 0.4826}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000579", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.215, "estimated_latency_ms": 23566, "estimated_pass_rate": 0.2873, "estimated_quality": 0.6311, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.7, "max_latency_ms": 25580, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2873, "quality": 0.6311, "quality_target_met": true, "utility": 36.587459}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.7,\"max_latency_ms\":25580,\"minimum_quality\":0.4826},\"task\":{\"ambiguity\":0.5984,\"category\":\"review\",\"complexity\":0.8453,\"consequence\":0.5291,\"evidence_need\":0.7125,\"family\":\"policy_compliance\",\"input_tokens\":1720,\"reversibility\":0.6227,\"risk\":0.696,\"summary\":\"Identify evidence required before approving background jobs for production.\"},\"workers\":[{\"credits_per_call\":3.971,\"id\":\"worker_a\",\"p95_latency_ms\":72785,\"pass_rate\":0.752,\"provider_group\":\"provider_z\",\"quality\":0.8441,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.215,\"id\":\"worker_b\",\"p95_latency_ms\":23566,\"pass_rate\":0.4848,\"provider_group\":\"provider_y\",\"quality\":0.7943,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.174,\"id\":\"worker_c\",\"p95_latency_ms\":59320,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.215,\"estimated_latency_ms\":23566,\"estimated_pass_rate\":0.2873,\"estimated_quality\":0.6311,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.7,\"max_latency_ms\":25580,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5984, "category": "review", "complexity": 0.8453, "consequence": 0.5291, "evidence_need": 0.7125, "family": "policy_compliance", "input_tokens": 1720, "request_id": "synthetic-policy_compliance-000579", "reversibility": 0.6227, "risk": 0.696, "summary": "Identify evidence required before approving background jobs for production."}, "workers": [{"credits_per_call": 3.971, "id": "worker_a", "p95_latency_ms": 72785, "pass_rate": 0.752, "provider_group": "provider_z", "quality": 0.8441, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.215, "id": "worker_b", "p95_latency_ms": 23566, "pass_rate": 0.4848, "provider_group": "provider_y", "quality": 0.7943, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.174, "id": "worker_c", "p95_latency_ms": 59320, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.384, "max_latency_ms": 91069, "minimum_quality": 0.7506}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000327", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.152, "estimated_latency_ms": 88921, "estimated_pass_rate": 0.5855, "estimated_quality": 0.8146, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.384, "max_latency_ms": 91069, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5855, "quality": 0.8146, "quality_target_met": true, "utility": 58.226477}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.384,\"max_latency_ms\":91069,\"minimum_quality\":0.7506},\"task\":{\"ambiguity\":0.7197,\"category\":\"review\",\"complexity\":0.6492,\"consequence\":0.5547,\"evidence_need\":0.7394,\"family\":\"policy_compliance\",\"input_tokens\":2480,\"reversibility\":0.3818,\"risk\":0.4598,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.849,\"id\":\"worker_a\",\"p95_latency_ms\":58957,\"pass_rate\":0.937,\"provider_group\":\"provider_y\",\"quality\":0.9695,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.933,\"id\":\"worker_b\",\"p95_latency_ms\":18446,\"pass_rate\":0.5181,\"provider_group\":\"provider_x\",\"quality\":0.7666,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.219,\"id\":\"worker_c\",\"p95_latency_ms\":70475,\"pass_rate\":0.8369,\"provider_group\":\"provider_z\",\"quality\":0.8345,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.152,\"estimated_latency_ms\":88921,\"estimated_pass_rate\":0.5855,\"estimated_quality\":0.8146,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.384,\"max_latency_ms\":91069,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7197, "category": "review", "complexity": 0.6492, "consequence": 0.5547, "evidence_need": 0.7394, "family": "policy_compliance", "input_tokens": 2480, "request_id": "synthetic-policy_compliance-000327", "reversibility": 0.3818, "risk": 0.4598, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 11.849, "id": "worker_a", "p95_latency_ms": 58957, "pass_rate": 0.937, "provider_group": "provider_y", "quality": 0.9695, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.933, "id": "worker_b", "p95_latency_ms": 18446, "pass_rate": 0.5181, "provider_group": "provider_x", "quality": 0.7666, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.219, "id": "worker_c", "p95_latency_ms": 70475, "pass_rate": 0.8369, "provider_group": "provider_z", "quality": 0.8345, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.267, "max_latency_ms": 53780, "minimum_quality": 0.7122}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000078", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.099, "estimated_latency_ms": 28952, "estimated_pass_rate": 0.3358, "estimated_quality": 0.556, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.267, "max_latency_ms": 53780, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3358, "quality": 0.556, "quality_target_met": false, "utility": 10.40431}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.267,\"max_latency_ms\":53780,\"minimum_quality\":0.7122},\"task\":{\"ambiguity\":0.7041,\"category\":\"review\",\"complexity\":0.5485,\"consequence\":0.9331,\"evidence_need\":0.9319,\"family\":\"policy_compliance\",\"input_tokens\":1320,\"reversibility\":0.5672,\"risk\":0.8076,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.099,\"id\":\"worker_a\",\"p95_latency_ms\":28952,\"pass_rate\":0.5166,\"provider_group\":\"provider_z\",\"quality\":0.6806,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.965,\"id\":\"worker_b\",\"p95_latency_ms\":85788,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9567,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.789,\"id\":\"worker_c\",\"p95_latency_ms\":54921,\"pass_rate\":0.8635,\"provider_group\":\"provider_y\",\"quality\":0.8745,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.099,\"estimated_latency_ms\":28952,\"estimated_pass_rate\":0.3358,\"estimated_quality\":0.556,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.267,\"max_latency_ms\":53780,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7041, "category": "review", "complexity": 0.5485, "consequence": 0.9331, "evidence_need": 0.9319, "family": "policy_compliance", "input_tokens": 1320, "request_id": "synthetic-policy_compliance-000078", "reversibility": 0.5672, "risk": 0.8076, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.099, "id": "worker_a", "p95_latency_ms": 28952, "pass_rate": 0.5166, "provider_group": "provider_z", "quality": 0.6806, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.965, "id": "worker_b", "p95_latency_ms": 85788, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9567, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.789, "id": "worker_c", "p95_latency_ms": 54921, "pass_rate": 0.8635, "provider_group": "provider_y", "quality": 0.8745, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.469, "max_latency_ms": 135679, "minimum_quality": 0.8108}, "family": "migration_planning", "id": "synthetic-migration_planning-000901", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.739, "estimated_latency_ms": 80893, "estimated_pass_rate": 0.7411, "estimated_quality": 0.8643, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.469, "max_latency_ms": 135679, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7411, "quality": 0.8643, "quality_target_met": true, "utility": 65.430619}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.469,\"max_latency_ms\":135679,\"minimum_quality\":0.8108},\"task\":{\"ambiguity\":0.7031,\"category\":\"coding\",\"complexity\":0.7029,\"consequence\":0.6495,\"evidence_need\":0.6796,\"family\":\"migration_planning\",\"input_tokens\":160,\"reversibility\":0.5914,\"risk\":0.628,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":3.491,\"id\":\"worker_a\",\"p95_latency_ms\":82718,\"pass_rate\":0.8766,\"provider_group\":\"provider_x\",\"quality\":0.869,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.048,\"id\":\"worker_b\",\"p95_latency_ms\":22210,\"pass_rate\":0.5846,\"provider_group\":\"provider_y\",\"quality\":0.699,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.691,\"id\":\"worker_c\",\"p95_latency_ms\":58683,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.739,\"estimated_latency_ms\":80893,\"estimated_pass_rate\":0.7411,\"estimated_quality\":0.8643,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.469,\"max_latency_ms\":135679,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7031, "category": "coding", "complexity": 0.7029, "consequence": 0.6495, "evidence_need": 0.6796, "family": "migration_planning", "input_tokens": 160, "request_id": "synthetic-migration_planning-000901", "reversibility": 0.5914, "risk": 0.628, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 3.491, "id": "worker_a", "p95_latency_ms": 82718, "pass_rate": 0.8766, "provider_group": "provider_x", "quality": 0.869, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.048, "id": "worker_b", "p95_latency_ms": 22210, "pass_rate": 0.5846, "provider_group": "provider_y", "quality": 0.699, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.691, "id": "worker_c", "p95_latency_ms": 58683, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.249, "max_latency_ms": 166885, "minimum_quality": 0.7976}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000942", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.792, "estimated_latency_ms": 121855, "estimated_pass_rate": 0.8349, "estimated_quality": 0.9184, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.249, "max_latency_ms": 166885, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8349, "quality": 0.9184, "quality_target_met": true, "utility": 68.888809}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.249,\"max_latency_ms\":166885,\"minimum_quality\":0.7976},\"task\":{\"ambiguity\":0.5647,\"category\":\"review\",\"complexity\":0.8394,\"consequence\":0.5237,\"evidence_need\":0.6375,\"family\":\"policy_compliance\",\"input_tokens\":4240,\"reversibility\":0.5449,\"risk\":0.7422,\"summary\":\"Audit a workflow for provider failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.75,\"id\":\"worker_a\",\"p95_latency_ms\":47062,\"pass_rate\":0.8855,\"provider_group\":\"provider_z\",\"quality\":0.8411,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.784,\"id\":\"worker_b\",\"p95_latency_ms\":27539,\"pass_rate\":0.5416,\"provider_group\":\"provider_y\",\"quality\":0.679,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.042,\"id\":\"worker_c\",\"p95_latency_ms\":74793,\"pass_rate\":0.8959,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.792,\"estimated_latency_ms\":121855,\"estimated_pass_rate\":0.8349,\"estimated_quality\":0.9184,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.249,\"max_latency_ms\":166885,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5647, "category": "review", "complexity": 0.8394, "consequence": 0.5237, "evidence_need": 0.6375, "family": "policy_compliance", "input_tokens": 4240, "request_id": "synthetic-policy_compliance-000942", "reversibility": 0.5449, "risk": 0.7422, "summary": "Audit a workflow for provider failover against explicit access requirements."}, "workers": [{"credits_per_call": 3.75, "id": "worker_a", "p95_latency_ms": 47062, "pass_rate": 0.8855, "provider_group": "provider_z", "quality": 0.8411, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.784, "id": "worker_b", "p95_latency_ms": 27539, "pass_rate": 0.5416, "provider_group": "provider_y", "quality": 0.679, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.042, "id": "worker_c", "p95_latency_ms": 74793, "pass_rate": 0.8959, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.681, "max_latency_ms": 166580, "minimum_quality": 0.9472}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000438", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.14, "estimated_latency_ms": 137035, "estimated_pass_rate": 0.7578, "estimated_quality": 0.9247, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.681, "max_latency_ms": 166580, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7578, "quality": 0.9247, "quality_target_met": false, "utility": 77.495414}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.681,\"max_latency_ms\":166580,\"minimum_quality\":0.9472},\"task\":{\"ambiguity\":0.5669,\"category\":\"review\",\"complexity\":0.8137,\"consequence\":0.6032,\"evidence_need\":0.609,\"family\":\"policy_compliance\",\"input_tokens\":5600,\"reversibility\":0.5718,\"risk\":0.8189,\"summary\":\"Audit a workflow for tenant-scoped caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.53,\"id\":\"worker_a\",\"p95_latency_ms\":64896,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.83,\"id\":\"worker_b\",\"p95_latency_ms\":25913,\"pass_rate\":0.4825,\"provider_group\":\"provider_y\",\"quality\":0.795,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.78,\"id\":\"worker_c\",\"p95_latency_ms\":46226,\"pass_rate\":0.8448,\"provider_group\":\"provider_z\",\"quality\":0.8555,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.14,\"estimated_latency_ms\":137035,\"estimated_pass_rate\":0.7578,\"estimated_quality\":0.9247,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.681,\"max_latency_ms\":166580,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5669, "category": "review", "complexity": 0.8137, "consequence": 0.6032, "evidence_need": 0.609, "family": "policy_compliance", "input_tokens": 5600, "request_id": "synthetic-policy_compliance-000438", "reversibility": 0.5718, "risk": 0.8189, "summary": "Audit a workflow for tenant-scoped caching against explicit access requirements."}, "workers": [{"credits_per_call": 10.53, "id": "worker_a", "p95_latency_ms": 64896, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.83, "id": "worker_b", "p95_latency_ms": 25913, "pass_rate": 0.4825, "provider_group": "provider_y", "quality": 0.795, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.78, "id": "worker_c", "p95_latency_ms": 46226, "pass_rate": 0.8448, "provider_group": "provider_z", "quality": 0.8555, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.133, "max_latency_ms": 62021, "minimum_quality": 0.6768}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000240", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.857, "estimated_latency_ms": 31156, "estimated_pass_rate": 0.3145, "estimated_quality": 0.692, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.133, "max_latency_ms": 62021, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3145, "quality": 0.692, "quality_target_met": true, "utility": 30.210356}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.133,\"max_latency_ms\":62021,\"minimum_quality\":0.6768},\"task\":{\"ambiguity\":0.5066,\"category\":\"review\",\"complexity\":0.4707,\"consequence\":0.7561,\"evidence_need\":0.9454,\"family\":\"policy_compliance\",\"input_tokens\":1640,\"reversibility\":0.5873,\"risk\":0.5991,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.857,\"id\":\"worker_a\",\"p95_latency_ms\":31156,\"pass_rate\":0.4547,\"provider_group\":\"provider_y\",\"quality\":0.7743,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.597,\"id\":\"worker_b\",\"p95_latency_ms\":45444,\"pass_rate\":0.7465,\"provider_group\":\"provider_z\",\"quality\":0.8034,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.279,\"id\":\"worker_c\",\"p95_latency_ms\":84653,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.857,\"estimated_latency_ms\":31156,\"estimated_pass_rate\":0.3145,\"estimated_quality\":0.692,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.133,\"max_latency_ms\":62021,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5066, "category": "review", "complexity": 0.4707, "consequence": 0.7561, "evidence_need": 0.9454, "family": "policy_compliance", "input_tokens": 1640, "request_id": "synthetic-policy_compliance-000240", "reversibility": 0.5873, "risk": 0.5991, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 0.857, "id": "worker_a", "p95_latency_ms": 31156, "pass_rate": 0.4547, "provider_group": "provider_y", "quality": 0.7743, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.597, "id": "worker_b", "p95_latency_ms": 45444, "pass_rate": 0.7465, "provider_group": "provider_z", "quality": 0.8034, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.279, "id": "worker_c", "p95_latency_ms": 84653, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.174, "max_latency_ms": 210984, "minimum_quality": 0.7846}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001233", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.352, "estimated_latency_ms": 183215, "estimated_pass_rate": 0.7308, "estimated_quality": 0.9155, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.174, "max_latency_ms": 210984, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7308, "quality": 0.9155, "quality_target_met": true, "utility": 72.977113}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.174,\"max_latency_ms\":210984,\"minimum_quality\":0.7846},\"task\":{\"ambiguity\":0.7818,\"category\":\"review\",\"complexity\":0.605,\"consequence\":0.6024,\"evidence_need\":0.5832,\"family\":\"policy_compliance\",\"input_tokens\":400,\"reversibility\":0.4672,\"risk\":0.6566,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":15.191,\"id\":\"worker_a\",\"p95_latency_ms\":77546,\"pass_rate\":0.9698,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.183,\"id\":\"worker_b\",\"p95_latency_ms\":75527,\"pass_rate\":0.7294,\"provider_group\":\"provider_z\",\"quality\":0.8967,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.978,\"id\":\"worker_c\",\"p95_latency_ms\":30142,\"pass_rate\":0.5099,\"provider_group\":\"provider_x\",\"quality\":0.7573,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.352,\"estimated_latency_ms\":183215,\"estimated_pass_rate\":0.7308,\"estimated_quality\":0.9155,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.174,\"max_latency_ms\":210984,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7818, "category": "review", "complexity": 0.605, "consequence": 0.6024, "evidence_need": 0.5832, "family": "policy_compliance", "input_tokens": 400, "request_id": "synthetic-policy_compliance-001233", "reversibility": 0.4672, "risk": 0.6566, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 15.191, "id": "worker_a", "p95_latency_ms": 77546, "pass_rate": 0.9698, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.183, "id": "worker_b", "p95_latency_ms": 75527, "pass_rate": 0.7294, "provider_group": "provider_z", "quality": 0.8967, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.978, "id": "worker_c", "p95_latency_ms": 30142, "pass_rate": 0.5099, "provider_group": "provider_x", "quality": 0.7573, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.414, "max_latency_ms": 220748, "minimum_quality": 0.8917}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000618", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.969, "estimated_latency_ms": 177895, "estimated_pass_rate": 0.7863, "estimated_quality": 0.9267, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.414, "max_latency_ms": 220748, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7863, "quality": 0.9267, "quality_target_met": true, "utility": 77.83008}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.414,\"max_latency_ms\":220748,\"minimum_quality\":0.8917},\"task\":{\"ambiguity\":0.525,\"category\":\"review\",\"complexity\":0.6257,\"consequence\":0.7794,\"evidence_need\":0.6718,\"family\":\"policy_compliance\",\"input_tokens\":2400,\"reversibility\":0.4447,\"risk\":0.7386,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.921,\"id\":\"worker_a\",\"p95_latency_ms\":74282,\"pass_rate\":0.9768,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.837,\"id\":\"worker_b\",\"p95_latency_ms\":71421,\"pass_rate\":0.799,\"provider_group\":\"provider_x\",\"quality\":0.8883,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.211,\"id\":\"worker_c\",\"p95_latency_ms\":32192,\"pass_rate\":0.5878,\"provider_group\":\"provider_y\",\"quality\":0.7713,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.969,\"estimated_latency_ms\":177895,\"estimated_pass_rate\":0.7863,\"estimated_quality\":0.9267,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.414,\"max_latency_ms\":220748,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.525, "category": "review", "complexity": 0.6257, "consequence": 0.7794, "evidence_need": 0.6718, "family": "policy_compliance", "input_tokens": 2400, "request_id": "synthetic-policy_compliance-000618", "reversibility": 0.4447, "risk": 0.7386, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 13.921, "id": "worker_a", "p95_latency_ms": 74282, "pass_rate": 0.9768, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.837, "id": "worker_b", "p95_latency_ms": 71421, "pass_rate": 0.799, "provider_group": "provider_x", "quality": 0.8883, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.211, "id": "worker_c", "p95_latency_ms": 32192, "pass_rate": 0.5878, "provider_group": "provider_y", "quality": 0.7713, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.356, "max_latency_ms": 217932, "minimum_quality": 0.8234}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000620", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.369, "estimated_latency_ms": 101315, "estimated_pass_rate": 0.7067, "estimated_quality": 0.8315, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.356, "max_latency_ms": 217932, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7067, "quality": 0.8315, "quality_target_met": true, "utility": 57.814942}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.356,\"max_latency_ms\":217932,\"minimum_quality\":0.8234},\"task\":{\"ambiguity\":0.6148,\"category\":\"analysis\",\"complexity\":0.7824,\"consequence\":0.2824,\"evidence_need\":0.6531,\"family\":\"performance_tuning\",\"input_tokens\":2320,\"reversibility\":0.6007,\"risk\":0.3618,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":14.653,\"id\":\"worker_a\",\"p95_latency_ms\":72443,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.547,\"id\":\"worker_b\",\"p95_latency_ms\":77798,\"pass_rate\":0.8204,\"provider_group\":\"provider_y\",\"quality\":0.9336,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.822,\"id\":\"worker_c\",\"p95_latency_ms\":23517,\"pass_rate\":0.4998,\"provider_group\":\"provider_x\",\"quality\":0.6909,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.369,\"estimated_latency_ms\":101315,\"estimated_pass_rate\":0.7067,\"estimated_quality\":0.8315,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.356,\"max_latency_ms\":217932,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6148, "category": "analysis", "complexity": 0.7824, "consequence": 0.2824, "evidence_need": 0.6531, "family": "performance_tuning", "input_tokens": 2320, "request_id": "synthetic-performance_tuning-000620", "reversibility": 0.6007, "risk": 0.3618, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 14.653, "id": "worker_a", "p95_latency_ms": 72443, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.547, "id": "worker_b", "p95_latency_ms": 77798, "pass_rate": 0.8204, "provider_group": "provider_y", "quality": 0.9336, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.822, "id": "worker_c", "p95_latency_ms": 23517, "pass_rate": 0.4998, "provider_group": "provider_x", "quality": 0.6909, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.486, "max_latency_ms": 139458, "minimum_quality": 0.8187}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000297", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.215, "estimated_latency_ms": 128864, "estimated_pass_rate": 0.7052, "estimated_quality": 0.8952, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.486, "max_latency_ms": 139458, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7052, "quality": 0.8952, "quality_target_met": true, "utility": 72.123672}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.486,\"max_latency_ms\":139458,\"minimum_quality\":0.8187},\"task\":{\"ambiguity\":0.5512,\"category\":\"review\",\"complexity\":0.5651,\"consequence\":0.732,\"evidence_need\":0.8908,\"family\":\"policy_compliance\",\"input_tokens\":3400,\"reversibility\":0.4389,\"risk\":0.576,\"summary\":\"Audit a workflow for model routing against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.275,\"id\":\"worker_a\",\"p95_latency_ms\":50916,\"pass_rate\":0.935,\"provider_group\":\"provider_y\",\"quality\":0.9286,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.005,\"id\":\"worker_b\",\"p95_latency_ms\":27126,\"pass_rate\":0.4739,\"provider_group\":\"provider_z\",\"quality\":0.7545,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.935,\"id\":\"worker_c\",\"p95_latency_ms\":50822,\"pass_rate\":0.7642,\"provider_group\":\"provider_x\",\"quality\":0.8454,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.215,\"estimated_latency_ms\":128864,\"estimated_pass_rate\":0.7052,\"estimated_quality\":0.8952,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.486,\"max_latency_ms\":139458,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5512, "category": "review", "complexity": 0.5651, "consequence": 0.732, "evidence_need": 0.8908, "family": "policy_compliance", "input_tokens": 3400, "request_id": "synthetic-policy_compliance-000297", "reversibility": 0.4389, "risk": 0.576, "summary": "Audit a workflow for model routing against explicit access requirements."}, "workers": [{"credits_per_call": 12.275, "id": "worker_a", "p95_latency_ms": 50916, "pass_rate": 0.935, "provider_group": "provider_y", "quality": 0.9286, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.005, "id": "worker_b", "p95_latency_ms": 27126, "pass_rate": 0.4739, "provider_group": "provider_z", "quality": 0.7545, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.935, "id": "worker_c", "p95_latency_ms": 50822, "pass_rate": 0.7642, "provider_group": "provider_x", "quality": 0.8454, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.106, "max_latency_ms": 49998, "minimum_quality": 0.7516}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000567", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.933, "estimated_latency_ms": 30812, "estimated_pass_rate": 0.4101, "estimated_quality": 0.5353, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.106, "max_latency_ms": 49998, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4101, "quality": 0.5353, "quality_target_met": false, "utility": 15.052475}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.106,\"max_latency_ms\":49998,\"minimum_quality\":0.7516},\"task\":{\"ambiguity\":0.5437,\"category\":\"review\",\"complexity\":0.7921,\"consequence\":0.7826,\"evidence_need\":0.8652,\"family\":\"policy_compliance\",\"input_tokens\":240,\"reversibility\":0.6324,\"risk\":0.6126,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":14.292,\"id\":\"worker_a\",\"p95_latency_ms\":81136,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.933,\"id\":\"worker_b\",\"p95_latency_ms\":30812,\"pass_rate\":0.5906,\"provider_group\":\"provider_y\",\"quality\":0.6792,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.182,\"id\":\"worker_c\",\"p95_latency_ms\":71207,\"pass_rate\":0.7619,\"provider_group\":\"provider_z\",\"quality\":0.8856,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.933,\"estimated_latency_ms\":30812,\"estimated_pass_rate\":0.4101,\"estimated_quality\":0.5353,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.106,\"max_latency_ms\":49998,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5437, "category": "review", "complexity": 0.7921, "consequence": 0.7826, "evidence_need": 0.8652, "family": "policy_compliance", "input_tokens": 240, "request_id": "synthetic-policy_compliance-000567", "reversibility": 0.6324, "risk": 0.6126, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 14.292, "id": "worker_a", "p95_latency_ms": 81136, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.933, "id": "worker_b", "p95_latency_ms": 30812, "pass_rate": 0.5906, "provider_group": "provider_y", "quality": 0.6792, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.182, "id": "worker_c", "p95_latency_ms": 71207, "pass_rate": 0.7619, "provider_group": "provider_z", "quality": 0.8856, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.691, "max_latency_ms": 38497, "minimum_quality": 0.6518}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000323", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.153, "estimated_latency_ms": 29570, "estimated_pass_rate": 0.3607, "estimated_quality": 0.6462, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.691, "max_latency_ms": 38497, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3607, "quality": 0.6462, "quality_target_met": false, "utility": 39.895881}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.691,\"max_latency_ms\":38497,\"minimum_quality\":0.6518},\"task\":{\"ambiguity\":0.3398,\"category\":\"analysis\",\"complexity\":0.5159,\"consequence\":0.5833,\"evidence_need\":0.5156,\"family\":\"performance_tuning\",\"input_tokens\":4240,\"reversibility\":0.8984,\"risk\":0.5868,\"summary\":\"Choose between throughput and tail-latency changes for provider failover.\"},\"workers\":[{\"credits_per_call\":1.153,\"id\":\"worker_a\",\"p95_latency_ms\":29570,\"pass_rate\":0.4924,\"provider_group\":\"provider_x\",\"quality\":0.7246,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":13.628,\"id\":\"worker_b\",\"p95_latency_ms\":57560,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.753,\"id\":\"worker_c\",\"p95_latency_ms\":57444,\"pass_rate\":0.8656,\"provider_group\":\"provider_y\",\"quality\":0.8644,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.153,\"estimated_latency_ms\":29570,\"estimated_pass_rate\":0.3607,\"estimated_quality\":0.6462,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.691,\"max_latency_ms\":38497,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3398, "category": "analysis", "complexity": 0.5159, "consequence": 0.5833, "evidence_need": 0.5156, "family": "performance_tuning", "input_tokens": 4240, "request_id": "synthetic-performance_tuning-000323", "reversibility": 0.8984, "risk": 0.5868, "summary": "Choose between throughput and tail-latency changes for provider failover."}, "workers": [{"credits_per_call": 1.153, "id": "worker_a", "p95_latency_ms": 29570, "pass_rate": 0.4924, "provider_group": "provider_x", "quality": 0.7246, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 13.628, "id": "worker_b", "p95_latency_ms": 57560, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.753, "id": "worker_c", "p95_latency_ms": 57444, "pass_rate": 0.8656, "provider_group": "provider_y", "quality": 0.8644, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.612, "max_latency_ms": 26802, "minimum_quality": 0.7766}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000476", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.119, "estimated_latency_ms": 20864, "estimated_pass_rate": 0.3572, "estimated_quality": 0.557, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.612, "max_latency_ms": 26802, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3572, "quality": 0.557, "quality_target_met": false, "utility": 25.164197}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.612,\"max_latency_ms\":26802,\"minimum_quality\":0.7766},\"task\":{\"ambiguity\":0.4713,\"category\":\"analysis\",\"complexity\":0.6198,\"consequence\":0.6307,\"evidence_need\":0.6129,\"family\":\"performance_tuning\",\"input_tokens\":1880,\"reversibility\":0.6911,\"risk\":0.2097,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":15.043,\"id\":\"worker_a\",\"p95_latency_ms\":55320,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9865,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.119,\"id\":\"worker_b\",\"p95_latency_ms\":20864,\"pass_rate\":0.4801,\"provider_group\":\"provider_z\",\"quality\":0.6358,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.909,\"id\":\"worker_c\",\"p95_latency_ms\":69812,\"pass_rate\":0.8442,\"provider_group\":\"provider_x\",\"quality\":0.9769,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.119,\"estimated_latency_ms\":20864,\"estimated_pass_rate\":0.3572,\"estimated_quality\":0.557,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.612,\"max_latency_ms\":26802,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4713, "category": "analysis", "complexity": 0.6198, "consequence": 0.6307, "evidence_need": 0.6129, "family": "performance_tuning", "input_tokens": 1880, "request_id": "synthetic-performance_tuning-000476", "reversibility": 0.6911, "risk": 0.2097, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 15.043, "id": "worker_a", "p95_latency_ms": 55320, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9865, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.119, "id": "worker_b", "p95_latency_ms": 20864, "pass_rate": 0.4801, "provider_group": "provider_z", "quality": 0.6358, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.909, "id": "worker_c", "p95_latency_ms": 69812, "pass_rate": 0.8442, "provider_group": "provider_x", "quality": 0.9769, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.242, "max_latency_ms": 54306, "minimum_quality": 0.6804}, "family": "migration_planning", "id": "synthetic-migration_planning-000568", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.998, "estimated_latency_ms": 31800, "estimated_pass_rate": 0.39, "estimated_quality": 0.6709, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.242, "max_latency_ms": 54306, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.39, "quality": 0.6709, "quality_target_met": false, "utility": 41.906933}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.242,\"max_latency_ms\":54306,\"minimum_quality\":0.6804},\"task\":{\"ambiguity\":0.363,\"category\":\"coding\",\"complexity\":0.739,\"consequence\":0.8425,\"evidence_need\":0.586,\"family\":\"migration_planning\",\"input_tokens\":720,\"reversibility\":0.4503,\"risk\":0.4962,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.998,\"id\":\"worker_a\",\"p95_latency_ms\":31800,\"pass_rate\":0.5414,\"provider_group\":\"provider_y\",\"quality\":0.7846,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.313,\"id\":\"worker_b\",\"p95_latency_ms\":76557,\"pass_rate\":0.8052,\"provider_group\":\"provider_x\",\"quality\":0.8067,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.061,\"id\":\"worker_c\",\"p95_latency_ms\":84656,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9207,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.998,\"estimated_latency_ms\":31800,\"estimated_pass_rate\":0.39,\"estimated_quality\":0.6709,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.242,\"max_latency_ms\":54306,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.363, "category": "coding", "complexity": 0.739, "consequence": 0.8425, "evidence_need": 0.586, "family": "migration_planning", "input_tokens": 720, "request_id": "synthetic-migration_planning-000568", "reversibility": 0.4503, "risk": 0.4962, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 0.998, "id": "worker_a", "p95_latency_ms": 31800, "pass_rate": 0.5414, "provider_group": "provider_y", "quality": 0.7846, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.313, "id": "worker_b", "p95_latency_ms": 76557, "pass_rate": 0.8052, "provider_group": "provider_x", "quality": 0.8067, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.061, "id": "worker_c", "p95_latency_ms": 84656, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9207, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.276, "max_latency_ms": 141281, "minimum_quality": 0.8096}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000875", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.453, "estimated_latency_ms": 103536, "estimated_pass_rate": 0.8221, "estimated_quality": 0.8406, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.276, "max_latency_ms": 141281, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.8221, "quality": 0.8406, "quality_target_met": true, "utility": 62.789504}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.276,\"max_latency_ms\":141281,\"minimum_quality\":0.8096},\"task\":{\"ambiguity\":0.7252,\"category\":\"analysis\",\"complexity\":0.6917,\"consequence\":0.3988,\"evidence_need\":0.7332,\"family\":\"performance_tuning\",\"input_tokens\":4160,\"reversibility\":0.6331,\"risk\":0.4538,\"summary\":\"Choose between throughput and tail-latency changes for model routing.\"},\"workers\":[{\"credits_per_call\":0.942,\"id\":\"worker_a\",\"p95_latency_ms\":32678,\"pass_rate\":0.5509,\"provider_group\":\"provider_x\",\"quality\":0.7396,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.511,\"id\":\"worker_b\",\"p95_latency_ms\":70858,\"pass_rate\":0.953,\"provider_group\":\"provider_y\",\"quality\":0.9233,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.126,\"id\":\"worker_c\",\"p95_latency_ms\":74272,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.453,\"estimated_latency_ms\":103536,\"estimated_pass_rate\":0.8221,\"estimated_quality\":0.8406,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.276,\"max_latency_ms\":141281,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7252, "category": "analysis", "complexity": 0.6917, "consequence": 0.3988, "evidence_need": 0.7332, "family": "performance_tuning", "input_tokens": 4160, "request_id": "synthetic-performance_tuning-000875", "reversibility": 0.6331, "risk": 0.4538, "summary": "Choose between throughput and tail-latency changes for model routing."}, "workers": [{"credits_per_call": 0.942, "id": "worker_a", "p95_latency_ms": 32678, "pass_rate": 0.5509, "provider_group": "provider_x", "quality": 0.7396, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.511, "id": "worker_b", "p95_latency_ms": 70858, "pass_rate": 0.953, "provider_group": "provider_y", "quality": 0.9233, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.126, "id": "worker_c", "p95_latency_ms": 74272, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.681, "max_latency_ms": 151091, "minimum_quality": 0.7258}, "family": "migration_planning", "id": "synthetic-migration_planning-000145", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.566, "estimated_latency_ms": 83850, "estimated_pass_rate": 0.581, "estimated_quality": 0.8297, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.681, "max_latency_ms": 151091, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.581, "quality": 0.8297, "quality_target_met": true, "utility": 62.851863}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.681,\"max_latency_ms\":151091,\"minimum_quality\":0.7258},\"task\":{\"ambiguity\":0.4766,\"category\":\"coding\",\"complexity\":0.7666,\"consequence\":0.6099,\"evidence_need\":0.8476,\"family\":\"migration_planning\",\"input_tokens\":2480,\"reversibility\":0.4616,\"risk\":0.6697,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.872,\"id\":\"worker_a\",\"p95_latency_ms\":27532,\"pass_rate\":0.4785,\"provider_group\":\"provider_y\",\"quality\":0.784,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.694,\"id\":\"worker_b\",\"p95_latency_ms\":56318,\"pass_rate\":0.823,\"provider_group\":\"provider_z\",\"quality\":0.8294,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.49,\"id\":\"worker_c\",\"p95_latency_ms\":90343,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.566,\"estimated_latency_ms\":83850,\"estimated_pass_rate\":0.581,\"estimated_quality\":0.8297,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.681,\"max_latency_ms\":151091,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4766, "category": "coding", "complexity": 0.7666, "consequence": 0.6099, "evidence_need": 0.8476, "family": "migration_planning", "input_tokens": 2480, "request_id": "synthetic-migration_planning-000145", "reversibility": 0.4616, "risk": 0.6697, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 0.872, "id": "worker_a", "p95_latency_ms": 27532, "pass_rate": 0.4785, "provider_group": "provider_y", "quality": 0.784, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.694, "id": "worker_b", "p95_latency_ms": 56318, "pass_rate": 0.823, "provider_group": "provider_z", "quality": 0.8294, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.49, "id": "worker_c", "p95_latency_ms": 90343, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.73, "max_latency_ms": 168981, "minimum_quality": 0.7969}, "family": "migration_planning", "id": "synthetic-migration_planning-000907", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.823, "estimated_latency_ms": 155755, "estimated_pass_rate": 0.6785, "estimated_quality": 0.9105, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.73, "max_latency_ms": 168981, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6785, "quality": 0.9105, "quality_target_met": true, "utility": 72.254146}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.73,\"max_latency_ms\":168981,\"minimum_quality\":0.7969},\"task\":{\"ambiguity\":0.6456,\"category\":\"coding\",\"complexity\":0.8749,\"consequence\":0.5192,\"evidence_need\":0.4614,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.3581,\"risk\":0.6613,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":3.852,\"id\":\"worker_a\",\"p95_latency_ms\":68075,\"pass_rate\":0.835,\"provider_group\":\"provider_y\",\"quality\":0.8114,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.924,\"id\":\"worker_b\",\"p95_latency_ms\":63215,\"pass_rate\":0.8863,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.047,\"id\":\"worker_c\",\"p95_latency_ms\":24465,\"pass_rate\":0.459,\"provider_group\":\"provider_x\",\"quality\":0.8004,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.823,\"estimated_latency_ms\":155755,\"estimated_pass_rate\":0.6785,\"estimated_quality\":0.9105,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.73,\"max_latency_ms\":168981,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6456, "category": "coding", "complexity": 0.8749, "consequence": 0.5192, "evidence_need": 0.4614, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-000907", "reversibility": 0.3581, "risk": 0.6613, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 3.852, "id": "worker_a", "p95_latency_ms": 68075, "pass_rate": 0.835, "provider_group": "provider_y", "quality": 0.8114, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.924, "id": "worker_b", "p95_latency_ms": 63215, "pass_rate": 0.8863, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.047, "id": "worker_c", "p95_latency_ms": 24465, "pass_rate": 0.459, "provider_group": "provider_x", "quality": 0.8004, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.427, "max_latency_ms": 44212, "minimum_quality": 0.55}, "family": "migration_planning", "id": "synthetic-migration_planning-000601", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.847, "estimated_latency_ms": 31392, "estimated_pass_rate": 0.4104, "estimated_quality": 0.6051, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.427, "max_latency_ms": 44212, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4104, "quality": 0.6051, "quality_target_met": true, "utility": 42.030165}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.427,\"max_latency_ms\":44212,\"minimum_quality\":0.55},\"task\":{\"ambiguity\":0.4163,\"category\":\"coding\",\"complexity\":0.5854,\"consequence\":0.7478,\"evidence_need\":0.4216,\"family\":\"migration_planning\",\"input_tokens\":4200,\"reversibility\":0.5912,\"risk\":0.5147,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.847,\"id\":\"worker_a\",\"p95_latency_ms\":31392,\"pass_rate\":0.5499,\"provider_group\":\"provider_z\",\"quality\":0.696,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.404,\"id\":\"worker_b\",\"p95_latency_ms\":87307,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.976,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.99,\"id\":\"worker_c\",\"p95_latency_ms\":64823,\"pass_rate\":0.7935,\"provider_group\":\"provider_x\",\"quality\":0.9083,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.847,\"estimated_latency_ms\":31392,\"estimated_pass_rate\":0.4104,\"estimated_quality\":0.6051,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.427,\"max_latency_ms\":44212,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4163, "category": "coding", "complexity": 0.5854, "consequence": 0.7478, "evidence_need": 0.4216, "family": "migration_planning", "input_tokens": 4200, "request_id": "synthetic-migration_planning-000601", "reversibility": 0.5912, "risk": 0.5147, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 0.847, "id": "worker_a", "p95_latency_ms": 31392, "pass_rate": 0.5499, "provider_group": "provider_z", "quality": 0.696, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.404, "id": "worker_b", "p95_latency_ms": 87307, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.976, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.99, "id": "worker_c", "p95_latency_ms": 64823, "pass_rate": 0.7935, "provider_group": "provider_x", "quality": 0.9083, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.235, "max_latency_ms": 104193, "minimum_quality": 0.8901}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000177", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.627, "estimated_latency_ms": 88683, "estimated_pass_rate": 0.7018, "estimated_quality": 0.8882, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.235, "max_latency_ms": 104193, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7018, "quality": 0.8882, "quality_target_met": false, "utility": 64.662198}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.235,\"max_latency_ms\":104193,\"minimum_quality\":0.8901},\"task\":{\"ambiguity\":0.4878,\"category\":\"review\",\"complexity\":0.4759,\"consequence\":0.5737,\"evidence_need\":0.8979,\"family\":\"policy_compliance\",\"input_tokens\":7720,\"reversibility\":0.529,\"risk\":0.5609,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":0.971,\"id\":\"worker_a\",\"p95_latency_ms\":18667,\"pass_rate\":0.5463,\"provider_group\":\"provider_y\",\"quality\":0.7188,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.656,\"id\":\"worker_b\",\"p95_latency_ms\":70016,\"pass_rate\":0.9698,\"provider_group\":\"provider_x\",\"quality\":0.9261,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.835,\"id\":\"worker_c\",\"p95_latency_ms\":71584,\"pass_rate\":0.8833,\"provider_group\":\"provider_z\",\"quality\":0.8137,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.627,\"estimated_latency_ms\":88683,\"estimated_pass_rate\":0.7018,\"estimated_quality\":0.8882,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.235,\"max_latency_ms\":104193,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4878, "category": "review", "complexity": 0.4759, "consequence": 0.5737, "evidence_need": 0.8979, "family": "policy_compliance", "input_tokens": 7720, "request_id": "synthetic-policy_compliance-000177", "reversibility": 0.529, "risk": 0.5609, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 0.971, "id": "worker_a", "p95_latency_ms": 18667, "pass_rate": 0.5463, "provider_group": "provider_y", "quality": 0.7188, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.656, "id": "worker_b", "p95_latency_ms": 70016, "pass_rate": 0.9698, "provider_group": "provider_x", "quality": 0.9261, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.835, "id": "worker_c", "p95_latency_ms": 71584, "pass_rate": 0.8833, "provider_group": "provider_z", "quality": 0.8137, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.568, "max_latency_ms": 25913, "minimum_quality": 0.5468}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000732", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.938, "estimated_latency_ms": 18320, "estimated_pass_rate": 0.4021, "estimated_quality": 0.5868, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.568, "max_latency_ms": 25913, "policy": "lean", "reason_codes": ["complex_task", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4021, "quality": 0.5868, "quality_target_met": true, "utility": 23.19601}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.568,\"max_latency_ms\":25913,\"minimum_quality\":0.5468},\"task\":{\"ambiguity\":0.4451,\"category\":\"review\",\"complexity\":0.824,\"consequence\":0.7884,\"evidence_need\":0.5942,\"family\":\"policy_compliance\",\"input_tokens\":7640,\"reversibility\":0.4079,\"risk\":0.8349,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.938,\"id\":\"worker_a\",\"p95_latency_ms\":18320,\"pass_rate\":0.5968,\"provider_group\":\"provider_y\",\"quality\":0.7458,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.474,\"id\":\"worker_b\",\"p95_latency_ms\":66354,\"pass_rate\":0.9169,\"provider_group\":\"provider_x\",\"quality\":0.9105,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.027,\"id\":\"worker_c\",\"p95_latency_ms\":47862,\"pass_rate\":0.8683,\"provider_group\":\"provider_z\",\"quality\":0.9069,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.938,\"estimated_latency_ms\":18320,\"estimated_pass_rate\":0.4021,\"estimated_quality\":0.5868,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.568,\"max_latency_ms\":25913,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4451, "category": "review", "complexity": 0.824, "consequence": 0.7884, "evidence_need": 0.5942, "family": "policy_compliance", "input_tokens": 7640, "request_id": "synthetic-policy_compliance-000732", "reversibility": 0.4079, "risk": 0.8349, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 0.938, "id": "worker_a", "p95_latency_ms": 18320, "pass_rate": 0.5968, "provider_group": "provider_y", "quality": 0.7458, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.474, "id": "worker_b", "p95_latency_ms": 66354, "pass_rate": 0.9169, "provider_group": "provider_x", "quality": 0.9105, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.027, "id": "worker_c", "p95_latency_ms": 47862, "pass_rate": 0.8683, "provider_group": "provider_z", "quality": 0.9069, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.766, "max_latency_ms": 170033, "minimum_quality": 0.8175}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000939", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.625, "estimated_latency_ms": 150765, "estimated_pass_rate": 0.7323, "estimated_quality": 0.912, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.766, "max_latency_ms": 170033, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7323, "quality": 0.912, "quality_target_met": true, "utility": 77.446234}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.766,\"max_latency_ms\":170033,\"minimum_quality\":0.8175},\"task\":{\"ambiguity\":0.521,\"category\":\"review\",\"complexity\":0.8284,\"consequence\":0.9021,\"evidence_need\":0.8857,\"family\":\"policy_compliance\",\"input_tokens\":4320,\"reversibility\":0.4816,\"risk\":0.5426,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":12.59,\"id\":\"worker_a\",\"p95_latency_ms\":62973,\"pass_rate\":0.8871,\"provider_group\":\"provider_y\",\"quality\":0.9722,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.189,\"id\":\"worker_b\",\"p95_latency_ms\":17611,\"pass_rate\":0.5764,\"provider_group\":\"provider_x\",\"quality\":0.7354,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.846,\"id\":\"worker_c\",\"p95_latency_ms\":70181,\"pass_rate\":0.8072,\"provider_group\":\"provider_z\",\"quality\":0.9042,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.625,\"estimated_latency_ms\":150765,\"estimated_pass_rate\":0.7323,\"estimated_quality\":0.912,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.766,\"max_latency_ms\":170033,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.521, "category": "review", "complexity": 0.8284, "consequence": 0.9021, "evidence_need": 0.8857, "family": "policy_compliance", "input_tokens": 4320, "request_id": "synthetic-policy_compliance-000939", "reversibility": 0.4816, "risk": 0.5426, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 12.59, "id": "worker_a", "p95_latency_ms": 62973, "pass_rate": 0.8871, "provider_group": "provider_y", "quality": 0.9722, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.189, "id": "worker_b", "p95_latency_ms": 17611, "pass_rate": 0.5764, "provider_group": "provider_x", "quality": 0.7354, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.846, "id": "worker_c", "p95_latency_ms": 70181, "pass_rate": 0.8072, "provider_group": "provider_z", "quality": 0.9042, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.269, "max_latency_ms": 25732, "minimum_quality": 0.5239}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000890", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.209, "estimated_latency_ms": 19048, "estimated_pass_rate": 0.3477, "estimated_quality": 0.5356, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.269, "max_latency_ms": 25732, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3477, "quality": 0.5356, "quality_target_met": true, "utility": 34.143358}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.269,\"max_latency_ms\":25732,\"minimum_quality\":0.5239},\"task\":{\"ambiguity\":0.3136,\"category\":\"analysis\",\"complexity\":0.6464,\"consequence\":0.367,\"evidence_need\":0.6103,\"family\":\"performance_tuning\",\"input_tokens\":5640,\"reversibility\":0.7803,\"risk\":0.555,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":13.337,\"id\":\"worker_a\",\"p95_latency_ms\":77968,\"pass_rate\":0.9683,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.709,\"id\":\"worker_b\",\"p95_latency_ms\":77665,\"pass_rate\":0.8478,\"provider_group\":\"provider_y\",\"quality\":0.9495,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.209,\"id\":\"worker_c\",\"p95_latency_ms\":19048,\"pass_rate\":0.4897,\"provider_group\":\"provider_z\",\"quality\":0.6335,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.209,\"estimated_latency_ms\":19048,\"estimated_pass_rate\":0.3477,\"estimated_quality\":0.5356,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.269,\"max_latency_ms\":25732,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3136, "category": "analysis", "complexity": 0.6464, "consequence": 0.367, "evidence_need": 0.6103, "family": "performance_tuning", "input_tokens": 5640, "request_id": "synthetic-performance_tuning-000890", "reversibility": 0.7803, "risk": 0.555, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 13.337, "id": "worker_a", "p95_latency_ms": 77968, "pass_rate": 0.9683, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.709, "id": "worker_b", "p95_latency_ms": 77665, "pass_rate": 0.8478, "provider_group": "provider_y", "quality": 0.9495, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.209, "id": "worker_c", "p95_latency_ms": 19048, "pass_rate": 0.4897, "provider_group": "provider_z", "quality": 0.6335, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.296, "max_latency_ms": 262980, "minimum_quality": 0.8792}, "family": "migration_planning", "id": "synthetic-migration_planning-000229", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.06, "estimated_latency_ms": 112233, "estimated_pass_rate": 0.6927, "estimated_quality": 0.8043, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.296, "max_latency_ms": 262980, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6927, "quality": 0.8043, "quality_target_met": false, "utility": 58.23805}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.296,\"max_latency_ms\":262980,\"minimum_quality\":0.8792},\"task\":{\"ambiguity\":0.6273,\"category\":\"coding\",\"complexity\":0.8849,\"consequence\":0.8397,\"evidence_need\":0.5415,\"family\":\"migration_planning\",\"input_tokens\":7760,\"reversibility\":0.4061,\"risk\":0.7365,\"summary\":\"Design compatibility testing for a background jobs schema migration.\"},\"workers\":[{\"credits_per_call\":13.01,\"id\":\"worker_a\",\"p95_latency_ms\":86974,\"pass_rate\":0.9164,\"provider_group\":\"provider_y\",\"quality\":0.9377,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.855,\"id\":\"worker_b\",\"p95_latency_ms\":81313,\"pass_rate\":0.737,\"provider_group\":\"provider_z\",\"quality\":0.9189,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.05,\"id\":\"worker_c\",\"p95_latency_ms\":25259,\"pass_rate\":0.5614,\"provider_group\":\"provider_x\",\"quality\":0.7163,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.06,\"estimated_latency_ms\":112233,\"estimated_pass_rate\":0.6927,\"estimated_quality\":0.8043,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.296,\"max_latency_ms\":262980,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6273, "category": "coding", "complexity": 0.8849, "consequence": 0.8397, "evidence_need": 0.5415, "family": "migration_planning", "input_tokens": 7760, "request_id": "synthetic-migration_planning-000229", "reversibility": 0.4061, "risk": 0.7365, "summary": "Design compatibility testing for a background jobs schema migration."}, "workers": [{"credits_per_call": 13.01, "id": "worker_a", "p95_latency_ms": 86974, "pass_rate": 0.9164, "provider_group": "provider_y", "quality": 0.9377, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.855, "id": "worker_b", "p95_latency_ms": 81313, "pass_rate": 0.737, "provider_group": "provider_z", "quality": 0.9189, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.05, "id": "worker_c", "p95_latency_ms": 25259, "pass_rate": 0.5614, "provider_group": "provider_x", "quality": 0.7163, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.973, "max_latency_ms": 140860, "minimum_quality": 0.897}, "family": "migration_planning", "id": "synthetic-migration_planning-000718", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.722, "estimated_latency_ms": 101459, "estimated_pass_rate": 0.575, "estimated_quality": 0.7653, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.973, "max_latency_ms": 140860, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.575, "quality": 0.7653, "quality_target_met": false, "utility": 51.064689}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.973,\"max_latency_ms\":140860,\"minimum_quality\":0.897},\"task\":{\"ambiguity\":0.5118,\"category\":\"coding\",\"complexity\":0.719,\"consequence\":0.5203,\"evidence_need\":0.577,\"family\":\"migration_planning\",\"input_tokens\":3200,\"reversibility\":0.2579,\"risk\":0.601,\"summary\":\"Plan a zero-downtime migration of background jobs with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.752,\"id\":\"worker_a\",\"p95_latency_ms\":71936,\"pass_rate\":0.8548,\"provider_group\":\"provider_y\",\"quality\":0.8019,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.342,\"id\":\"worker_b\",\"p95_latency_ms\":72381,\"pass_rate\":0.8881,\"provider_group\":\"provider_z\",\"quality\":0.959,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.97,\"id\":\"worker_c\",\"p95_latency_ms\":29523,\"pass_rate\":0.4766,\"provider_group\":\"provider_x\",\"quality\":0.7056,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.722,\"estimated_latency_ms\":101459,\"estimated_pass_rate\":0.575,\"estimated_quality\":0.7653,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.973,\"max_latency_ms\":140860,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5118, "category": "coding", "complexity": 0.719, "consequence": 0.5203, "evidence_need": 0.577, "family": "migration_planning", "input_tokens": 3200, "request_id": "synthetic-migration_planning-000718", "reversibility": 0.2579, "risk": 0.601, "summary": "Plan a zero-downtime migration of background jobs with rollback checkpoints."}, "workers": [{"credits_per_call": 2.752, "id": "worker_a", "p95_latency_ms": 71936, "pass_rate": 0.8548, "provider_group": "provider_y", "quality": 0.8019, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.342, "id": "worker_b", "p95_latency_ms": 72381, "pass_rate": 0.8881, "provider_group": "provider_z", "quality": 0.959, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.97, "id": "worker_c", "p95_latency_ms": 29523, "pass_rate": 0.4766, "provider_group": "provider_x", "quality": 0.7056, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.847, "max_latency_ms": 192032, "minimum_quality": 0.6901}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001188", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.855, "estimated_latency_ms": 107556, "estimated_pass_rate": 0.5951, "estimated_quality": 0.8407, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.847, "max_latency_ms": 192032, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5951, "quality": 0.8407, "quality_target_met": true, "utility": 63.858995}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.847,\"max_latency_ms\":192032,\"minimum_quality\":0.6901},\"task\":{\"ambiguity\":0.3742,\"category\":\"review\",\"complexity\":0.7402,\"consequence\":0.903,\"evidence_need\":0.7729,\"family\":\"policy_compliance\",\"input_tokens\":1440,\"reversibility\":0.3971,\"risk\":0.6075,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.666,\"id\":\"worker_a\",\"p95_latency_ms\":74601,\"pass_rate\":0.8835,\"provider_group\":\"provider_z\",\"quality\":0.8096,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.018,\"id\":\"worker_b\",\"p95_latency_ms\":58727,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9684,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.189,\"id\":\"worker_c\",\"p95_latency_ms\":32955,\"pass_rate\":0.4813,\"provider_group\":\"provider_x\",\"quality\":0.7925,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.855,\"estimated_latency_ms\":107556,\"estimated_pass_rate\":0.5951,\"estimated_quality\":0.8407,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.847,\"max_latency_ms\":192032,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3742, "category": "review", "complexity": 0.7402, "consequence": 0.903, "evidence_need": 0.7729, "family": "policy_compliance", "input_tokens": 1440, "request_id": "synthetic-policy_compliance-001188", "reversibility": 0.3971, "risk": 0.6075, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 3.666, "id": "worker_a", "p95_latency_ms": 74601, "pass_rate": 0.8835, "provider_group": "provider_z", "quality": 0.8096, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.018, "id": "worker_b", "p95_latency_ms": 58727, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9684, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.189, "id": "worker_c", "p95_latency_ms": 32955, "pass_rate": 0.4813, "provider_group": "provider_x", "quality": 0.7925, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.871, "max_latency_ms": 209548, "minimum_quality": 0.7812}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001041", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.43, "estimated_latency_ms": 160000, "estimated_pass_rate": 0.699, "estimated_quality": 0.9047, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.871, "max_latency_ms": 209548, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.699, "quality": 0.9047, "quality_target_met": true, "utility": 75.599538}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.871,\"max_latency_ms\":209548,\"minimum_quality\":0.7812},\"task\":{\"ambiguity\":0.5435,\"category\":\"review\",\"complexity\":0.7669,\"consequence\":0.8996,\"evidence_need\":0.8388,\"family\":\"policy_compliance\",\"input_tokens\":760,\"reversibility\":0.3083,\"risk\":0.5684,\"summary\":\"Identify evidence required before approving prompt caching for production.\"},\"workers\":[{\"credits_per_call\":2.943,\"id\":\"worker_a\",\"p95_latency_ms\":49442,\"pass_rate\":0.7202,\"provider_group\":\"provider_x\",\"quality\":0.831,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.56,\"id\":\"worker_b\",\"p95_latency_ms\":89969,\"pass_rate\":0.9727,\"provider_group\":\"provider_y\",\"quality\":0.9758,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.927,\"id\":\"worker_c\",\"p95_latency_ms\":20589,\"pass_rate\":0.4747,\"provider_group\":\"provider_z\",\"quality\":0.7415,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.43,\"estimated_latency_ms\":160000,\"estimated_pass_rate\":0.699,\"estimated_quality\":0.9047,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.871,\"max_latency_ms\":209548,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5435, "category": "review", "complexity": 0.7669, "consequence": 0.8996, "evidence_need": 0.8388, "family": "policy_compliance", "input_tokens": 760, "request_id": "synthetic-policy_compliance-001041", "reversibility": 0.3083, "risk": 0.5684, "summary": "Identify evidence required before approving prompt caching for production."}, "workers": [{"credits_per_call": 2.943, "id": "worker_a", "p95_latency_ms": 49442, "pass_rate": 0.7202, "provider_group": "provider_x", "quality": 0.831, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.56, "id": "worker_b", "p95_latency_ms": 89969, "pass_rate": 0.9727, "provider_group": "provider_y", "quality": 0.9758, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.927, "id": "worker_c", "p95_latency_ms": 20589, "pass_rate": 0.4747, "provider_group": "provider_z", "quality": 0.7415, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.28, "max_latency_ms": 234837, "minimum_quality": 0.778}, "family": "migration_planning", "id": "synthetic-migration_planning-001063", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.15, "estimated_latency_ms": 97311, "estimated_pass_rate": 0.6504, "estimated_quality": 0.7531, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.28, "max_latency_ms": 234837, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6504, "quality": 0.7531, "quality_target_met": false, "utility": 54.487475}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.28,\"max_latency_ms\":234837,\"minimum_quality\":0.778},\"task\":{\"ambiguity\":0.4756,\"category\":\"coding\",\"complexity\":0.7066,\"consequence\":0.6741,\"evidence_need\":0.6077,\"family\":\"migration_planning\",\"input_tokens\":6960,\"reversibility\":0.3738,\"risk\":0.5012,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":4.122,\"id\":\"worker_a\",\"p95_latency_ms\":78218,\"pass_rate\":0.7198,\"provider_group\":\"provider_y\",\"quality\":0.7985,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.028,\"id\":\"worker_b\",\"p95_latency_ms\":19093,\"pass_rate\":0.5729,\"provider_group\":\"provider_z\",\"quality\":0.6905,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.586,\"id\":\"worker_c\",\"p95_latency_ms\":80568,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9391,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.15,\"estimated_latency_ms\":97311,\"estimated_pass_rate\":0.6504,\"estimated_quality\":0.7531,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.28,\"max_latency_ms\":234837,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4756, "category": "coding", "complexity": 0.7066, "consequence": 0.6741, "evidence_need": 0.6077, "family": "migration_planning", "input_tokens": 6960, "request_id": "synthetic-migration_planning-001063", "reversibility": 0.3738, "risk": 0.5012, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 4.122, "id": "worker_a", "p95_latency_ms": 78218, "pass_rate": 0.7198, "provider_group": "provider_y", "quality": 0.7985, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.028, "id": "worker_b", "p95_latency_ms": 19093, "pass_rate": 0.5729, "provider_group": "provider_z", "quality": 0.6905, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.586, "id": "worker_c", "p95_latency_ms": 80568, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9391, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.696, "max_latency_ms": 27539, "minimum_quality": 0.7252}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000615", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.166, "estimated_latency_ms": 24466, "estimated_pass_rate": 0.4039, "estimated_quality": 0.5855, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.696, "max_latency_ms": 27539, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4039, "quality": 0.5855, "quality_target_met": false, "utility": 32.697994}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.696,\"max_latency_ms\":27539,\"minimum_quality\":0.7252},\"task\":{\"ambiguity\":0.5936,\"category\":\"review\",\"complexity\":0.5849,\"consequence\":0.5357,\"evidence_need\":0.9,\"family\":\"policy_compliance\",\"input_tokens\":7680,\"reversibility\":0.6934,\"risk\":0.7894,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":3.261,\"id\":\"worker_a\",\"p95_latency_ms\":45796,\"pass_rate\":0.7481,\"provider_group\":\"provider_y\",\"quality\":0.9055,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.611,\"id\":\"worker_b\",\"p95_latency_ms\":66602,\"pass_rate\":0.9042,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.166,\"id\":\"worker_c\",\"p95_latency_ms\":24466,\"pass_rate\":0.579,\"provider_group\":\"provider_z\",\"quality\":0.708,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.166,\"estimated_latency_ms\":24466,\"estimated_pass_rate\":0.4039,\"estimated_quality\":0.5855,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.696,\"max_latency_ms\":27539,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5936, "category": "review", "complexity": 0.5849, "consequence": 0.5357, "evidence_need": 0.9, "family": "policy_compliance", "input_tokens": 7680, "request_id": "synthetic-policy_compliance-000615", "reversibility": 0.6934, "risk": 0.7894, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 3.261, "id": "worker_a", "p95_latency_ms": 45796, "pass_rate": 0.7481, "provider_group": "provider_y", "quality": 0.9055, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.611, "id": "worker_b", "p95_latency_ms": 66602, "pass_rate": 0.9042, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.166, "id": "worker_c", "p95_latency_ms": 24466, "pass_rate": 0.579, "provider_group": "provider_z", "quality": 0.708, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.694, "max_latency_ms": 159149, "minimum_quality": 0.9123}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001115", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.649, "estimated_latency_ms": 153279, "estimated_pass_rate": 0.713, "estimated_quality": 0.9287, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.694, "max_latency_ms": 159149, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.713, "quality": 0.9287, "quality_target_met": true, "utility": 70.442339}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.694,\"max_latency_ms\":159149,\"minimum_quality\":0.9123},\"task\":{\"ambiguity\":0.5036,\"category\":\"analysis\",\"complexity\":0.8881,\"consequence\":0.3858,\"evidence_need\":0.7097,\"family\":\"performance_tuning\",\"input_tokens\":5280,\"reversibility\":0.8995,\"risk\":0.2854,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.989,\"id\":\"worker_a\",\"p95_latency_ms\":19291,\"pass_rate\":0.4252,\"provider_group\":\"provider_z\",\"quality\":0.7267,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.895,\"id\":\"worker_b\",\"p95_latency_ms\":72073,\"pass_rate\":0.8799,\"provider_group\":\"provider_y\",\"quality\":0.9348,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.765,\"id\":\"worker_c\",\"p95_latency_ms\":61915,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.649,\"estimated_latency_ms\":153279,\"estimated_pass_rate\":0.713,\"estimated_quality\":0.9287,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.694,\"max_latency_ms\":159149,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5036, "category": "analysis", "complexity": 0.8881, "consequence": 0.3858, "evidence_need": 0.7097, "family": "performance_tuning", "input_tokens": 5280, "request_id": "synthetic-performance_tuning-001115", "reversibility": 0.8995, "risk": 0.2854, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 0.989, "id": "worker_a", "p95_latency_ms": 19291, "pass_rate": 0.4252, "provider_group": "provider_z", "quality": 0.7267, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.895, "id": "worker_b", "p95_latency_ms": 72073, "pass_rate": 0.8799, "provider_group": "provider_y", "quality": 0.9348, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.765, "id": "worker_c", "p95_latency_ms": 61915, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.168, "max_latency_ms": 52638, "minimum_quality": 0.5297}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000639", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.958, "estimated_latency_ms": 26444, "estimated_pass_rate": 0.332, "estimated_quality": 0.6035, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.168, "max_latency_ms": 52638, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.332, "quality": 0.6035, "quality_target_met": true, "utility": 40.675521}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.168,\"max_latency_ms\":52638,\"minimum_quality\":0.5297},\"task\":{\"ambiguity\":0.5061,\"category\":\"review\",\"complexity\":0.8375,\"consequence\":0.6808,\"evidence_need\":0.654,\"family\":\"policy_compliance\",\"input_tokens\":3280,\"reversibility\":0.4189,\"risk\":0.6961,\"summary\":\"Review whether schema evolution satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.937,\"id\":\"worker_a\",\"p95_latency_ms\":85369,\"pass_rate\":0.8988,\"provider_group\":\"provider_x\",\"quality\":0.9104,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.839,\"id\":\"worker_b\",\"p95_latency_ms\":53381,\"pass_rate\":0.7951,\"provider_group\":\"provider_z\",\"quality\":0.8057,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.958,\"id\":\"worker_c\",\"p95_latency_ms\":26444,\"pass_rate\":0.5216,\"provider_group\":\"provider_y\",\"quality\":0.759,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.958,\"estimated_latency_ms\":26444,\"estimated_pass_rate\":0.332,\"estimated_quality\":0.6035,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.168,\"max_latency_ms\":52638,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5061, "category": "review", "complexity": 0.8375, "consequence": 0.6808, "evidence_need": 0.654, "family": "policy_compliance", "input_tokens": 3280, "request_id": "synthetic-policy_compliance-000639", "reversibility": 0.4189, "risk": 0.6961, "summary": "Review whether schema evolution satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.937, "id": "worker_a", "p95_latency_ms": 85369, "pass_rate": 0.8988, "provider_group": "provider_x", "quality": 0.9104, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.839, "id": "worker_b", "p95_latency_ms": 53381, "pass_rate": 0.7951, "provider_group": "provider_z", "quality": 0.8057, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.958, "id": "worker_c", "p95_latency_ms": 26444, "pass_rate": 0.5216, "provider_group": "provider_y", "quality": 0.759, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.698, "max_latency_ms": 33083, "minimum_quality": 0.5722}, "family": "migration_planning", "id": "synthetic-migration_planning-000190", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.068, "estimated_latency_ms": 24332, "estimated_pass_rate": 0.3251, "estimated_quality": 0.5505, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.698, "max_latency_ms": 33083, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3251, "quality": 0.5505, "quality_target_met": false, "utility": 32.984397}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.698,\"max_latency_ms\":33083,\"minimum_quality\":0.5722},\"task\":{\"ambiguity\":0.6457,\"category\":\"coding\",\"complexity\":0.8153,\"consequence\":0.5079,\"evidence_need\":0.7225,\"family\":\"migration_planning\",\"input_tokens\":3640,\"reversibility\":0.3742,\"risk\":0.5651,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.068,\"id\":\"worker_a\",\"p95_latency_ms\":24332,\"pass_rate\":0.5121,\"provider_group\":\"provider_x\",\"quality\":0.702,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.063,\"id\":\"worker_b\",\"p95_latency_ms\":64659,\"pass_rate\":0.8764,\"provider_group\":\"provider_z\",\"quality\":0.8096,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.508,\"id\":\"worker_c\",\"p95_latency_ms\":58808,\"pass_rate\":0.9772,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.068,\"estimated_latency_ms\":24332,\"estimated_pass_rate\":0.3251,\"estimated_quality\":0.5505,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.698,\"max_latency_ms\":33083,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6457, "category": "coding", "complexity": 0.8153, "consequence": 0.5079, "evidence_need": 0.7225, "family": "migration_planning", "input_tokens": 3640, "request_id": "synthetic-migration_planning-000190", "reversibility": 0.3742, "risk": 0.5651, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 1.068, "id": "worker_a", "p95_latency_ms": 24332, "pass_rate": 0.5121, "provider_group": "provider_x", "quality": 0.702, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.063, "id": "worker_b", "p95_latency_ms": 64659, "pass_rate": 0.8764, "provider_group": "provider_z", "quality": 0.8096, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.508, "id": "worker_c", "p95_latency_ms": 58808, "pass_rate": 0.9772, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 6.068, "max_latency_ms": 194768, "minimum_quality": 0.8998}, "family": "migration_planning", "id": "synthetic-migration_planning-001216", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.822, "estimated_latency_ms": 90784, "estimated_pass_rate": 0.5458, "estimated_quality": 0.76, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.068, "max_latency_ms": 194768, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5458, "quality": 0.76, "quality_target_met": false, "utility": 49.934912}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":6.068,\"max_latency_ms\":194768,\"minimum_quality\":0.8998},\"task\":{\"ambiguity\":0.6777,\"category\":\"coding\",\"complexity\":0.8549,\"consequence\":0.8252,\"evidence_need\":0.84,\"family\":\"migration_planning\",\"input_tokens\":3400,\"reversibility\":0.4395,\"risk\":0.582,\"summary\":\"Review a staged migration for artifact storage and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.737,\"id\":\"worker_a\",\"p95_latency_ms\":78835,\"pass_rate\":0.8933,\"provider_group\":\"provider_z\",\"quality\":0.9533,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.982,\"id\":\"worker_b\",\"p95_latency_ms\":69059,\"pass_rate\":0.8505,\"provider_group\":\"provider_x\",\"quality\":0.7973,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.84,\"id\":\"worker_c\",\"p95_latency_ms\":21725,\"pass_rate\":0.4656,\"provider_group\":\"provider_y\",\"quality\":0.6773,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.822,\"estimated_latency_ms\":90784,\"estimated_pass_rate\":0.5458,\"estimated_quality\":0.76,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.068,\"max_latency_ms\":194768,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6777, "category": "coding", "complexity": 0.8549, "consequence": 0.8252, "evidence_need": 0.84, "family": "migration_planning", "input_tokens": 3400, "request_id": "synthetic-migration_planning-001216", "reversibility": 0.4395, "risk": 0.582, "summary": "Review a staged migration for artifact storage and identify irreversible steps."}, "workers": [{"credits_per_call": 14.737, "id": "worker_a", "p95_latency_ms": 78835, "pass_rate": 0.8933, "provider_group": "provider_z", "quality": 0.9533, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.982, "id": "worker_b", "p95_latency_ms": 69059, "pass_rate": 0.8505, "provider_group": "provider_x", "quality": 0.7973, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.84, "id": "worker_c", "p95_latency_ms": 21725, "pass_rate": 0.4656, "provider_group": "provider_y", "quality": 0.6773, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.943, "max_latency_ms": 202292, "minimum_quality": 0.8128}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001107", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.129, "estimated_latency_ms": 173493, "estimated_pass_rate": 0.7607, "estimated_quality": 0.8983, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.943, "max_latency_ms": 202292, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7607, "quality": 0.8983, "quality_target_met": true, "utility": 74.912313}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.943,\"max_latency_ms\":202292,\"minimum_quality\":0.8128},\"task\":{\"ambiguity\":0.4326,\"category\":\"review\",\"complexity\":0.8435,\"consequence\":0.8675,\"evidence_need\":0.9014,\"family\":\"policy_compliance\",\"input_tokens\":1200,\"reversibility\":0.631,\"risk\":0.4508,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":14.037,\"id\":\"worker_a\",\"p95_latency_ms\":63739,\"pass_rate\":0.9851,\"provider_group\":\"provider_y\",\"quality\":0.922,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.095,\"id\":\"worker_b\",\"p95_latency_ms\":30732,\"pass_rate\":0.5482,\"provider_group\":\"provider_x\",\"quality\":0.7403,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.997,\"id\":\"worker_c\",\"p95_latency_ms\":79022,\"pass_rate\":0.8363,\"provider_group\":\"provider_z\",\"quality\":0.9075,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.129,\"estimated_latency_ms\":173493,\"estimated_pass_rate\":0.7607,\"estimated_quality\":0.8983,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.943,\"max_latency_ms\":202292,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4326, "category": "review", "complexity": 0.8435, "consequence": 0.8675, "evidence_need": 0.9014, "family": "policy_compliance", "input_tokens": 1200, "request_id": "synthetic-policy_compliance-001107", "reversibility": 0.631, "risk": 0.4508, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 14.037, "id": "worker_a", "p95_latency_ms": 63739, "pass_rate": 0.9851, "provider_group": "provider_y", "quality": 0.922, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.095, "id": "worker_b", "p95_latency_ms": 30732, "pass_rate": 0.5482, "provider_group": "provider_x", "quality": 0.7403, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.997, "id": "worker_c", "p95_latency_ms": 79022, "pass_rate": 0.8363, "provider_group": "provider_z", "quality": 0.9075, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.873, "max_latency_ms": 192562, "minimum_quality": 0.826}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000606", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.909, "estimated_latency_ms": 139981, "estimated_pass_rate": 0.7585, "estimated_quality": 0.913, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.873, "max_latency_ms": 192562, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7585, "quality": 0.913, "quality_target_met": true, "utility": 72.767692}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.873,\"max_latency_ms\":192562,\"minimum_quality\":0.826},\"task\":{\"ambiguity\":0.4664,\"category\":\"review\",\"complexity\":0.6526,\"consequence\":0.7964,\"evidence_need\":0.7866,\"family\":\"policy_compliance\",\"input_tokens\":5160,\"reversibility\":0.6012,\"risk\":0.496,\"summary\":\"Identify evidence required before approving provider failover for production.\"},\"workers\":[{\"credits_per_call\":3.03,\"id\":\"worker_a\",\"p95_latency_ms\":45037,\"pass_rate\":0.8595,\"provider_group\":\"provider_y\",\"quality\":0.8437,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.199,\"id\":\"worker_b\",\"p95_latency_ms\":29947,\"pass_rate\":0.5382,\"provider_group\":\"provider_z\",\"quality\":0.7468,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.68,\"id\":\"worker_c\",\"p95_latency_ms\":64997,\"pass_rate\":0.9802,\"provider_group\":\"provider_x\",\"quality\":0.9588,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.909,\"estimated_latency_ms\":139981,\"estimated_pass_rate\":0.7585,\"estimated_quality\":0.913,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.873,\"max_latency_ms\":192562,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4664, "category": "review", "complexity": 0.6526, "consequence": 0.7964, "evidence_need": 0.7866, "family": "policy_compliance", "input_tokens": 5160, "request_id": "synthetic-policy_compliance-000606", "reversibility": 0.6012, "risk": 0.496, "summary": "Identify evidence required before approving provider failover for production."}, "workers": [{"credits_per_call": 3.03, "id": "worker_a", "p95_latency_ms": 45037, "pass_rate": 0.8595, "provider_group": "provider_y", "quality": 0.8437, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.199, "id": "worker_b", "p95_latency_ms": 29947, "pass_rate": 0.5382, "provider_group": "provider_z", "quality": 0.7468, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.68, "id": "worker_c", "p95_latency_ms": 64997, "pass_rate": 0.9802, "provider_group": "provider_x", "quality": 0.9588, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.682, "max_latency_ms": 218968, "minimum_quality": 0.8094}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001209", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.732, "estimated_latency_ms": 176431, "estimated_pass_rate": 0.7021, "estimated_quality": 0.8769, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 25.682, "max_latency_ms": 218968, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7021, "quality": 0.8769, "quality_target_met": true, "utility": 78.203729}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.682,\"max_latency_ms\":218968,\"minimum_quality\":0.8094},\"task\":{\"ambiguity\":0.7908,\"category\":\"review\",\"complexity\":0.6708,\"consequence\":0.8907,\"evidence_need\":0.8425,\"family\":\"policy_compliance\",\"input_tokens\":2560,\"reversibility\":0.5882,\"risk\":0.7625,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":2.794,\"id\":\"worker_a\",\"p95_latency_ms\":67204,\"pass_rate\":0.7419,\"provider_group\":\"provider_x\",\"quality\":0.7987,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.214,\"id\":\"worker_b\",\"p95_latency_ms\":25538,\"pass_rate\":0.5285,\"provider_group\":\"provider_z\",\"quality\":0.6958,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.724,\"id\":\"worker_c\",\"p95_latency_ms\":83689,\"pass_rate\":0.9091,\"provider_group\":\"provider_y\",\"quality\":0.9477,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.732,\"estimated_latency_ms\":176431,\"estimated_pass_rate\":0.7021,\"estimated_quality\":0.8769,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.682,\"max_latency_ms\":218968,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7908, "category": "review", "complexity": 0.6708, "consequence": 0.8907, "evidence_need": 0.8425, "family": "policy_compliance", "input_tokens": 2560, "request_id": "synthetic-policy_compliance-001209", "reversibility": 0.5882, "risk": 0.7625, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 2.794, "id": "worker_a", "p95_latency_ms": 67204, "pass_rate": 0.7419, "provider_group": "provider_x", "quality": 0.7987, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.214, "id": "worker_b", "p95_latency_ms": 25538, "pass_rate": 0.5285, "provider_group": "provider_z", "quality": 0.6958, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.724, "id": "worker_c", "p95_latency_ms": 83689, "pass_rate": 0.9091, "provider_group": "provider_y", "quality": 0.9477, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.458, "max_latency_ms": 157047, "minimum_quality": 0.8846}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000990", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.828, "estimated_latency_ms": 116566, "estimated_pass_rate": 0.7474, "estimated_quality": 0.8612, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.458, "max_latency_ms": 157047, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7474, "quality": 0.8612, "quality_target_met": false, "utility": 73.085746}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.458,\"max_latency_ms\":157047,\"minimum_quality\":0.8846},\"task\":{\"ambiguity\":0.6341,\"category\":\"review\",\"complexity\":0.7452,\"consequence\":0.7882,\"evidence_need\":0.6059,\"family\":\"policy_compliance\",\"input_tokens\":1360,\"reversibility\":0.5143,\"risk\":0.691,\"summary\":\"Identify evidence required before approving prompt caching for production.\"},\"workers\":[{\"credits_per_call\":0.798,\"id\":\"worker_a\",\"p95_latency_ms\":19767,\"pass_rate\":0.577,\"provider_group\":\"provider_y\",\"quality\":0.6842,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.101,\"id\":\"worker_b\",\"p95_latency_ms\":47129,\"pass_rate\":0.7207,\"provider_group\":\"provider_z\",\"quality\":0.7951,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.929,\"id\":\"worker_c\",\"p95_latency_ms\":49670,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.93,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.828,\"estimated_latency_ms\":116566,\"estimated_pass_rate\":0.7474,\"estimated_quality\":0.8612,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.458,\"max_latency_ms\":157047,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6341, "category": "review", "complexity": 0.7452, "consequence": 0.7882, "evidence_need": 0.6059, "family": "policy_compliance", "input_tokens": 1360, "request_id": "synthetic-policy_compliance-000990", "reversibility": 0.5143, "risk": 0.691, "summary": "Identify evidence required before approving prompt caching for production."}, "workers": [{"credits_per_call": 0.798, "id": "worker_a", "p95_latency_ms": 19767, "pass_rate": 0.577, "provider_group": "provider_y", "quality": 0.6842, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.101, "id": "worker_b", "p95_latency_ms": 47129, "pass_rate": 0.7207, "provider_group": "provider_z", "quality": 0.7951, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.929, "id": "worker_c", "p95_latency_ms": 49670, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.93, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.462, "max_latency_ms": 242047, "minimum_quality": 0.8686}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000644", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.051, "estimated_latency_ms": 154201, "estimated_pass_rate": 0.8929, "estimated_quality": 0.9402, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.462, "max_latency_ms": 242047, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8929, "quality": 0.9402, "quality_target_met": true, "utility": 70.799556}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.462,\"max_latency_ms\":242047,\"minimum_quality\":0.8686},\"task\":{\"ambiguity\":0.4981,\"category\":\"analysis\",\"complexity\":0.6261,\"consequence\":0.3505,\"evidence_need\":0.5955,\"family\":\"performance_tuning\",\"input_tokens\":2080,\"reversibility\":0.8276,\"risk\":0.4839,\"summary\":\"Diagnose latency variance in credit reservation and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.809,\"id\":\"worker_a\",\"p95_latency_ms\":24407,\"pass_rate\":0.4949,\"provider_group\":\"provider_z\",\"quality\":0.7391,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.331,\"id\":\"worker_b\",\"p95_latency_ms\":72029,\"pass_rate\":0.7985,\"provider_group\":\"provider_x\",\"quality\":0.9008,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.72,\"id\":\"worker_c\",\"p95_latency_ms\":82172,\"pass_rate\":0.9775,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.051,\"estimated_latency_ms\":154201,\"estimated_pass_rate\":0.8929,\"estimated_quality\":0.9402,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.462,\"max_latency_ms\":242047,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4981, "category": "analysis", "complexity": 0.6261, "consequence": 0.3505, "evidence_need": 0.5955, "family": "performance_tuning", "input_tokens": 2080, "request_id": "synthetic-performance_tuning-000644", "reversibility": 0.8276, "risk": 0.4839, "summary": "Diagnose latency variance in credit reservation and rank the next experiments."}, "workers": [{"credits_per_call": 0.809, "id": "worker_a", "p95_latency_ms": 24407, "pass_rate": 0.4949, "provider_group": "provider_z", "quality": 0.7391, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.331, "id": "worker_b", "p95_latency_ms": 72029, "pass_rate": 0.7985, "provider_group": "provider_x", "quality": 0.9008, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.72, "id": "worker_c", "p95_latency_ms": 82172, "pass_rate": 0.9775, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.45, "max_latency_ms": 35029, "minimum_quality": 0.6671}, "family": "migration_planning", "id": "synthetic-migration_planning-000523", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.892, "estimated_latency_ms": 30242, "estimated_pass_rate": 0.4278, "estimated_quality": 0.6739, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.45, "max_latency_ms": 35029, "policy": "lean", "reason_codes": ["latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4278, "quality": 0.6739, "quality_target_met": true, "utility": 45.926307}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.45,\"max_latency_ms\":35029,\"minimum_quality\":0.6671},\"task\":{\"ambiguity\":0.4378,\"category\":\"coding\",\"complexity\":0.6212,\"consequence\":0.5939,\"evidence_need\":0.6184,\"family\":\"migration_planning\",\"input_tokens\":1440,\"reversibility\":0.5579,\"risk\":0.5688,\"summary\":\"Design compatibility testing for a artifact storage schema migration.\"},\"workers\":[{\"credits_per_call\":2.76,\"id\":\"worker_a\",\"p95_latency_ms\":71224,\"pass_rate\":0.8204,\"provider_group\":\"provider_y\",\"quality\":0.8802,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.116,\"id\":\"worker_b\",\"p95_latency_ms\":50221,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9589,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.892,\"id\":\"worker_c\",\"p95_latency_ms\":30242,\"pass_rate\":0.5774,\"provider_group\":\"provider_z\",\"quality\":0.7766,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.892,\"estimated_latency_ms\":30242,\"estimated_pass_rate\":0.4278,\"estimated_quality\":0.6739,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.45,\"max_latency_ms\":35029,\"policy\":\"lean\",\"reason_codes\":[\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4378, "category": "coding", "complexity": 0.6212, "consequence": 0.5939, "evidence_need": 0.6184, "family": "migration_planning", "input_tokens": 1440, "request_id": "synthetic-migration_planning-000523", "reversibility": 0.5579, "risk": 0.5688, "summary": "Design compatibility testing for a artifact storage schema migration."}, "workers": [{"credits_per_call": 2.76, "id": "worker_a", "p95_latency_ms": 71224, "pass_rate": 0.8204, "provider_group": "provider_y", "quality": 0.8802, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.116, "id": "worker_b", "p95_latency_ms": 50221, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9589, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.892, "id": "worker_c", "p95_latency_ms": 30242, "pass_rate": 0.5774, "provider_group": "provider_z", "quality": 0.7766, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.963, "max_latency_ms": 123282, "minimum_quality": 0.7912}, "family": "migration_planning", "id": "synthetic-migration_planning-000940", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.666, "estimated_latency_ms": 102747, "estimated_pass_rate": 0.6607, "estimated_quality": 0.8313, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.963, "max_latency_ms": 123282, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6607, "quality": 0.8313, "quality_target_met": true, "utility": 61.94817}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.963,\"max_latency_ms\":123282,\"minimum_quality\":0.7912},\"task\":{\"ambiguity\":0.4373,\"category\":\"coding\",\"complexity\":0.9278,\"consequence\":0.4455,\"evidence_need\":0.7346,\"family\":\"migration_planning\",\"input_tokens\":2120,\"reversibility\":0.4406,\"risk\":0.731,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.45,\"id\":\"worker_a\",\"p95_latency_ms\":70646,\"pass_rate\":0.8604,\"provider_group\":\"provider_x\",\"quality\":0.9043,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.216,\"id\":\"worker_b\",\"p95_latency_ms\":32101,\"pass_rate\":0.5535,\"provider_group\":\"provider_z\",\"quality\":0.7737,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.946,\"id\":\"worker_c\",\"p95_latency_ms\":76408,\"pass_rate\":0.924,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.666,\"estimated_latency_ms\":102747,\"estimated_pass_rate\":0.6607,\"estimated_quality\":0.8313,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.963,\"max_latency_ms\":123282,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4373, "category": "coding", "complexity": 0.9278, "consequence": 0.4455, "evidence_need": 0.7346, "family": "migration_planning", "input_tokens": 2120, "request_id": "synthetic-migration_planning-000940", "reversibility": 0.4406, "risk": 0.731, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 3.45, "id": "worker_a", "p95_latency_ms": 70646, "pass_rate": 0.8604, "provider_group": "provider_x", "quality": 0.9043, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.216, "id": "worker_b", "p95_latency_ms": 32101, "pass_rate": 0.5535, "provider_group": "provider_z", "quality": 0.7737, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.946, "id": "worker_c", "p95_latency_ms": 76408, "pass_rate": 0.924, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.645, "max_latency_ms": 42550, "minimum_quality": 0.647}, "family": "migration_planning", "id": "synthetic-migration_planning-000796", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.912, "estimated_latency_ms": 25860, "estimated_pass_rate": 0.3131, "estimated_quality": 0.6477, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.645, "max_latency_ms": 42550, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3131, "quality": 0.6477, "quality_target_met": true, "utility": 42.87236}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.645,\"max_latency_ms\":42550,\"minimum_quality\":0.647},\"task\":{\"ambiguity\":0.5086,\"category\":\"coding\",\"complexity\":0.6365,\"consequence\":0.4006,\"evidence_need\":0.6609,\"family\":\"migration_planning\",\"input_tokens\":6320,\"reversibility\":0.4984,\"risk\":0.6222,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.829,\"id\":\"worker_a\",\"p95_latency_ms\":72813,\"pass_rate\":0.7419,\"provider_group\":\"provider_z\",\"quality\":0.8685,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.912,\"id\":\"worker_b\",\"p95_latency_ms\":25860,\"pass_rate\":0.4741,\"provider_group\":\"provider_x\",\"quality\":0.7618,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.531,\"id\":\"worker_c\",\"p95_latency_ms\":83740,\"pass_rate\":0.9337,\"provider_group\":\"provider_y\",\"quality\":0.9233,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.912,\"estimated_latency_ms\":25860,\"estimated_pass_rate\":0.3131,\"estimated_quality\":0.6477,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.645,\"max_latency_ms\":42550,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5086, "category": "coding", "complexity": 0.6365, "consequence": 0.4006, "evidence_need": 0.6609, "family": "migration_planning", "input_tokens": 6320, "request_id": "synthetic-migration_planning-000796", "reversibility": 0.4984, "risk": 0.6222, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 2.829, "id": "worker_a", "p95_latency_ms": 72813, "pass_rate": 0.7419, "provider_group": "provider_z", "quality": 0.8685, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.912, "id": "worker_b", "p95_latency_ms": 25860, "pass_rate": 0.4741, "provider_group": "provider_x", "quality": 0.7618, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.531, "id": "worker_c", "p95_latency_ms": 83740, "pass_rate": 0.9337, "provider_group": "provider_y", "quality": 0.9233, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.506, "max_latency_ms": 227240, "minimum_quality": 0.8922}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001143", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.679, "estimated_latency_ms": 167012, "estimated_pass_rate": 0.7534, "estimated_quality": 0.9061, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.506, "max_latency_ms": 227240, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7534, "quality": 0.9061, "quality_target_met": true, "utility": 77.035087}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.506,\"max_latency_ms\":227240,\"minimum_quality\":0.8922},\"task\":{\"ambiguity\":0.6802,\"category\":\"review\",\"complexity\":0.8394,\"consequence\":0.8605,\"evidence_need\":0.8559,\"family\":\"policy_compliance\",\"input_tokens\":1440,\"reversibility\":0.485,\"risk\":0.4928,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.704,\"id\":\"worker_a\",\"p95_latency_ms\":47323,\"pass_rate\":0.8694,\"provider_group\":\"provider_x\",\"quality\":0.8851,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.191,\"id\":\"worker_b\",\"p95_latency_ms\":28903,\"pass_rate\":0.5226,\"provider_group\":\"provider_z\",\"quality\":0.7575,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.784,\"id\":\"worker_c\",\"p95_latency_ms\":90786,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9387,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.679,\"estimated_latency_ms\":167012,\"estimated_pass_rate\":0.7534,\"estimated_quality\":0.9061,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.506,\"max_latency_ms\":227240,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6802, "category": "review", "complexity": 0.8394, "consequence": 0.8605, "evidence_need": 0.8559, "family": "policy_compliance", "input_tokens": 1440, "request_id": "synthetic-policy_compliance-001143", "reversibility": 0.485, "risk": 0.4928, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.704, "id": "worker_a", "p95_latency_ms": 47323, "pass_rate": 0.8694, "provider_group": "provider_x", "quality": 0.8851, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.191, "id": "worker_b", "p95_latency_ms": 28903, "pass_rate": 0.5226, "provider_group": "provider_z", "quality": 0.7575, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.784, "id": "worker_c", "p95_latency_ms": 90786, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9387, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.507, "max_latency_ms": 51801, "minimum_quality": 0.5495}, "family": "migration_planning", "id": "synthetic-migration_planning-000706", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.133, "estimated_latency_ms": 26427, "estimated_pass_rate": 0.3563, "estimated_quality": 0.6989, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.507, "max_latency_ms": 51801, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3563, "quality": 0.6989, "quality_target_met": true, "utility": 46.06742}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.507,\"max_latency_ms\":51801,\"minimum_quality\":0.5495},\"task\":{\"ambiguity\":0.4892,\"category\":\"coding\",\"complexity\":0.6518,\"consequence\":0.4421,\"evidence_need\":0.5606,\"family\":\"migration_planning\",\"input_tokens\":5680,\"reversibility\":0.4008,\"risk\":0.4819,\"summary\":\"Plan a zero-downtime migration of prompt caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.407,\"id\":\"worker_a\",\"p95_latency_ms\":68403,\"pass_rate\":0.8704,\"provider_group\":\"provider_y\",\"quality\":0.8871,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.133,\"id\":\"worker_b\",\"p95_latency_ms\":26427,\"pass_rate\":0.5062,\"provider_group\":\"provider_x\",\"quality\":0.8043,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.062,\"id\":\"worker_c\",\"p95_latency_ms\":57925,\"pass_rate\":0.9578,\"provider_group\":\"provider_z\",\"quality\":0.9156,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.133,\"estimated_latency_ms\":26427,\"estimated_pass_rate\":0.3563,\"estimated_quality\":0.6989,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.507,\"max_latency_ms\":51801,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4892, "category": "coding", "complexity": 0.6518, "consequence": 0.4421, "evidence_need": 0.5606, "family": "migration_planning", "input_tokens": 5680, "request_id": "synthetic-migration_planning-000706", "reversibility": 0.4008, "risk": 0.4819, "summary": "Plan a zero-downtime migration of prompt caching with rollback checkpoints."}, "workers": [{"credits_per_call": 3.407, "id": "worker_a", "p95_latency_ms": 68403, "pass_rate": 0.8704, "provider_group": "provider_y", "quality": 0.8871, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.133, "id": "worker_b", "p95_latency_ms": 26427, "pass_rate": 0.5062, "provider_group": "provider_x", "quality": 0.8043, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.062, "id": "worker_c", "p95_latency_ms": 57925, "pass_rate": 0.9578, "provider_group": "provider_z", "quality": 0.9156, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.171, "max_latency_ms": 220632, "minimum_quality": 0.9487}, "family": "migration_planning", "id": "synthetic-migration_planning-000496", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.732, "estimated_latency_ms": 158093, "estimated_pass_rate": 0.7264, "estimated_quality": 0.9158, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.171, "max_latency_ms": 220632, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7264, "quality": 0.9158, "quality_target_met": false, "utility": 71.570949}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.171,\"max_latency_ms\":220632,\"minimum_quality\":0.9487},\"task\":{\"ambiguity\":0.3477,\"category\":\"coding\",\"complexity\":0.8192,\"consequence\":0.4417,\"evidence_need\":0.623,\"family\":\"migration_planning\",\"input_tokens\":6440,\"reversibility\":0.315,\"risk\":0.7137,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":3.077,\"id\":\"worker_a\",\"p95_latency_ms\":48162,\"pass_rate\":0.7526,\"provider_group\":\"provider_x\",\"quality\":0.8452,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.77,\"id\":\"worker_b\",\"p95_latency_ms\":86574,\"pass_rate\":0.9171,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.885,\"id\":\"worker_c\",\"p95_latency_ms\":23357,\"pass_rate\":0.545,\"provider_group\":\"provider_y\",\"quality\":0.7912,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.732,\"estimated_latency_ms\":158093,\"estimated_pass_rate\":0.7264,\"estimated_quality\":0.9158,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.171,\"max_latency_ms\":220632,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3477, "category": "coding", "complexity": 0.8192, "consequence": 0.4417, "evidence_need": 0.623, "family": "migration_planning", "input_tokens": 6440, "request_id": "synthetic-migration_planning-000496", "reversibility": 0.315, "risk": 0.7137, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 3.077, "id": "worker_a", "p95_latency_ms": 48162, "pass_rate": 0.7526, "provider_group": "provider_x", "quality": 0.8452, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.77, "id": "worker_b", "p95_latency_ms": 86574, "pass_rate": 0.9171, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.885, "id": "worker_c", "p95_latency_ms": 23357, "pass_rate": 0.545, "provider_group": "provider_y", "quality": 0.7912, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.831, "max_latency_ms": 227888, "minimum_quality": 0.8463}, "family": "migration_planning", "id": "synthetic-migration_planning-000610", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.313, "estimated_latency_ms": 185306, "estimated_pass_rate": 0.7363, "estimated_quality": 0.9171, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.831, "max_latency_ms": 227888, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7363, "quality": 0.9171, "quality_target_met": true, "utility": 73.113093}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.831,\"max_latency_ms\":227888,\"minimum_quality\":0.8463},\"task\":{\"ambiguity\":0.6072,\"category\":\"coding\",\"complexity\":0.7453,\"consequence\":0.5216,\"evidence_need\":0.7184,\"family\":\"migration_planning\",\"input_tokens\":5000,\"reversibility\":0.2687,\"risk\":0.5022,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":10.53,\"id\":\"worker_a\",\"p95_latency_ms\":89632,\"pass_rate\":0.9934,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.001,\"id\":\"worker_b\",\"p95_latency_ms\":66067,\"pass_rate\":0.7473,\"provider_group\":\"provider_z\",\"quality\":0.8262,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.782,\"id\":\"worker_c\",\"p95_latency_ms\":29607,\"pass_rate\":0.546,\"provider_group\":\"provider_x\",\"quality\":0.7573,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.313,\"estimated_latency_ms\":185306,\"estimated_pass_rate\":0.7363,\"estimated_quality\":0.9171,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.831,\"max_latency_ms\":227888,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6072, "category": "coding", "complexity": 0.7453, "consequence": 0.5216, "evidence_need": 0.7184, "family": "migration_planning", "input_tokens": 5000, "request_id": "synthetic-migration_planning-000610", "reversibility": 0.2687, "risk": 0.5022, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 10.53, "id": "worker_a", "p95_latency_ms": 89632, "pass_rate": 0.9934, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.001, "id": "worker_b", "p95_latency_ms": 66067, "pass_rate": 0.7473, "provider_group": "provider_z", "quality": 0.8262, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.782, "id": "worker_c", "p95_latency_ms": 29607, "pass_rate": 0.546, "provider_group": "provider_x", "quality": 0.7573, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 13.004, "max_latency_ms": 128853, "minimum_quality": 0.7432}, "family": "migration_planning", "id": "synthetic-migration_planning-000319", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.754, "estimated_latency_ms": 93373, "estimated_pass_rate": 0.8274, "estimated_quality": 0.8665, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.004, "max_latency_ms": 128853, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8274, "quality": 0.8665, "quality_target_met": true, "utility": 63.236001}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":13.004,\"max_latency_ms\":128853,\"minimum_quality\":0.7432},\"task\":{\"ambiguity\":0.5237,\"category\":\"coding\",\"complexity\":0.7961,\"consequence\":0.811,\"evidence_need\":0.7099,\"family\":\"migration_planning\",\"input_tokens\":7160,\"reversibility\":0.3285,\"risk\":0.4811,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":4.182,\"id\":\"worker_a\",\"p95_latency_ms\":49160,\"pass_rate\":0.8764,\"provider_group\":\"provider_y\",\"quality\":0.8013,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.869,\"id\":\"worker_b\",\"p95_latency_ms\":62310,\"pass_rate\":0.9433,\"provider_group\":\"provider_x\",\"quality\":0.9471,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.885,\"id\":\"worker_c\",\"p95_latency_ms\":31063,\"pass_rate\":0.5995,\"provider_group\":\"provider_z\",\"quality\":0.801,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.754,\"estimated_latency_ms\":93373,\"estimated_pass_rate\":0.8274,\"estimated_quality\":0.8665,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.004,\"max_latency_ms\":128853,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5237, "category": "coding", "complexity": 0.7961, "consequence": 0.811, "evidence_need": 0.7099, "family": "migration_planning", "input_tokens": 7160, "request_id": "synthetic-migration_planning-000319", "reversibility": 0.3285, "risk": 0.4811, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 4.182, "id": "worker_a", "p95_latency_ms": 49160, "pass_rate": 0.8764, "provider_group": "provider_y", "quality": 0.8013, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.869, "id": "worker_b", "p95_latency_ms": 62310, "pass_rate": 0.9433, "provider_group": "provider_x", "quality": 0.9471, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.885, "id": "worker_c", "p95_latency_ms": 31063, "pass_rate": 0.5995, "provider_group": "provider_z", "quality": 0.801, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.488, "max_latency_ms": 205042, "minimum_quality": 0.792}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000153", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.065, "estimated_latency_ms": 155156, "estimated_pass_rate": 0.7192, "estimated_quality": 0.9228, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.488, "max_latency_ms": 205042, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7192, "quality": 0.9228, "quality_target_met": true, "utility": 74.532165}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.488,\"max_latency_ms\":205042,\"minimum_quality\":0.792},\"task\":{\"ambiguity\":0.4118,\"category\":\"review\",\"complexity\":0.7515,\"consequence\":0.6921,\"evidence_need\":0.7518,\"family\":\"policy_compliance\",\"input_tokens\":1440,\"reversibility\":0.348,\"risk\":0.6221,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.556,\"id\":\"worker_a\",\"p95_latency_ms\":64196,\"pass_rate\":0.8998,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.029,\"id\":\"worker_b\",\"p95_latency_ms\":24002,\"pass_rate\":0.5162,\"provider_group\":\"provider_x\",\"quality\":0.7666,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.48,\"id\":\"worker_c\",\"p95_latency_ms\":66958,\"pass_rate\":0.7666,\"provider_group\":\"provider_z\",\"quality\":0.9105,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.065,\"estimated_latency_ms\":155156,\"estimated_pass_rate\":0.7192,\"estimated_quality\":0.9228,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.488,\"max_latency_ms\":205042,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4118, "category": "review", "complexity": 0.7515, "consequence": 0.6921, "evidence_need": 0.7518, "family": "policy_compliance", "input_tokens": 1440, "request_id": "synthetic-policy_compliance-000153", "reversibility": 0.348, "risk": 0.6221, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.556, "id": "worker_a", "p95_latency_ms": 64196, "pass_rate": 0.8998, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.029, "id": "worker_b", "p95_latency_ms": 24002, "pass_rate": 0.5162, "provider_group": "provider_x", "quality": 0.7666, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.48, "id": "worker_c", "p95_latency_ms": 66958, "pass_rate": 0.7666, "provider_group": "provider_z", "quality": 0.9105, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.242, "max_latency_ms": 38293, "minimum_quality": 0.6104}, "family": "migration_planning", "id": "synthetic-migration_planning-000661", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.074, "estimated_latency_ms": 29623, "estimated_pass_rate": 0.3681, "estimated_quality": 0.6103, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.242, "max_latency_ms": 38293, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3681, "quality": 0.6103, "quality_target_met": false, "utility": 36.766415}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.242,\"max_latency_ms\":38293,\"minimum_quality\":0.6104},\"task\":{\"ambiguity\":0.3715,\"category\":\"coding\",\"complexity\":0.9129,\"consequence\":0.4644,\"evidence_need\":0.5729,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.5601,\"risk\":0.44,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.954,\"id\":\"worker_a\",\"p95_latency_ms\":61955,\"pass_rate\":0.843,\"provider_group\":\"provider_x\",\"quality\":0.8079,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.074,\"id\":\"worker_b\",\"p95_latency_ms\":29623,\"pass_rate\":0.5353,\"provider_group\":\"provider_z\",\"quality\":0.752,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.761,\"id\":\"worker_c\",\"p95_latency_ms\":70104,\"pass_rate\":0.9077,\"provider_group\":\"provider_y\",\"quality\":0.9148,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.074,\"estimated_latency_ms\":29623,\"estimated_pass_rate\":0.3681,\"estimated_quality\":0.6103,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.242,\"max_latency_ms\":38293,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3715, "category": "coding", "complexity": 0.9129, "consequence": 0.4644, "evidence_need": 0.5729, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-000661", "reversibility": 0.5601, "risk": 0.44, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 2.954, "id": "worker_a", "p95_latency_ms": 61955, "pass_rate": 0.843, "provider_group": "provider_x", "quality": 0.8079, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.074, "id": "worker_b", "p95_latency_ms": 29623, "pass_rate": 0.5353, "provider_group": "provider_z", "quality": 0.752, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.761, "id": "worker_c", "p95_latency_ms": 70104, "pass_rate": 0.9077, "provider_group": "provider_y", "quality": 0.9148, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.052, "max_latency_ms": 164617, "minimum_quality": 0.9103}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000672", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.198, "estimated_latency_ms": 160652, "estimated_pass_rate": 0.7667, "estimated_quality": 0.9266, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 26.052, "max_latency_ms": 164617, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7667, "quality": 0.9266, "quality_target_met": true, "utility": 79.777576}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.052,\"max_latency_ms\":164617,\"minimum_quality\":0.9103},\"task\":{\"ambiguity\":0.796,\"category\":\"review\",\"complexity\":0.6333,\"consequence\":0.6484,\"evidence_need\":0.8608,\"family\":\"policy_compliance\",\"input_tokens\":4840,\"reversibility\":0.387,\"risk\":0.7749,\"summary\":\"Identify evidence required before approving private credentials for production.\"},\"workers\":[{\"credits_per_call\":3.583,\"id\":\"worker_a\",\"p95_latency_ms\":78431,\"pass_rate\":0.7276,\"provider_group\":\"provider_z\",\"quality\":0.9225,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.704,\"id\":\"worker_b\",\"p95_latency_ms\":53052,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.911,\"id\":\"worker_c\",\"p95_latency_ms\":29169,\"pass_rate\":0.5538,\"provider_group\":\"provider_y\",\"quality\":0.7675,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.198,\"estimated_latency_ms\":160652,\"estimated_pass_rate\":0.7667,\"estimated_quality\":0.9266,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.052,\"max_latency_ms\":164617,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.796, "category": "review", "complexity": 0.6333, "consequence": 0.6484, "evidence_need": 0.8608, "family": "policy_compliance", "input_tokens": 4840, "request_id": "synthetic-policy_compliance-000672", "reversibility": 0.387, "risk": 0.7749, "summary": "Identify evidence required before approving private credentials for production."}, "workers": [{"credits_per_call": 3.583, "id": "worker_a", "p95_latency_ms": 78431, "pass_rate": 0.7276, "provider_group": "provider_z", "quality": 0.9225, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.704, "id": "worker_b", "p95_latency_ms": 53052, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.911, "id": "worker_c", "p95_latency_ms": 29169, "pass_rate": 0.5538, "provider_group": "provider_y", "quality": 0.7675, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.332, "max_latency_ms": 208227, "minimum_quality": 0.931}, "family": "migration_planning", "id": "synthetic-migration_planning-000133", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.046, "estimated_latency_ms": 178244, "estimated_pass_rate": 0.7675, "estimated_quality": 0.8782, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 22.332, "max_latency_ms": 208227, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7675, "quality": 0.8782, "quality_target_met": false, "utility": 73.591815}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.332,\"max_latency_ms\":208227,\"minimum_quality\":0.931},\"task\":{\"ambiguity\":0.666,\"category\":\"coding\",\"complexity\":0.9048,\"consequence\":0.7765,\"evidence_need\":0.407,\"family\":\"migration_planning\",\"input_tokens\":3960,\"reversibility\":0.2853,\"risk\":0.7332,\"summary\":\"Plan a zero-downtime migration of prompt caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.225,\"id\":\"worker_a\",\"p95_latency_ms\":81290,\"pass_rate\":0.8372,\"provider_group\":\"provider_x\",\"quality\":0.8172,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.604,\"id\":\"worker_b\",\"p95_latency_ms\":66004,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9197,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.217,\"id\":\"worker_c\",\"p95_latency_ms\":30950,\"pass_rate\":0.5757,\"provider_group\":\"provider_z\",\"quality\":0.8027,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.046,\"estimated_latency_ms\":178244,\"estimated_pass_rate\":0.7675,\"estimated_quality\":0.8782,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.332,\"max_latency_ms\":208227,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.666, "category": "coding", "complexity": 0.9048, "consequence": 0.7765, "evidence_need": 0.407, "family": "migration_planning", "input_tokens": 3960, "request_id": "synthetic-migration_planning-000133", "reversibility": 0.2853, "risk": 0.7332, "summary": "Plan a zero-downtime migration of prompt caching with rollback checkpoints."}, "workers": [{"credits_per_call": 4.225, "id": "worker_a", "p95_latency_ms": 81290, "pass_rate": 0.8372, "provider_group": "provider_x", "quality": 0.8172, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.604, "id": "worker_b", "p95_latency_ms": 66004, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9197, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.217, "id": "worker_c", "p95_latency_ms": 30950, "pass_rate": 0.5757, "provider_group": "provider_z", "quality": 0.8027, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.054, "max_latency_ms": 151471, "minimum_quality": 0.849}, "family": "migration_planning", "id": "synthetic-migration_planning-001255", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.087, "estimated_latency_ms": 127625, "estimated_pass_rate": 0.759, "estimated_quality": 0.8968, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.054, "max_latency_ms": 151471, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.759, "quality": 0.8968, "quality_target_met": true, "utility": 73.755917}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.054,\"max_latency_ms\":151471,\"minimum_quality\":0.849},\"task\":{\"ambiguity\":0.5646,\"category\":\"coding\",\"complexity\":0.7881,\"consequence\":0.6454,\"evidence_need\":0.5141,\"family\":\"migration_planning\",\"input_tokens\":6440,\"reversibility\":0.5717,\"risk\":0.6444,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":9.955,\"id\":\"worker_a\",\"p95_latency_ms\":55720,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9638,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.147,\"id\":\"worker_b\",\"p95_latency_ms\":22212,\"pass_rate\":0.5477,\"provider_group\":\"provider_y\",\"quality\":0.7435,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.985,\"id\":\"worker_c\",\"p95_latency_ms\":49693,\"pass_rate\":0.8194,\"provider_group\":\"provider_x\",\"quality\":0.8331,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.087,\"estimated_latency_ms\":127625,\"estimated_pass_rate\":0.759,\"estimated_quality\":0.8968,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.054,\"max_latency_ms\":151471,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5646, "category": "coding", "complexity": 0.7881, "consequence": 0.6454, "evidence_need": 0.5141, "family": "migration_planning", "input_tokens": 6440, "request_id": "synthetic-migration_planning-001255", "reversibility": 0.5717, "risk": 0.6444, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 9.955, "id": "worker_a", "p95_latency_ms": 55720, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9638, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.147, "id": "worker_b", "p95_latency_ms": 22212, "pass_rate": 0.5477, "provider_group": "provider_y", "quality": 0.7435, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.985, "id": "worker_c", "p95_latency_ms": 49693, "pass_rate": 0.8194, "provider_group": "provider_x", "quality": 0.8331, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.373, "max_latency_ms": 236325, "minimum_quality": 0.8795}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000933", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.731, "estimated_latency_ms": 91547, "estimated_pass_rate": 0.8294, "estimated_quality": 0.922, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.373, "max_latency_ms": 236325, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8294, "quality": 0.922, "quality_target_met": true, "utility": 68.958438}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.373,\"max_latency_ms\":236325,\"minimum_quality\":0.8795},\"task\":{\"ambiguity\":0.6979,\"category\":\"review\",\"complexity\":0.4885,\"consequence\":0.9055,\"evidence_need\":0.9205,\"family\":\"policy_compliance\",\"input_tokens\":2760,\"reversibility\":0.3113,\"risk\":0.5117,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.774,\"id\":\"worker_a\",\"p95_latency_ms\":67921,\"pass_rate\":0.9355,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.256,\"id\":\"worker_b\",\"p95_latency_ms\":77058,\"pass_rate\":0.8649,\"provider_group\":\"provider_x\",\"quality\":0.8444,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.957,\"id\":\"worker_c\",\"p95_latency_ms\":23626,\"pass_rate\":0.5682,\"provider_group\":\"provider_y\",\"quality\":0.7561,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.731,\"estimated_latency_ms\":91547,\"estimated_pass_rate\":0.8294,\"estimated_quality\":0.922,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.373,\"max_latency_ms\":236325,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6979, "category": "review", "complexity": 0.4885, "consequence": 0.9055, "evidence_need": 0.9205, "family": "policy_compliance", "input_tokens": 2760, "request_id": "synthetic-policy_compliance-000933", "reversibility": 0.3113, "risk": 0.5117, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 10.774, "id": "worker_a", "p95_latency_ms": 67921, "pass_rate": 0.9355, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.256, "id": "worker_b", "p95_latency_ms": 77058, "pass_rate": 0.8649, "provider_group": "provider_x", "quality": 0.8444, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.957, "id": "worker_c", "p95_latency_ms": 23626, "pass_rate": 0.5682, "provider_group": "provider_y", "quality": 0.7561, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.894, "max_latency_ms": 148465, "minimum_quality": 0.833}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001245", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.821, "estimated_latency_ms": 92797, "estimated_pass_rate": 0.5507, "estimated_quality": 0.8075, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.894, "max_latency_ms": 148465, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5507, "quality": 0.8075, "quality_target_met": false, "utility": 58.745812}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.894,\"max_latency_ms\":148465,\"minimum_quality\":0.833},\"task\":{\"ambiguity\":0.7509,\"category\":\"review\",\"complexity\":0.503,\"consequence\":0.7145,\"evidence_need\":0.8543,\"family\":\"policy_compliance\",\"input_tokens\":6760,\"reversibility\":0.661,\"risk\":0.5499,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.803,\"id\":\"worker_a\",\"p95_latency_ms\":27863,\"pass_rate\":0.4762,\"provider_group\":\"provider_z\",\"quality\":0.7453,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.018,\"id\":\"worker_b\",\"p95_latency_ms\":64934,\"pass_rate\":0.7659,\"provider_group\":\"provider_y\",\"quality\":0.818,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.037,\"id\":\"worker_c\",\"p95_latency_ms\":91626,\"pass_rate\":0.9051,\"provider_group\":\"provider_x\",\"quality\":0.9909,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.821,\"estimated_latency_ms\":92797,\"estimated_pass_rate\":0.5507,\"estimated_quality\":0.8075,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.894,\"max_latency_ms\":148465,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7509, "category": "review", "complexity": 0.503, "consequence": 0.7145, "evidence_need": 0.8543, "family": "policy_compliance", "input_tokens": 6760, "request_id": "synthetic-policy_compliance-001245", "reversibility": 0.661, "risk": 0.5499, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.803, "id": "worker_a", "p95_latency_ms": 27863, "pass_rate": 0.4762, "provider_group": "provider_z", "quality": 0.7453, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.018, "id": "worker_b", "p95_latency_ms": 64934, "pass_rate": 0.7659, "provider_group": "provider_y", "quality": 0.818, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.037, "id": "worker_c", "p95_latency_ms": 91626, "pass_rate": 0.9051, "provider_group": "provider_x", "quality": 0.9909, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.369, "max_latency_ms": 36626, "minimum_quality": 0.6067}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000215", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.095, "estimated_latency_ms": 28472, "estimated_pass_rate": 0.4005, "estimated_quality": 0.5489, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.369, "max_latency_ms": 36626, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4005, "quality": 0.5489, "quality_target_met": false, "utility": 35.042373}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.369,\"max_latency_ms\":36626,\"minimum_quality\":0.6067},\"task\":{\"ambiguity\":0.6587,\"category\":\"analysis\",\"complexity\":0.5993,\"consequence\":0.4142,\"evidence_need\":0.5826,\"family\":\"performance_tuning\",\"input_tokens\":1520,\"reversibility\":0.6133,\"risk\":0.3815,\"summary\":\"Review a proposed optimization for data retention for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.095,\"id\":\"worker_a\",\"p95_latency_ms\":28472,\"pass_rate\":0.5492,\"provider_group\":\"provider_z\",\"quality\":0.649,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":12.754,\"id\":\"worker_b\",\"p95_latency_ms\":91242,\"pass_rate\":0.9893,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.155,\"id\":\"worker_c\",\"p95_latency_ms\":81193,\"pass_rate\":0.9209,\"provider_group\":\"provider_x\",\"quality\":0.8729,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.095,\"estimated_latency_ms\":28472,\"estimated_pass_rate\":0.4005,\"estimated_quality\":0.5489,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.369,\"max_latency_ms\":36626,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6587, "category": "analysis", "complexity": 0.5993, "consequence": 0.4142, "evidence_need": 0.5826, "family": "performance_tuning", "input_tokens": 1520, "request_id": "synthetic-performance_tuning-000215", "reversibility": 0.6133, "risk": 0.3815, "summary": "Review a proposed optimization for data retention for hidden regressions."}, "workers": [{"credits_per_call": 1.095, "id": "worker_a", "p95_latency_ms": 28472, "pass_rate": 0.5492, "provider_group": "provider_z", "quality": 0.649, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 12.754, "id": "worker_b", "p95_latency_ms": 91242, "pass_rate": 0.9893, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.155, "id": "worker_c", "p95_latency_ms": 81193, "pass_rate": 0.9209, "provider_group": "provider_x", "quality": 0.8729, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.605, "max_latency_ms": 85431, "minimum_quality": 0.7266}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000897", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.772, "estimated_latency_ms": 78335, "estimated_pass_rate": 0.7397, "estimated_quality": 0.8934, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.605, "max_latency_ms": 85431, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7397, "quality": 0.8934, "quality_target_met": true, "utility": 67.893792}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.605,\"max_latency_ms\":85431,\"minimum_quality\":0.7266},\"task\":{\"ambiguity\":0.7531,\"category\":\"review\",\"complexity\":0.7026,\"consequence\":0.8925,\"evidence_need\":0.6394,\"family\":\"policy_compliance\",\"input_tokens\":3920,\"reversibility\":0.6181,\"risk\":0.7245,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":11.942,\"id\":\"worker_a\",\"p95_latency_ms\":57814,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.83,\"id\":\"worker_b\",\"p95_latency_ms\":20521,\"pass_rate\":0.544,\"provider_group\":\"provider_x\",\"quality\":0.7989,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.174,\"id\":\"worker_c\",\"p95_latency_ms\":63687,\"pass_rate\":0.8568,\"provider_group\":\"provider_y\",\"quality\":0.8714,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.772,\"estimated_latency_ms\":78335,\"estimated_pass_rate\":0.7397,\"estimated_quality\":0.8934,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.605,\"max_latency_ms\":85431,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7531, "category": "review", "complexity": 0.7026, "consequence": 0.8925, "evidence_need": 0.6394, "family": "policy_compliance", "input_tokens": 3920, "request_id": "synthetic-policy_compliance-000897", "reversibility": 0.6181, "risk": 0.7245, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 11.942, "id": "worker_a", "p95_latency_ms": 57814, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.83, "id": "worker_b", "p95_latency_ms": 20521, "pass_rate": 0.544, "provider_group": "provider_x", "quality": 0.7989, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.174, "id": "worker_c", "p95_latency_ms": 63687, "pass_rate": 0.8568, "provider_group": "provider_y", "quality": 0.8714, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.219, "max_latency_ms": 154468, "minimum_quality": 0.7208}, "family": "migration_planning", "id": "synthetic-migration_planning-000805", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.387, "estimated_latency_ms": 85439, "estimated_pass_rate": 0.8625, "estimated_quality": 0.8924, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.219, "max_latency_ms": 154468, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8625, "quality": 0.8924, "quality_target_met": true, "utility": 65.879516}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.219,\"max_latency_ms\":154468,\"minimum_quality\":0.7208},\"task\":{\"ambiguity\":0.4302,\"category\":\"coding\",\"complexity\":0.6956,\"consequence\":0.7224,\"evidence_need\":0.5436,\"family\":\"migration_planning\",\"input_tokens\":3120,\"reversibility\":0.4064,\"risk\":0.6804,\"summary\":\"Review a staged migration for credit reservation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.218,\"id\":\"worker_a\",\"p95_latency_ms\":52996,\"pass_rate\":0.8685,\"provider_group\":\"provider_y\",\"quality\":0.8135,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.586,\"id\":\"worker_b\",\"p95_latency_ms\":53208,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.801,\"id\":\"worker_c\",\"p95_latency_ms\":32231,\"pass_rate\":0.5281,\"provider_group\":\"provider_x\",\"quality\":0.7425,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.387,\"estimated_latency_ms\":85439,\"estimated_pass_rate\":0.8625,\"estimated_quality\":0.8924,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.219,\"max_latency_ms\":154468,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4302, "category": "coding", "complexity": 0.6956, "consequence": 0.7224, "evidence_need": 0.5436, "family": "migration_planning", "input_tokens": 3120, "request_id": "synthetic-migration_planning-000805", "reversibility": 0.4064, "risk": 0.6804, "summary": "Review a staged migration for credit reservation and identify irreversible steps."}, "workers": [{"credits_per_call": 3.218, "id": "worker_a", "p95_latency_ms": 52996, "pass_rate": 0.8685, "provider_group": "provider_y", "quality": 0.8135, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 13.586, "id": "worker_b", "p95_latency_ms": 53208, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.801, "id": "worker_c", "p95_latency_ms": 32231, "pass_rate": 0.5281, "provider_group": "provider_x", "quality": 0.7425, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.145, "max_latency_ms": 92005, "minimum_quality": 0.7858}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000399", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.01, "estimated_latency_ms": 84598, "estimated_pass_rate": 0.7009, "estimated_quality": 0.9013, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.145, "max_latency_ms": 92005, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7009, "quality": 0.9013, "quality_target_met": true, "utility": 65.705614}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.145,\"max_latency_ms\":92005,\"minimum_quality\":0.7858},\"task\":{\"ambiguity\":0.5653,\"category\":\"review\",\"complexity\":0.8085,\"consequence\":0.8883,\"evidence_need\":0.8894,\"family\":\"policy_compliance\",\"input_tokens\":7120,\"reversibility\":0.3877,\"risk\":0.5287,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.002,\"id\":\"worker_a\",\"p95_latency_ms\":59910,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9646,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.008,\"id\":\"worker_b\",\"p95_latency_ms\":24688,\"pass_rate\":0.5566,\"provider_group\":\"provider_z\",\"quality\":0.7789,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.782,\"id\":\"worker_c\",\"p95_latency_ms\":53219,\"pass_rate\":0.8781,\"provider_group\":\"provider_x\",\"quality\":0.8541,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.01,\"estimated_latency_ms\":84598,\"estimated_pass_rate\":0.7009,\"estimated_quality\":0.9013,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.145,\"max_latency_ms\":92005,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5653, "category": "review", "complexity": 0.8085, "consequence": 0.8883, "evidence_need": 0.8894, "family": "policy_compliance", "input_tokens": 7120, "request_id": "synthetic-policy_compliance-000399", "reversibility": 0.3877, "risk": 0.5287, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 12.002, "id": "worker_a", "p95_latency_ms": 59910, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9646, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.008, "id": "worker_b", "p95_latency_ms": 24688, "pass_rate": 0.5566, "provider_group": "provider_z", "quality": 0.7789, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.782, "id": "worker_c", "p95_latency_ms": 53219, "pass_rate": 0.8781, "provider_group": "provider_x", "quality": 0.8541, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.029, "max_latency_ms": 42236, "minimum_quality": 0.6009}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000272", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.84, "estimated_latency_ms": 32472, "estimated_pass_rate": 0.2695, "estimated_quality": 0.5144, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.029, "max_latency_ms": 42236, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2695, "quality": 0.5144, "quality_target_met": false, "utility": 29.057363}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.029,\"max_latency_ms\":42236,\"minimum_quality\":0.6009},\"task\":{\"ambiguity\":0.563,\"category\":\"analysis\",\"complexity\":0.8092,\"consequence\":0.2951,\"evidence_need\":0.5894,\"family\":\"performance_tuning\",\"input_tokens\":4560,\"reversibility\":0.8779,\"risk\":0.3193,\"summary\":\"Diagnose latency variance in schema evolution and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.333,\"id\":\"worker_a\",\"p95_latency_ms\":76386,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.878,\"id\":\"worker_b\",\"p95_latency_ms\":70971,\"pass_rate\":0.9009,\"provider_group\":\"provider_x\",\"quality\":0.9109,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.84,\"id\":\"worker_c\",\"p95_latency_ms\":32472,\"pass_rate\":0.4297,\"provider_group\":\"provider_z\",\"quality\":0.6415,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.84,\"estimated_latency_ms\":32472,\"estimated_pass_rate\":0.2695,\"estimated_quality\":0.5144,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.029,\"max_latency_ms\":42236,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.563, "category": "analysis", "complexity": 0.8092, "consequence": 0.2951, "evidence_need": 0.5894, "family": "performance_tuning", "input_tokens": 4560, "request_id": "synthetic-performance_tuning-000272", "reversibility": 0.8779, "risk": 0.3193, "summary": "Diagnose latency variance in schema evolution and rank the next experiments."}, "workers": [{"credits_per_call": 11.333, "id": "worker_a", "p95_latency_ms": 76386, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.878, "id": "worker_b", "p95_latency_ms": 70971, "pass_rate": 0.9009, "provider_group": "provider_x", "quality": 0.9109, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.84, "id": "worker_c", "p95_latency_ms": 32472, "pass_rate": 0.4297, "provider_group": "provider_z", "quality": 0.6415, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.311, "max_latency_ms": 221425, "minimum_quality": 0.7096}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000110", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.479, "estimated_latency_ms": 152215, "estimated_pass_rate": 0.9141, "estimated_quality": 0.9479, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.311, "max_latency_ms": 221425, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9141, "quality": 0.9479, "quality_target_met": true, "utility": 71.188928}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.311,\"max_latency_ms\":221425,\"minimum_quality\":0.7096},\"task\":{\"ambiguity\":0.3838,\"category\":\"analysis\",\"complexity\":0.7007,\"consequence\":0.5831,\"evidence_need\":0.5907,\"family\":\"performance_tuning\",\"input_tokens\":680,\"reversibility\":0.5692,\"risk\":0.4008,\"summary\":\"Choose between throughput and tail-latency changes for database failover.\"},\"workers\":[{\"credits_per_call\":0.783,\"id\":\"worker_a\",\"p95_latency_ms\":30230,\"pass_rate\":0.5133,\"provider_group\":\"provider_z\",\"quality\":0.6625,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.799,\"id\":\"worker_b\",\"p95_latency_ms\":61818,\"pass_rate\":0.9366,\"provider_group\":\"provider_x\",\"quality\":0.8906,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.68,\"id\":\"worker_c\",\"p95_latency_ms\":90397,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.479,\"estimated_latency_ms\":152215,\"estimated_pass_rate\":0.9141,\"estimated_quality\":0.9479,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.311,\"max_latency_ms\":221425,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3838, "category": "analysis", "complexity": 0.7007, "consequence": 0.5831, "evidence_need": 0.5907, "family": "performance_tuning", "input_tokens": 680, "request_id": "synthetic-performance_tuning-000110", "reversibility": 0.5692, "risk": 0.4008, "summary": "Choose between throughput and tail-latency changes for database failover."}, "workers": [{"credits_per_call": 0.783, "id": "worker_a", "p95_latency_ms": 30230, "pass_rate": 0.5133, "provider_group": "provider_z", "quality": 0.6625, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.799, "id": "worker_b", "p95_latency_ms": 61818, "pass_rate": 0.9366, "provider_group": "provider_x", "quality": 0.8906, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.68, "id": "worker_c", "p95_latency_ms": 90397, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.759, "max_latency_ms": 233081, "minimum_quality": 0.7641}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001260", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.745, "estimated_latency_ms": 78959, "estimated_pass_rate": 0.6536, "estimated_quality": 0.814, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.759, "max_latency_ms": 233081, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6536, "quality": 0.814, "quality_target_met": true, "utility": 64.5445}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.759,\"max_latency_ms\":233081,\"minimum_quality\":0.7641},\"task\":{\"ambiguity\":0.6246,\"category\":\"review\",\"complexity\":0.6462,\"consequence\":0.6262,\"evidence_need\":0.5549,\"family\":\"policy_compliance\",\"input_tokens\":6000,\"reversibility\":0.4948,\"risk\":0.8142,\"summary\":\"Review whether webhook delivery satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":15.344,\"id\":\"worker_a\",\"p95_latency_ms\":89499,\"pass_rate\":0.921,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.814,\"id\":\"worker_b\",\"p95_latency_ms\":25739,\"pass_rate\":0.5796,\"provider_group\":\"provider_z\",\"quality\":0.7938,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.931,\"id\":\"worker_c\",\"p95_latency_ms\":53220,\"pass_rate\":0.7711,\"provider_group\":\"provider_y\",\"quality\":0.8765,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.745,\"estimated_latency_ms\":78959,\"estimated_pass_rate\":0.6536,\"estimated_quality\":0.814,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.759,\"max_latency_ms\":233081,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6246, "category": "review", "complexity": 0.6462, "consequence": 0.6262, "evidence_need": 0.5549, "family": "policy_compliance", "input_tokens": 6000, "request_id": "synthetic-policy_compliance-001260", "reversibility": 0.4948, "risk": 0.8142, "summary": "Review whether webhook delivery satisfies a documented retention policy."}, "workers": [{"credits_per_call": 15.344, "id": "worker_a", "p95_latency_ms": 89499, "pass_rate": 0.921, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.814, "id": "worker_b", "p95_latency_ms": 25739, "pass_rate": 0.5796, "provider_group": "provider_z", "quality": 0.7938, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.931, "id": "worker_c", "p95_latency_ms": 53220, "pass_rate": 0.7711, "provider_group": "provider_y", "quality": 0.8765, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.61, "max_latency_ms": 30033, "minimum_quality": 0.6907}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000596", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.846, "estimated_latency_ms": 19319, "estimated_pass_rate": 0.2607, "estimated_quality": 0.516, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.61, "max_latency_ms": 30033, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2607, "quality": 0.516, "quality_target_met": false, "utility": 26.115591}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.61,\"max_latency_ms\":30033,\"minimum_quality\":0.6907},\"task\":{\"ambiguity\":0.7424,\"category\":\"analysis\",\"complexity\":0.805,\"consequence\":0.3473,\"evidence_need\":0.5023,\"family\":\"performance_tuning\",\"input_tokens\":2520,\"reversibility\":0.8699,\"risk\":0.4159,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":0.846,\"id\":\"worker_a\",\"p95_latency_ms\":19319,\"pass_rate\":0.4418,\"provider_group\":\"provider_y\",\"quality\":0.6614,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.305,\"id\":\"worker_b\",\"p95_latency_ms\":74911,\"pass_rate\":0.7894,\"provider_group\":\"provider_z\",\"quality\":0.9381,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":9.959,\"id\":\"worker_c\",\"p95_latency_ms\":59619,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.846,\"estimated_latency_ms\":19319,\"estimated_pass_rate\":0.2607,\"estimated_quality\":0.516,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.61,\"max_latency_ms\":30033,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7424, "category": "analysis", "complexity": 0.805, "consequence": 0.3473, "evidence_need": 0.5023, "family": "performance_tuning", "input_tokens": 2520, "request_id": "synthetic-performance_tuning-000596", "reversibility": 0.8699, "risk": 0.4159, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 0.846, "id": "worker_a", "p95_latency_ms": 19319, "pass_rate": 0.4418, "provider_group": "provider_y", "quality": 0.6614, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.305, "id": "worker_b", "p95_latency_ms": 74911, "pass_rate": 0.7894, "provider_group": "provider_z", "quality": 0.9381, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 9.959, "id": "worker_c", "p95_latency_ms": 59619, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.565, "max_latency_ms": 35666, "minimum_quality": 0.7783}, "family": "migration_planning", "id": "synthetic-migration_planning-000169", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.792, "estimated_latency_ms": 18098, "estimated_pass_rate": 0.3458, "estimated_quality": 0.5399, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.565, "max_latency_ms": 35666, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3458, "quality": 0.5399, "quality_target_met": false, "utility": 25.588473}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.565,\"max_latency_ms\":35666,\"minimum_quality\":0.7783},\"task\":{\"ambiguity\":0.562,\"category\":\"coding\",\"complexity\":0.9429,\"consequence\":0.6319,\"evidence_need\":0.6947,\"family\":\"migration_planning\",\"input_tokens\":6080,\"reversibility\":0.4081,\"risk\":0.3653,\"summary\":\"Plan a zero-downtime migration of rate limiting with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.117,\"id\":\"worker_a\",\"p95_latency_ms\":89002,\"pass_rate\":0.9304,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.018,\"id\":\"worker_b\",\"p95_latency_ms\":75073,\"pass_rate\":0.8491,\"provider_group\":\"provider_y\",\"quality\":0.8808,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.792,\"id\":\"worker_c\",\"p95_latency_ms\":18098,\"pass_rate\":0.5248,\"provider_group\":\"provider_z\",\"quality\":0.6944,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.792,\"estimated_latency_ms\":18098,\"estimated_pass_rate\":0.3458,\"estimated_quality\":0.5399,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.565,\"max_latency_ms\":35666,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.562, "category": "coding", "complexity": 0.9429, "consequence": 0.6319, "evidence_need": 0.6947, "family": "migration_planning", "input_tokens": 6080, "request_id": "synthetic-migration_planning-000169", "reversibility": 0.4081, "risk": 0.3653, "summary": "Plan a zero-downtime migration of rate limiting with rollback checkpoints."}, "workers": [{"credits_per_call": 12.117, "id": "worker_a", "p95_latency_ms": 89002, "pass_rate": 0.9304, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.018, "id": "worker_b", "p95_latency_ms": 75073, "pass_rate": 0.8491, "provider_group": "provider_y", "quality": 0.8808, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.792, "id": "worker_c", "p95_latency_ms": 18098, "pass_rate": 0.5248, "provider_group": "provider_z", "quality": 0.6944, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.194, "max_latency_ms": 116969, "minimum_quality": 0.7899}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000630", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.145, "estimated_latency_ms": 91533, "estimated_pass_rate": 0.6335, "estimated_quality": 0.7877, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.194, "max_latency_ms": 116969, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6335, "quality": 0.7877, "quality_target_met": false, "utility": 54.762195}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.194,\"max_latency_ms\":116969,\"minimum_quality\":0.7899},\"task\":{\"ambiguity\":0.6279,\"category\":\"review\",\"complexity\":0.7511,\"consequence\":0.6559,\"evidence_need\":0.6583,\"family\":\"policy_compliance\",\"input_tokens\":4400,\"reversibility\":0.3746,\"risk\":0.5135,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":15.225,\"id\":\"worker_a\",\"p95_latency_ms\":65106,\"pass_rate\":0.9322,\"provider_group\":\"provider_x\",\"quality\":0.9421,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.004,\"id\":\"worker_b\",\"p95_latency_ms\":20326,\"pass_rate\":0.4867,\"provider_group\":\"provider_z\",\"quality\":0.6876,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.141,\"id\":\"worker_c\",\"p95_latency_ms\":71207,\"pass_rate\":0.735,\"provider_group\":\"provider_y\",\"quality\":0.8816,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.145,\"estimated_latency_ms\":91533,\"estimated_pass_rate\":0.6335,\"estimated_quality\":0.7877,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.194,\"max_latency_ms\":116969,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6279, "category": "review", "complexity": 0.7511, "consequence": 0.6559, "evidence_need": 0.6583, "family": "policy_compliance", "input_tokens": 4400, "request_id": "synthetic-policy_compliance-000630", "reversibility": 0.3746, "risk": 0.5135, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 15.225, "id": "worker_a", "p95_latency_ms": 65106, "pass_rate": 0.9322, "provider_group": "provider_x", "quality": 0.9421, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.004, "id": "worker_b", "p95_latency_ms": 20326, "pass_rate": 0.4867, "provider_group": "provider_z", "quality": 0.6876, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.141, "id": "worker_c", "p95_latency_ms": 71207, "pass_rate": 0.735, "provider_group": "provider_y", "quality": 0.8816, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.284, "max_latency_ms": 250094, "minimum_quality": 0.7496}, "family": "migration_planning", "id": "synthetic-migration_planning-000961", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.298, "estimated_latency_ms": 99309, "estimated_pass_rate": 0.7285, "estimated_quality": 0.827, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.284, "max_latency_ms": 250094, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7285, "quality": 0.827, "quality_target_met": true, "utility": 63.082113}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.284,\"max_latency_ms\":250094,\"minimum_quality\":0.7496},\"task\":{\"ambiguity\":0.4906,\"category\":\"coding\",\"complexity\":0.8387,\"consequence\":0.7516,\"evidence_need\":0.6683,\"family\":\"migration_planning\",\"input_tokens\":6840,\"reversibility\":0.2703,\"risk\":0.4827,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":13.611,\"id\":\"worker_a\",\"p95_latency_ms\":79848,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9422,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.281,\"id\":\"worker_b\",\"p95_latency_ms\":72354,\"pass_rate\":0.8361,\"provider_group\":\"provider_z\",\"quality\":0.9221,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.017,\"id\":\"worker_c\",\"p95_latency_ms\":26955,\"pass_rate\":0.5698,\"provider_group\":\"provider_y\",\"quality\":0.6891,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.298,\"estimated_latency_ms\":99309,\"estimated_pass_rate\":0.7285,\"estimated_quality\":0.827,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.284,\"max_latency_ms\":250094,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4906, "category": "coding", "complexity": 0.8387, "consequence": 0.7516, "evidence_need": 0.6683, "family": "migration_planning", "input_tokens": 6840, "request_id": "synthetic-migration_planning-000961", "reversibility": 0.2703, "risk": 0.4827, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 13.611, "id": "worker_a", "p95_latency_ms": 79848, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9422, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.281, "id": "worker_b", "p95_latency_ms": 72354, "pass_rate": 0.8361, "provider_group": "provider_z", "quality": 0.9221, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.017, "id": "worker_c", "p95_latency_ms": 26955, "pass_rate": 0.5698, "provider_group": "provider_y", "quality": 0.6891, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.863, "max_latency_ms": 53025, "minimum_quality": 0.5506}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000149", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.048, "estimated_latency_ms": 27328, "estimated_pass_rate": 0.2869, "estimated_quality": 0.4798, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.863, "max_latency_ms": 53025, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2869, "quality": 0.4798, "quality_target_met": false, "utility": 28.491145}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.863,\"max_latency_ms\":53025,\"minimum_quality\":0.5506},\"task\":{\"ambiguity\":0.5255,\"category\":\"analysis\",\"complexity\":0.8225,\"consequence\":0.2511,\"evidence_need\":0.6274,\"family\":\"performance_tuning\",\"input_tokens\":3440,\"reversibility\":0.7061,\"risk\":0.5263,\"summary\":\"Choose between throughput and tail-latency changes for artifact storage.\"},\"workers\":[{\"credits_per_call\":11.118,\"id\":\"worker_a\",\"p95_latency_ms\":77767,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.124,\"id\":\"worker_b\",\"p95_latency_ms\":68075,\"pass_rate\":0.889,\"provider_group\":\"provider_y\",\"quality\":0.859,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.048,\"id\":\"worker_c\",\"p95_latency_ms\":27328,\"pass_rate\":0.4625,\"provider_group\":\"provider_x\",\"quality\":0.6217,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.048,\"estimated_latency_ms\":27328,\"estimated_pass_rate\":0.2869,\"estimated_quality\":0.4798,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.863,\"max_latency_ms\":53025,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5255, "category": "analysis", "complexity": 0.8225, "consequence": 0.2511, "evidence_need": 0.6274, "family": "performance_tuning", "input_tokens": 3440, "request_id": "synthetic-performance_tuning-000149", "reversibility": 0.7061, "risk": 0.5263, "summary": "Choose between throughput and tail-latency changes for artifact storage."}, "workers": [{"credits_per_call": 11.118, "id": "worker_a", "p95_latency_ms": 77767, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.124, "id": "worker_b", "p95_latency_ms": 68075, "pass_rate": 0.889, "provider_group": "provider_y", "quality": 0.859, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.048, "id": "worker_c", "p95_latency_ms": 27328, "pass_rate": 0.4625, "provider_group": "provider_x", "quality": 0.6217, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.941, "max_latency_ms": 181685, "minimum_quality": 0.916}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001238", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.821, "estimated_latency_ms": 133208, "estimated_pass_rate": 0.7299, "estimated_quality": 0.9191, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 26.941, "max_latency_ms": 181685, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7299, "quality": 0.9191, "quality_target_met": true, "utility": 72.661727}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.941,\"max_latency_ms\":181685,\"minimum_quality\":0.916},\"task\":{\"ambiguity\":0.6852,\"category\":\"analysis\",\"complexity\":0.6471,\"consequence\":0.4754,\"evidence_need\":0.6693,\"family\":\"performance_tuning\",\"input_tokens\":3520,\"reversibility\":0.7013,\"risk\":0.4535,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.004,\"id\":\"worker_a\",\"p95_latency_ms\":29262,\"pass_rate\":0.4567,\"provider_group\":\"provider_z\",\"quality\":0.694,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.434,\"id\":\"worker_b\",\"p95_latency_ms\":50324,\"pass_rate\":0.8657,\"provider_group\":\"provider_y\",\"quality\":0.8696,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.383,\"id\":\"worker_c\",\"p95_latency_ms\":53622,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.821,\"estimated_latency_ms\":133208,\"estimated_pass_rate\":0.7299,\"estimated_quality\":0.9191,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.941,\"max_latency_ms\":181685,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6852, "category": "analysis", "complexity": 0.6471, "consequence": 0.4754, "evidence_need": 0.6693, "family": "performance_tuning", "input_tokens": 3520, "request_id": "synthetic-performance_tuning-001238", "reversibility": 0.7013, "risk": 0.4535, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 1.004, "id": "worker_a", "p95_latency_ms": 29262, "pass_rate": 0.4567, "provider_group": "provider_z", "quality": 0.694, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.434, "id": "worker_b", "p95_latency_ms": 50324, "pass_rate": 0.8657, "provider_group": "provider_y", "quality": 0.8696, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.383, "id": "worker_c", "p95_latency_ms": 53622, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.053, "max_latency_ms": 171405, "minimum_quality": 0.8616}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001047", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.242, "estimated_latency_ms": 86722, "estimated_pass_rate": 0.7772, "estimated_quality": 0.8423, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.053, "max_latency_ms": 171405, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7772, "quality": 0.8423, "quality_target_met": false, "utility": 62.589635}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.053,\"max_latency_ms\":171405,\"minimum_quality\":0.8616},\"task\":{\"ambiguity\":0.4684,\"category\":\"review\",\"complexity\":0.5806,\"consequence\":0.8168,\"evidence_need\":0.5607,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.5783,\"risk\":0.6172,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.182,\"id\":\"worker_a\",\"p95_latency_ms\":29069,\"pass_rate\":0.5277,\"provider_group\":\"provider_z\",\"quality\":0.7346,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.201,\"id\":\"worker_b\",\"p95_latency_ms\":49255,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9807,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.06,\"id\":\"worker_c\",\"p95_latency_ms\":57653,\"pass_rate\":0.8721,\"provider_group\":\"provider_y\",\"quality\":0.9058,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.242,\"estimated_latency_ms\":86722,\"estimated_pass_rate\":0.7772,\"estimated_quality\":0.8423,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.053,\"max_latency_ms\":171405,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4684, "category": "review", "complexity": 0.5806, "consequence": 0.8168, "evidence_need": 0.5607, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-001047", "reversibility": 0.5783, "risk": 0.6172, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.182, "id": "worker_a", "p95_latency_ms": 29069, "pass_rate": 0.5277, "provider_group": "provider_z", "quality": 0.7346, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.201, "id": "worker_b", "p95_latency_ms": 49255, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9807, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.06, "id": "worker_c", "p95_latency_ms": 57653, "pass_rate": 0.8721, "provider_group": "provider_y", "quality": 0.9058, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.679, "max_latency_ms": 27805, "minimum_quality": 0.7425}, "family": "migration_planning", "id": "synthetic-migration_planning-000265", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.063, "estimated_latency_ms": 20340, "estimated_pass_rate": 0.4598, "estimated_quality": 0.63, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.679, "max_latency_ms": 27805, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4598, "quality": 0.63, "quality_target_met": false, "utility": 37.186741}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.679,\"max_latency_ms\":27805,\"minimum_quality\":0.7425},\"task\":{\"ambiguity\":0.5579,\"category\":\"coding\",\"complexity\":0.6492,\"consequence\":0.8161,\"evidence_need\":0.8298,\"family\":\"migration_planning\",\"input_tokens\":4400,\"reversibility\":0.3613,\"risk\":0.371,\"summary\":\"Review a staged migration for credit reservation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.192,\"id\":\"worker_a\",\"p95_latency_ms\":48677,\"pass_rate\":0.8579,\"provider_group\":\"provider_x\",\"quality\":0.8982,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.561,\"id\":\"worker_b\",\"p95_latency_ms\":53979,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9691,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.063,\"id\":\"worker_c\",\"p95_latency_ms\":20340,\"pass_rate\":0.6056,\"provider_group\":\"provider_y\",\"quality\":0.7316,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.063,\"estimated_latency_ms\":20340,\"estimated_pass_rate\":0.4598,\"estimated_quality\":0.63,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.679,\"max_latency_ms\":27805,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5579, "category": "coding", "complexity": 0.6492, "consequence": 0.8161, "evidence_need": 0.8298, "family": "migration_planning", "input_tokens": 4400, "request_id": "synthetic-migration_planning-000265", "reversibility": 0.3613, "risk": 0.371, "summary": "Review a staged migration for credit reservation and identify irreversible steps."}, "workers": [{"credits_per_call": 3.192, "id": "worker_a", "p95_latency_ms": 48677, "pass_rate": 0.8579, "provider_group": "provider_x", "quality": 0.8982, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.561, "id": "worker_b", "p95_latency_ms": 53979, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9691, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.063, "id": "worker_c", "p95_latency_ms": 20340, "pass_rate": 0.6056, "provider_group": "provider_y", "quality": 0.7316, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.677, "max_latency_ms": 192071, "minimum_quality": 0.7503}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000762", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.956, "estimated_latency_ms": 98410, "estimated_pass_rate": 0.7931, "estimated_quality": 0.7948, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.677, "max_latency_ms": 192071, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7931, "quality": 0.7948, "quality_target_met": true, "utility": 61.883917}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.677,\"max_latency_ms\":192071,\"minimum_quality\":0.7503},\"task\":{\"ambiguity\":0.4699,\"category\":\"review\",\"complexity\":0.4044,\"consequence\":0.7095,\"evidence_need\":0.5502,\"family\":\"policy_compliance\",\"input_tokens\":2840,\"reversibility\":0.5794,\"risk\":0.5758,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.114,\"id\":\"worker_a\",\"p95_latency_ms\":22542,\"pass_rate\":0.576,\"provider_group\":\"provider_z\",\"quality\":0.6798,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.361,\"id\":\"worker_b\",\"p95_latency_ms\":84506,\"pass_rate\":0.9223,\"provider_group\":\"provider_y\",\"quality\":0.9117,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.842,\"id\":\"worker_c\",\"p95_latency_ms\":75868,\"pass_rate\":0.8711,\"provider_group\":\"provider_x\",\"quality\":0.8082,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.956,\"estimated_latency_ms\":98410,\"estimated_pass_rate\":0.7931,\"estimated_quality\":0.7948,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.677,\"max_latency_ms\":192071,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4699, "category": "review", "complexity": 0.4044, "consequence": 0.7095, "evidence_need": 0.5502, "family": "policy_compliance", "input_tokens": 2840, "request_id": "synthetic-policy_compliance-000762", "reversibility": 0.5794, "risk": 0.5758, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.114, "id": "worker_a", "p95_latency_ms": 22542, "pass_rate": 0.576, "provider_group": "provider_z", "quality": 0.6798, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.361, "id": "worker_b", "p95_latency_ms": 84506, "pass_rate": 0.9223, "provider_group": "provider_y", "quality": 0.9117, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.842, "id": "worker_c", "p95_latency_ms": 75868, "pass_rate": 0.8711, "provider_group": "provider_x", "quality": 0.8082, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.085, "max_latency_ms": 205506, "minimum_quality": 0.7946}, "family": "migration_planning", "id": "synthetic-migration_planning-001168", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.932, "estimated_latency_ms": 162470, "estimated_pass_rate": 0.7677, "estimated_quality": 0.9196, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.085, "max_latency_ms": 205506, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7677, "quality": 0.9196, "quality_target_met": true, "utility": 72.562211}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.085,\"max_latency_ms\":205506,\"minimum_quality\":0.7946},\"task\":{\"ambiguity\":0.7437,\"category\":\"coding\",\"complexity\":0.6347,\"consequence\":0.5418,\"evidence_need\":0.4549,\"family\":\"migration_planning\",\"input_tokens\":5960,\"reversibility\":0.5578,\"risk\":0.534,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":3.857,\"id\":\"worker_a\",\"p95_latency_ms\":53362,\"pass_rate\":0.841,\"provider_group\":\"provider_y\",\"quality\":0.8778,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.03,\"id\":\"worker_b\",\"p95_latency_ms\":76433,\"pass_rate\":0.953,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.045,\"id\":\"worker_c\",\"p95_latency_ms\":32675,\"pass_rate\":0.5961,\"provider_group\":\"provider_z\",\"quality\":0.7664,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.932,\"estimated_latency_ms\":162470,\"estimated_pass_rate\":0.7677,\"estimated_quality\":0.9196,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.085,\"max_latency_ms\":205506,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7437, "category": "coding", "complexity": 0.6347, "consequence": 0.5418, "evidence_need": 0.4549, "family": "migration_planning", "input_tokens": 5960, "request_id": "synthetic-migration_planning-001168", "reversibility": 0.5578, "risk": 0.534, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 3.857, "id": "worker_a", "p95_latency_ms": 53362, "pass_rate": 0.841, "provider_group": "provider_y", "quality": 0.8778, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.03, "id": "worker_b", "p95_latency_ms": 76433, "pass_rate": 0.953, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.045, "id": "worker_c", "p95_latency_ms": 32675, "pass_rate": 0.5961, "provider_group": "provider_z", "quality": 0.7664, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.309, "max_latency_ms": 171501, "minimum_quality": 0.7838}, "family": "migration_planning", "id": "synthetic-migration_planning-000244", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.617, "estimated_latency_ms": 161771, "estimated_pass_rate": 0.7301, "estimated_quality": 0.9289, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.309, "max_latency_ms": 171501, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7301, "quality": 0.9289, "quality_target_met": true, "utility": 73.687821}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.309,\"max_latency_ms\":171501,\"minimum_quality\":0.7838},\"task\":{\"ambiguity\":0.7223,\"category\":\"coding\",\"complexity\":0.5789,\"consequence\":0.7183,\"evidence_need\":0.6645,\"family\":\"migration_planning\",\"input_tokens\":7080,\"reversibility\":0.5054,\"risk\":0.5523,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.003,\"id\":\"worker_a\",\"p95_latency_ms\":24421,\"pass_rate\":0.5155,\"provider_group\":\"provider_y\",\"quality\":0.8004,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.811,\"id\":\"worker_b\",\"p95_latency_ms\":57769,\"pass_rate\":0.9637,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.803,\"id\":\"worker_c\",\"p95_latency_ms\":79581,\"pass_rate\":0.7377,\"provider_group\":\"provider_z\",\"quality\":0.9029,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.617,\"estimated_latency_ms\":161771,\"estimated_pass_rate\":0.7301,\"estimated_quality\":0.9289,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.309,\"max_latency_ms\":171501,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7223, "category": "coding", "complexity": 0.5789, "consequence": 0.7183, "evidence_need": 0.6645, "family": "migration_planning", "input_tokens": 7080, "request_id": "synthetic-migration_planning-000244", "reversibility": 0.5054, "risk": 0.5523, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 1.003, "id": "worker_a", "p95_latency_ms": 24421, "pass_rate": 0.5155, "provider_group": "provider_y", "quality": 0.8004, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.811, "id": "worker_b", "p95_latency_ms": 57769, "pass_rate": 0.9637, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.803, "id": "worker_c", "p95_latency_ms": 79581, "pass_rate": 0.7377, "provider_group": "provider_z", "quality": 0.9029, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.534, "max_latency_ms": 209827, "minimum_quality": 0.8987}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001172", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.667, "estimated_latency_ms": 169888, "estimated_pass_rate": 0.7345, "estimated_quality": 0.9252, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.534, "max_latency_ms": 209827, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7345, "quality": 0.9252, "quality_target_met": true, "utility": 69.827928}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.534,\"max_latency_ms\":209827,\"minimum_quality\":0.8987},\"task\":{\"ambiguity\":0.3702,\"category\":\"analysis\",\"complexity\":0.8504,\"consequence\":0.3933,\"evidence_need\":0.7119,\"family\":\"performance_tuning\",\"input_tokens\":6120,\"reversibility\":0.843,\"risk\":0.2711,\"summary\":\"Diagnose latency variance in schema evolution and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":10.613,\"id\":\"worker_a\",\"p95_latency_ms\":82443,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9924,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.149,\"id\":\"worker_b\",\"p95_latency_ms\":26958,\"pass_rate\":0.492,\"provider_group\":\"provider_x\",\"quality\":0.7259,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.905,\"id\":\"worker_c\",\"p95_latency_ms\":60487,\"pass_rate\":0.8537,\"provider_group\":\"provider_z\",\"quality\":0.9086,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.667,\"estimated_latency_ms\":169888,\"estimated_pass_rate\":0.7345,\"estimated_quality\":0.9252,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.534,\"max_latency_ms\":209827,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3702, "category": "analysis", "complexity": 0.8504, "consequence": 0.3933, "evidence_need": 0.7119, "family": "performance_tuning", "input_tokens": 6120, "request_id": "synthetic-performance_tuning-001172", "reversibility": 0.843, "risk": 0.2711, "summary": "Diagnose latency variance in schema evolution and rank the next experiments."}, "workers": [{"credits_per_call": 10.613, "id": "worker_a", "p95_latency_ms": 82443, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9924, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.149, "id": "worker_b", "p95_latency_ms": 26958, "pass_rate": 0.492, "provider_group": "provider_x", "quality": 0.7259, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.905, "id": "worker_c", "p95_latency_ms": 60487, "pass_rate": 0.8537, "provider_group": "provider_z", "quality": 0.9086, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.539, "max_latency_ms": 34949, "minimum_quality": 0.6986}, "family": "migration_planning", "id": "synthetic-migration_planning-000160", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.079, "estimated_latency_ms": 22525, "estimated_pass_rate": 0.3045, "estimated_quality": 0.6894, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.539, "max_latency_ms": 34949, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3045, "quality": 0.6894, "quality_target_met": false, "utility": 36.825384}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.539,\"max_latency_ms\":34949,\"minimum_quality\":0.6986},\"task\":{\"ambiguity\":0.4337,\"category\":\"coding\",\"complexity\":0.6833,\"consequence\":0.4623,\"evidence_need\":0.4532,\"family\":\"migration_planning\",\"input_tokens\":400,\"reversibility\":0.2598,\"risk\":0.4403,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":2.968,\"id\":\"worker_a\",\"p95_latency_ms\":72509,\"pass_rate\":0.7692,\"provider_group\":\"provider_z\",\"quality\":0.8825,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.079,\"id\":\"worker_b\",\"p95_latency_ms\":22525,\"pass_rate\":0.4504,\"provider_group\":\"provider_x\",\"quality\":0.7938,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.451,\"id\":\"worker_c\",\"p95_latency_ms\":51529,\"pass_rate\":0.9306,\"provider_group\":\"provider_y\",\"quality\":0.9549,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.079,\"estimated_latency_ms\":22525,\"estimated_pass_rate\":0.3045,\"estimated_quality\":0.6894,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.539,\"max_latency_ms\":34949,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4337, "category": "coding", "complexity": 0.6833, "consequence": 0.4623, "evidence_need": 0.4532, "family": "migration_planning", "input_tokens": 400, "request_id": "synthetic-migration_planning-000160", "reversibility": 0.2598, "risk": 0.4403, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 2.968, "id": "worker_a", "p95_latency_ms": 72509, "pass_rate": 0.7692, "provider_group": "provider_z", "quality": 0.8825, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.079, "id": "worker_b", "p95_latency_ms": 22525, "pass_rate": 0.4504, "provider_group": "provider_x", "quality": 0.7938, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.451, "id": "worker_c", "p95_latency_ms": 51529, "pass_rate": 0.9306, "provider_group": "provider_y", "quality": 0.9549, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.489, "max_latency_ms": 144746, "minimum_quality": 0.7748}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000714", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.916, "estimated_latency_ms": 101699, "estimated_pass_rate": 0.6362, "estimated_quality": 0.7976, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.489, "max_latency_ms": 144746, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6362, "quality": 0.7976, "quality_target_met": true, "utility": 61.05663}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.489,\"max_latency_ms\":144746,\"minimum_quality\":0.7748},\"task\":{\"ambiguity\":0.5271,\"category\":\"review\",\"complexity\":0.646,\"consequence\":0.8409,\"evidence_need\":0.6897,\"family\":\"policy_compliance\",\"input_tokens\":3280,\"reversibility\":0.6248,\"risk\":0.7635,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":0.921,\"id\":\"worker_a\",\"p95_latency_ms\":30266,\"pass_rate\":0.5541,\"provider_group\":\"provider_y\",\"quality\":0.7348,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.995,\"id\":\"worker_b\",\"p95_latency_ms\":71433,\"pass_rate\":0.7508,\"provider_group\":\"provider_x\",\"quality\":0.8923,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.243,\"id\":\"worker_c\",\"p95_latency_ms\":55309,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9823,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.916,\"estimated_latency_ms\":101699,\"estimated_pass_rate\":0.6362,\"estimated_quality\":0.7976,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.489,\"max_latency_ms\":144746,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5271, "category": "review", "complexity": 0.646, "consequence": 0.8409, "evidence_need": 0.6897, "family": "policy_compliance", "input_tokens": 3280, "request_id": "synthetic-policy_compliance-000714", "reversibility": 0.6248, "risk": 0.7635, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 0.921, "id": "worker_a", "p95_latency_ms": 30266, "pass_rate": 0.5541, "provider_group": "provider_y", "quality": 0.7348, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.995, "id": "worker_b", "p95_latency_ms": 71433, "pass_rate": 0.7508, "provider_group": "provider_x", "quality": 0.8923, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.243, "id": "worker_c", "p95_latency_ms": 55309, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9823, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.129, "max_latency_ms": 248157, "minimum_quality": 0.9592}, "family": "migration_planning", "id": "synthetic-migration_planning-001144", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.214, "estimated_latency_ms": 173698, "estimated_pass_rate": 0.7588, "estimated_quality": 0.8993, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 15.129, "max_latency_ms": 248157, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7588, "quality": 0.8993, "quality_target_met": false, "utility": 69.91389}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.129,\"max_latency_ms\":248157,\"minimum_quality\":0.9592},\"task\":{\"ambiguity\":0.5257,\"category\":\"coding\",\"complexity\":0.7102,\"consequence\":0.8006,\"evidence_need\":0.6159,\"family\":\"migration_planning\",\"input_tokens\":1840,\"reversibility\":0.3678,\"risk\":0.517,\"summary\":\"Review a staged migration for credit reservation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.873,\"id\":\"worker_a\",\"p95_latency_ms\":30239,\"pass_rate\":0.6091,\"provider_group\":\"provider_y\",\"quality\":0.7352,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.448,\"id\":\"worker_b\",\"p95_latency_ms\":71134,\"pass_rate\":0.7963,\"provider_group\":\"provider_z\",\"quality\":0.8408,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":9.893,\"id\":\"worker_c\",\"p95_latency_ms\":72325,\"pass_rate\":0.9454,\"provider_group\":\"provider_x\",\"quality\":0.9661,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.214,\"estimated_latency_ms\":173698,\"estimated_pass_rate\":0.7588,\"estimated_quality\":0.8993,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":15.129,\"max_latency_ms\":248157,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5257, "category": "coding", "complexity": 0.7102, "consequence": 0.8006, "evidence_need": 0.6159, "family": "migration_planning", "input_tokens": 1840, "request_id": "synthetic-migration_planning-001144", "reversibility": 0.3678, "risk": 0.517, "summary": "Review a staged migration for credit reservation and identify irreversible steps."}, "workers": [{"credits_per_call": 0.873, "id": "worker_a", "p95_latency_ms": 30239, "pass_rate": 0.6091, "provider_group": "provider_y", "quality": 0.7352, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.448, "id": "worker_b", "p95_latency_ms": 71134, "pass_rate": 0.7963, "provider_group": "provider_z", "quality": 0.8408, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 9.893, "id": "worker_c", "p95_latency_ms": 72325, "pass_rate": 0.9454, "provider_group": "provider_x", "quality": 0.9661, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.778, "max_latency_ms": 240884, "minimum_quality": 0.895}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000726", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.812, "estimated_latency_ms": 171198, "estimated_pass_rate": 0.7411, "estimated_quality": 0.9153, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.778, "max_latency_ms": 240884, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7411, "quality": 0.9153, "quality_target_met": true, "utility": 76.605472}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.778,\"max_latency_ms\":240884,\"minimum_quality\":0.895},\"task\":{\"ambiguity\":0.3705,\"category\":\"review\",\"complexity\":0.8476,\"consequence\":0.6212,\"evidence_need\":0.569,\"family\":\"policy_compliance\",\"input_tokens\":7440,\"reversibility\":0.3944,\"risk\":0.8187,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.293,\"id\":\"worker_a\",\"p95_latency_ms\":89095,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.557,\"id\":\"worker_b\",\"p95_latency_ms\":57555,\"pass_rate\":0.8283,\"provider_group\":\"provider_x\",\"quality\":0.832,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.962,\"id\":\"worker_c\",\"p95_latency_ms\":24548,\"pass_rate\":0.451,\"provider_group\":\"provider_y\",\"quality\":0.7593,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.812,\"estimated_latency_ms\":171198,\"estimated_pass_rate\":0.7411,\"estimated_quality\":0.9153,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.778,\"max_latency_ms\":240884,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3705, "category": "review", "complexity": 0.8476, "consequence": 0.6212, "evidence_need": 0.569, "family": "policy_compliance", "input_tokens": 7440, "request_id": "synthetic-policy_compliance-000726", "reversibility": 0.3944, "risk": 0.8187, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 10.293, "id": "worker_a", "p95_latency_ms": 89095, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.557, "id": "worker_b", "p95_latency_ms": 57555, "pass_rate": 0.8283, "provider_group": "provider_x", "quality": 0.832, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.962, "id": "worker_c", "p95_latency_ms": 24548, "pass_rate": 0.451, "provider_group": "provider_y", "quality": 0.7593, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.142, "max_latency_ms": 162854, "minimum_quality": 0.9305}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001134", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.722, "estimated_latency_ms": 148409, "estimated_pass_rate": 0.7595, "estimated_quality": 0.9141, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.142, "max_latency_ms": 162854, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7595, "quality": 0.9141, "quality_target_met": false, "utility": 74.585676}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.142,\"max_latency_ms\":162854,\"minimum_quality\":0.9305},\"task\":{\"ambiguity\":0.7268,\"category\":\"review\",\"complexity\":0.5929,\"consequence\":0.5428,\"evidence_need\":0.6805,\"family\":\"policy_compliance\",\"input_tokens\":5920,\"reversibility\":0.3255,\"risk\":0.7178,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.814,\"id\":\"worker_a\",\"p95_latency_ms\":44620,\"pass_rate\":0.7208,\"provider_group\":\"provider_y\",\"quality\":0.887,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.791,\"id\":\"worker_b\",\"p95_latency_ms\":28656,\"pass_rate\":0.5547,\"provider_group\":\"provider_z\",\"quality\":0.8006,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.117,\"id\":\"worker_c\",\"p95_latency_ms\":75133,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9694,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.722,\"estimated_latency_ms\":148409,\"estimated_pass_rate\":0.7595,\"estimated_quality\":0.9141,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.142,\"max_latency_ms\":162854,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7268, "category": "review", "complexity": 0.5929, "consequence": 0.5428, "evidence_need": 0.6805, "family": "policy_compliance", "input_tokens": 5920, "request_id": "synthetic-policy_compliance-001134", "reversibility": 0.3255, "risk": 0.7178, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 3.814, "id": "worker_a", "p95_latency_ms": 44620, "pass_rate": 0.7208, "provider_group": "provider_y", "quality": 0.887, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.791, "id": "worker_b", "p95_latency_ms": 28656, "pass_rate": 0.5547, "provider_group": "provider_z", "quality": 0.8006, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.117, "id": "worker_c", "p95_latency_ms": 75133, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9694, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.795, "max_latency_ms": 164766, "minimum_quality": 0.8443}, "family": "migration_planning", "id": "synthetic-migration_planning-000247", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.437, "estimated_latency_ms": 99042, "estimated_pass_rate": 0.6261, "estimated_quality": 0.7991, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.795, "max_latency_ms": 164766, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6261, "quality": 0.7991, "quality_target_met": false, "utility": 59.698841}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.795,\"max_latency_ms\":164766,\"minimum_quality\":0.8443},\"task\":{\"ambiguity\":0.5819,\"category\":\"coding\",\"complexity\":0.7139,\"consequence\":0.7345,\"evidence_need\":0.7641,\"family\":\"migration_planning\",\"input_tokens\":3480,\"reversibility\":0.3821,\"risk\":0.601,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.017,\"id\":\"worker_a\",\"p95_latency_ms\":20952,\"pass_rate\":0.5385,\"provider_group\":\"provider_z\",\"quality\":0.7074,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.42,\"id\":\"worker_b\",\"p95_latency_ms\":78090,\"pass_rate\":0.8474,\"provider_group\":\"provider_y\",\"quality\":0.8616,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.726,\"id\":\"worker_c\",\"p95_latency_ms\":77220,\"pass_rate\":0.9139,\"provider_group\":\"provider_x\",\"quality\":0.974,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.437,\"estimated_latency_ms\":99042,\"estimated_pass_rate\":0.6261,\"estimated_quality\":0.7991,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.795,\"max_latency_ms\":164766,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5819, "category": "coding", "complexity": 0.7139, "consequence": 0.7345, "evidence_need": 0.7641, "family": "migration_planning", "input_tokens": 3480, "request_id": "synthetic-migration_planning-000247", "reversibility": 0.3821, "risk": 0.601, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 1.017, "id": "worker_a", "p95_latency_ms": 20952, "pass_rate": 0.5385, "provider_group": "provider_z", "quality": 0.7074, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.42, "id": "worker_b", "p95_latency_ms": 78090, "pass_rate": 0.8474, "provider_group": "provider_y", "quality": 0.8616, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.726, "id": "worker_c", "p95_latency_ms": 77220, "pass_rate": 0.9139, "provider_group": "provider_x", "quality": 0.974, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.953, "max_latency_ms": 162782, "minimum_quality": 0.8166}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001091", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.419, "estimated_latency_ms": 107195, "estimated_pass_rate": 0.8855, "estimated_quality": 0.9132, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.953, "max_latency_ms": 162782, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8855, "quality": 0.9132, "quality_target_met": true, "utility": 67.025261}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.953,\"max_latency_ms\":162782,\"minimum_quality\":0.8166},\"task\":{\"ambiguity\":0.5437,\"category\":\"analysis\",\"complexity\":0.806,\"consequence\":0.2521,\"evidence_need\":0.5071,\"family\":\"performance_tuning\",\"input_tokens\":280,\"reversibility\":0.5907,\"risk\":0.5392,\"summary\":\"Choose between throughput and tail-latency changes for streaming inference.\"},\"workers\":[{\"credits_per_call\":1.019,\"id\":\"worker_a\",\"p95_latency_ms\":30313,\"pass_rate\":0.4,\"provider_group\":\"provider_y\",\"quality\":0.6361,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.462,\"id\":\"worker_b\",\"p95_latency_ms\":53285,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.957,\"id\":\"worker_c\",\"p95_latency_ms\":53910,\"pass_rate\":0.7955,\"provider_group\":\"provider_x\",\"quality\":0.869,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.419,\"estimated_latency_ms\":107195,\"estimated_pass_rate\":0.8855,\"estimated_quality\":0.9132,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.953,\"max_latency_ms\":162782,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5437, "category": "analysis", "complexity": 0.806, "consequence": 0.2521, "evidence_need": 0.5071, "family": "performance_tuning", "input_tokens": 280, "request_id": "synthetic-performance_tuning-001091", "reversibility": 0.5907, "risk": 0.5392, "summary": "Choose between throughput and tail-latency changes for streaming inference."}, "workers": [{"credits_per_call": 1.019, "id": "worker_a", "p95_latency_ms": 30313, "pass_rate": 0.4, "provider_group": "provider_y", "quality": 0.6361, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.462, "id": "worker_b", "p95_latency_ms": 53285, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.957, "id": "worker_c", "p95_latency_ms": 53910, "pass_rate": 0.7955, "provider_group": "provider_x", "quality": 0.869, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.492, "max_latency_ms": 42173, "minimum_quality": 0.7271}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000594", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.164, "estimated_latency_ms": 21673, "estimated_pass_rate": 0.3952, "estimated_quality": 0.6006, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.492, "max_latency_ms": 42173, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3952, "quality": 0.6006, "quality_target_met": false, "utility": 20.154049}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.492,\"max_latency_ms\":42173,\"minimum_quality\":0.7271},\"task\":{\"ambiguity\":0.3857,\"category\":\"review\",\"complexity\":0.4969,\"consequence\":0.7783,\"evidence_need\":0.7148,\"family\":\"policy_compliance\",\"input_tokens\":7640,\"reversibility\":0.6393,\"risk\":0.6889,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":2.798,\"id\":\"worker_a\",\"p95_latency_ms\":62488,\"pass_rate\":0.8493,\"provider_group\":\"provider_y\",\"quality\":0.8288,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.673,\"id\":\"worker_b\",\"p95_latency_ms\":64872,\"pass_rate\":0.9341,\"provider_group\":\"provider_z\",\"quality\":0.9421,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.164,\"id\":\"worker_c\",\"p95_latency_ms\":21673,\"pass_rate\":0.5365,\"provider_group\":\"provider_x\",\"quality\":0.686,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.164,\"estimated_latency_ms\":21673,\"estimated_pass_rate\":0.3952,\"estimated_quality\":0.6006,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.492,\"max_latency_ms\":42173,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3857, "category": "review", "complexity": 0.4969, "consequence": 0.7783, "evidence_need": 0.7148, "family": "policy_compliance", "input_tokens": 7640, "request_id": "synthetic-policy_compliance-000594", "reversibility": 0.6393, "risk": 0.6889, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 2.798, "id": "worker_a", "p95_latency_ms": 62488, "pass_rate": 0.8493, "provider_group": "provider_y", "quality": 0.8288, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 11.673, "id": "worker_b", "p95_latency_ms": 64872, "pass_rate": 0.9341, "provider_group": "provider_z", "quality": 0.9421, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.164, "id": "worker_c", "p95_latency_ms": 21673, "pass_rate": 0.5365, "provider_group": "provider_x", "quality": 0.686, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.779, "max_latency_ms": 39413, "minimum_quality": 0.5458}, "family": "migration_planning", "id": "synthetic-migration_planning-000271", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.176, "estimated_latency_ms": 29960, "estimated_pass_rate": 0.4043, "estimated_quality": 0.6577, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.779, "max_latency_ms": 39413, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4043, "quality": 0.6577, "quality_target_met": true, "utility": 42.028283}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.779,\"max_latency_ms\":39413,\"minimum_quality\":0.5458},\"task\":{\"ambiguity\":0.4489,\"category\":\"coding\",\"complexity\":0.9007,\"consequence\":0.6805,\"evidence_need\":0.4787,\"family\":\"migration_planning\",\"input_tokens\":360,\"reversibility\":0.3634,\"risk\":0.4718,\"summary\":\"Review a staged migration for database failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.176,\"id\":\"worker_a\",\"p95_latency_ms\":29960,\"pass_rate\":0.5786,\"provider_group\":\"provider_x\",\"quality\":0.8047,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.135,\"id\":\"worker_b\",\"p95_latency_ms\":49495,\"pass_rate\":0.9563,\"provider_group\":\"provider_z\",\"quality\":0.9653,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.194,\"id\":\"worker_c\",\"p95_latency_ms\":77649,\"pass_rate\":0.8352,\"provider_group\":\"provider_y\",\"quality\":0.83,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.176,\"estimated_latency_ms\":29960,\"estimated_pass_rate\":0.4043,\"estimated_quality\":0.6577,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.779,\"max_latency_ms\":39413,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4489, "category": "coding", "complexity": 0.9007, "consequence": 0.6805, "evidence_need": 0.4787, "family": "migration_planning", "input_tokens": 360, "request_id": "synthetic-migration_planning-000271", "reversibility": 0.3634, "risk": 0.4718, "summary": "Review a staged migration for database failover and identify irreversible steps."}, "workers": [{"credits_per_call": 1.176, "id": "worker_a", "p95_latency_ms": 29960, "pass_rate": 0.5786, "provider_group": "provider_x", "quality": 0.8047, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.135, "id": "worker_b", "p95_latency_ms": 49495, "pass_rate": 0.9563, "provider_group": "provider_z", "quality": 0.9653, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.194, "id": "worker_c", "p95_latency_ms": 77649, "pass_rate": 0.8352, "provider_group": "provider_y", "quality": 0.83, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 1.425, "max_latency_ms": 39295, "minimum_quality": 0.5639}, "family": "migration_planning", "id": "synthetic-migration_planning-000880", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.887, "estimated_latency_ms": 23961, "estimated_pass_rate": 0.2848, "estimated_quality": 0.6207, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.425, "max_latency_ms": 39295, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2848, "quality": 0.6207, "quality_target_met": true, "utility": 37.932632}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":1.425,\"max_latency_ms\":39295,\"minimum_quality\":0.5639},\"task\":{\"ambiguity\":0.3491,\"category\":\"coding\",\"complexity\":0.8359,\"consequence\":0.6481,\"evidence_need\":0.4856,\"family\":\"migration_planning\",\"input_tokens\":2320,\"reversibility\":0.3028,\"risk\":0.6918,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.887,\"id\":\"worker_a\",\"p95_latency_ms\":23961,\"pass_rate\":0.462,\"provider_group\":\"provider_x\",\"quality\":0.7651,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.182,\"id\":\"worker_b\",\"p95_latency_ms\":64849,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.784,\"id\":\"worker_c\",\"p95_latency_ms\":67173,\"pass_rate\":0.8092,\"provider_group\":\"provider_z\",\"quality\":0.8107,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.887,\"estimated_latency_ms\":23961,\"estimated_pass_rate\":0.2848,\"estimated_quality\":0.6207,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.425,\"max_latency_ms\":39295,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3491, "category": "coding", "complexity": 0.8359, "consequence": 0.6481, "evidence_need": 0.4856, "family": "migration_planning", "input_tokens": 2320, "request_id": "synthetic-migration_planning-000880", "reversibility": 0.3028, "risk": 0.6918, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 0.887, "id": "worker_a", "p95_latency_ms": 23961, "pass_rate": 0.462, "provider_group": "provider_x", "quality": 0.7651, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.182, "id": "worker_b", "p95_latency_ms": 64849, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.784, "id": "worker_c", "p95_latency_ms": 67173, "pass_rate": 0.8092, "provider_group": "provider_z", "quality": 0.8107, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.203, "max_latency_ms": 150236, "minimum_quality": 0.8232}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001203", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.501, "estimated_latency_ms": 87077, "estimated_pass_rate": 0.7081, "estimated_quality": 0.7932, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.203, "max_latency_ms": 150236, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7081, "quality": 0.7932, "quality_target_met": false, "utility": 58.866702}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.203,\"max_latency_ms\":150236,\"minimum_quality\":0.8232},\"task\":{\"ambiguity\":0.4239,\"category\":\"review\",\"complexity\":0.4831,\"consequence\":0.8799,\"evidence_need\":0.6895,\"family\":\"policy_compliance\",\"input_tokens\":680,\"reversibility\":0.5262,\"risk\":0.5789,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.545,\"id\":\"worker_a\",\"p95_latency_ms\":63159,\"pass_rate\":0.7667,\"provider_group\":\"provider_x\",\"quality\":0.8091,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.956,\"id\":\"worker_b\",\"p95_latency_ms\":23918,\"pass_rate\":0.5686,\"provider_group\":\"provider_z\",\"quality\":0.6885,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.606,\"id\":\"worker_c\",\"p95_latency_ms\":50515,\"pass_rate\":0.9677,\"provider_group\":\"provider_y\",\"quality\":0.9948,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.501,\"estimated_latency_ms\":87077,\"estimated_pass_rate\":0.7081,\"estimated_quality\":0.7932,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.203,\"max_latency_ms\":150236,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4239, "category": "review", "complexity": 0.4831, "consequence": 0.8799, "evidence_need": 0.6895, "family": "policy_compliance", "input_tokens": 680, "request_id": "synthetic-policy_compliance-001203", "reversibility": 0.5262, "risk": 0.5789, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.545, "id": "worker_a", "p95_latency_ms": 63159, "pass_rate": 0.7667, "provider_group": "provider_x", "quality": 0.8091, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.956, "id": "worker_b", "p95_latency_ms": 23918, "pass_rate": 0.5686, "provider_group": "provider_z", "quality": 0.6885, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.606, "id": "worker_c", "p95_latency_ms": 50515, "pass_rate": 0.9677, "provider_group": "provider_y", "quality": 0.9948, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.297, "max_latency_ms": 206510, "minimum_quality": 0.7873}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000635", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.947, "estimated_latency_ms": 115999, "estimated_pass_rate": 0.8838, "estimated_quality": 0.9413, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.297, "max_latency_ms": 206510, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8838, "quality": 0.9413, "quality_target_met": true, "utility": 68.947092}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.297,\"max_latency_ms\":206510,\"minimum_quality\":0.7873},\"task\":{\"ambiguity\":0.4491,\"category\":\"analysis\",\"complexity\":0.6328,\"consequence\":0.261,\"evidence_need\":0.3483,\"family\":\"performance_tuning\",\"input_tokens\":3640,\"reversibility\":0.6774,\"risk\":0.3601,\"summary\":\"Review a proposed optimization for credit reservation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":4.086,\"id\":\"worker_a\",\"p95_latency_ms\":63733,\"pass_rate\":0.8595,\"provider_group\":\"provider_x\",\"quality\":0.8996,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.877,\"id\":\"worker_b\",\"p95_latency_ms\":32277,\"pass_rate\":0.3921,\"provider_group\":\"provider_z\",\"quality\":0.7268,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.861,\"id\":\"worker_c\",\"p95_latency_ms\":52266,\"pass_rate\":0.9506,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.947,\"estimated_latency_ms\":115999,\"estimated_pass_rate\":0.8838,\"estimated_quality\":0.9413,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.297,\"max_latency_ms\":206510,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4491, "category": "analysis", "complexity": 0.6328, "consequence": 0.261, "evidence_need": 0.3483, "family": "performance_tuning", "input_tokens": 3640, "request_id": "synthetic-performance_tuning-000635", "reversibility": 0.6774, "risk": 0.3601, "summary": "Review a proposed optimization for credit reservation for hidden regressions."}, "workers": [{"credits_per_call": 4.086, "id": "worker_a", "p95_latency_ms": 63733, "pass_rate": 0.8595, "provider_group": "provider_x", "quality": 0.8996, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.877, "id": "worker_b", "p95_latency_ms": 32277, "pass_rate": 0.3921, "provider_group": "provider_z", "quality": 0.7268, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.861, "id": "worker_c", "p95_latency_ms": 52266, "pass_rate": 0.9506, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.59, "max_latency_ms": 135951, "minimum_quality": 0.7031}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000026", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.87, "estimated_latency_ms": 102822, "estimated_pass_rate": 0.8531, "estimated_quality": 0.8772, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.59, "max_latency_ms": 135951, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8531, "quality": 0.8772, "quality_target_met": true, "utility": 63.565793}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.59,\"max_latency_ms\":135951,\"minimum_quality\":0.7031},\"task\":{\"ambiguity\":0.3151,\"category\":\"analysis\",\"complexity\":0.8841,\"consequence\":0.446,\"evidence_need\":0.7226,\"family\":\"performance_tuning\",\"input_tokens\":7480,\"reversibility\":0.7608,\"risk\":0.5987,\"summary\":\"Choose between throughput and tail-latency changes for streaming inference.\"},\"workers\":[{\"credits_per_call\":1.019,\"id\":\"worker_a\",\"p95_latency_ms\":23969,\"pass_rate\":0.511,\"provider_group\":\"provider_z\",\"quality\":0.7166,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.525,\"id\":\"worker_b\",\"p95_latency_ms\":55889,\"pass_rate\":0.934,\"provider_group\":\"provider_x\",\"quality\":0.8803,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.851,\"id\":\"worker_c\",\"p95_latency_ms\":78853,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.87,\"estimated_latency_ms\":102822,\"estimated_pass_rate\":0.8531,\"estimated_quality\":0.8772,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.59,\"max_latency_ms\":135951,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3151, "category": "analysis", "complexity": 0.8841, "consequence": 0.446, "evidence_need": 0.7226, "family": "performance_tuning", "input_tokens": 7480, "request_id": "synthetic-performance_tuning-000026", "reversibility": 0.7608, "risk": 0.5987, "summary": "Choose between throughput and tail-latency changes for streaming inference."}, "workers": [{"credits_per_call": 1.019, "id": "worker_a", "p95_latency_ms": 23969, "pass_rate": 0.511, "provider_group": "provider_z", "quality": 0.7166, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.525, "id": "worker_b", "p95_latency_ms": 55889, "pass_rate": 0.934, "provider_group": "provider_x", "quality": 0.8803, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.851, "id": "worker_c", "p95_latency_ms": 78853, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.414, "max_latency_ms": 109116, "minimum_quality": 0.8195}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000860", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.923, "estimated_latency_ms": 98020, "estimated_pass_rate": 0.666, "estimated_quality": 0.8328, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.414, "max_latency_ms": 109116, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.666, "quality": 0.8328, "quality_target_met": true, "utility": 55.249698}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.414,\"max_latency_ms\":109116,\"minimum_quality\":0.8195},\"task\":{\"ambiguity\":0.7144,\"category\":\"analysis\",\"complexity\":0.8996,\"consequence\":0.5457,\"evidence_need\":0.5069,\"family\":\"performance_tuning\",\"input_tokens\":6240,\"reversibility\":0.7957,\"risk\":0.417,\"summary\":\"Choose between throughput and tail-latency changes for observability.\"},\"workers\":[{\"credits_per_call\":3.891,\"id\":\"worker_a\",\"p95_latency_ms\":67041,\"pass_rate\":0.7873,\"provider_group\":\"provider_x\",\"quality\":0.9672,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.032,\"id\":\"worker_b\",\"p95_latency_ms\":30979,\"pass_rate\":0.5282,\"provider_group\":\"provider_z\",\"quality\":0.6973,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":13.238,\"id\":\"worker_c\",\"p95_latency_ms\":54948,\"pass_rate\":0.9574,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.923,\"estimated_latency_ms\":98020,\"estimated_pass_rate\":0.666,\"estimated_quality\":0.8328,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.414,\"max_latency_ms\":109116,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7144, "category": "analysis", "complexity": 0.8996, "consequence": 0.5457, "evidence_need": 0.5069, "family": "performance_tuning", "input_tokens": 6240, "request_id": "synthetic-performance_tuning-000860", "reversibility": 0.7957, "risk": 0.417, "summary": "Choose between throughput and tail-latency changes for observability."}, "workers": [{"credits_per_call": 3.891, "id": "worker_a", "p95_latency_ms": 67041, "pass_rate": 0.7873, "provider_group": "provider_x", "quality": 0.9672, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.032, "id": "worker_b", "p95_latency_ms": 30979, "pass_rate": 0.5282, "provider_group": "provider_z", "quality": 0.6973, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 13.238, "id": "worker_c", "p95_latency_ms": 54948, "pass_rate": 0.9574, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.106, "max_latency_ms": 219801, "minimum_quality": 0.8946}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000989", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.379, "estimated_latency_ms": 154703, "estimated_pass_rate": 0.9012, "estimated_quality": 0.9368, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.106, "max_latency_ms": 219801, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9012, "quality": 0.9368, "quality_target_met": true, "utility": 68.431625}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.106,\"max_latency_ms\":219801,\"minimum_quality\":0.8946},\"task\":{\"ambiguity\":0.4852,\"category\":\"analysis\",\"complexity\":0.7168,\"consequence\":0.3348,\"evidence_need\":0.6004,\"family\":\"performance_tuning\",\"input_tokens\":7800,\"reversibility\":0.6145,\"risk\":0.3011,\"summary\":\"Review a proposed optimization for streaming inference for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.038,\"id\":\"worker_a\",\"p95_latency_ms\":70103,\"pass_rate\":0.8289,\"provider_group\":\"provider_z\",\"quality\":0.8789,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.341,\"id\":\"worker_b\",\"p95_latency_ms\":84600,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.91,\"id\":\"worker_c\",\"p95_latency_ms\":28506,\"pass_rate\":0.4673,\"provider_group\":\"provider_x\",\"quality\":0.6745,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.379,\"estimated_latency_ms\":154703,\"estimated_pass_rate\":0.9012,\"estimated_quality\":0.9368,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.106,\"max_latency_ms\":219801,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4852, "category": "analysis", "complexity": 0.7168, "consequence": 0.3348, "evidence_need": 0.6004, "family": "performance_tuning", "input_tokens": 7800, "request_id": "synthetic-performance_tuning-000989", "reversibility": 0.6145, "risk": 0.3011, "summary": "Review a proposed optimization for streaming inference for hidden regressions."}, "workers": [{"credits_per_call": 3.038, "id": "worker_a", "p95_latency_ms": 70103, "pass_rate": 0.8289, "provider_group": "provider_z", "quality": 0.8789, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.341, "id": "worker_b", "p95_latency_ms": 84600, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.91, "id": "worker_c", "p95_latency_ms": 28506, "pass_rate": 0.4673, "provider_group": "provider_x", "quality": 0.6745, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.266, "max_latency_ms": 205190, "minimum_quality": 0.7988}, "family": "migration_planning", "id": "synthetic-migration_planning-000337", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.218, "estimated_latency_ms": 81347, "estimated_pass_rate": 0.5951, "estimated_quality": 0.805, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.266, "max_latency_ms": 205190, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5951, "quality": 0.805, "quality_target_met": true, "utility": 59.207183}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.266,\"max_latency_ms\":205190,\"minimum_quality\":0.7988},\"task\":{\"ambiguity\":0.5699,\"category\":\"coding\",\"complexity\":0.6862,\"consequence\":0.7113,\"evidence_need\":0.7236,\"family\":\"migration_planning\",\"input_tokens\":3240,\"reversibility\":0.3576,\"risk\":0.6022,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.011,\"id\":\"worker_a\",\"p95_latency_ms\":51619,\"pass_rate\":0.8335,\"provider_group\":\"provider_y\",\"quality\":0.9144,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.401,\"id\":\"worker_b\",\"p95_latency_ms\":53457,\"pass_rate\":0.9302,\"provider_group\":\"provider_x\",\"quality\":0.9767,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.207,\"id\":\"worker_c\",\"p95_latency_ms\":29728,\"pass_rate\":0.4593,\"provider_group\":\"provider_z\",\"quality\":0.7026,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.218,\"estimated_latency_ms\":81347,\"estimated_pass_rate\":0.5951,\"estimated_quality\":0.805,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.266,\"max_latency_ms\":205190,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5699, "category": "coding", "complexity": 0.6862, "consequence": 0.7113, "evidence_need": 0.7236, "family": "migration_planning", "input_tokens": 3240, "request_id": "synthetic-migration_planning-000337", "reversibility": 0.3576, "risk": 0.6022, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 3.011, "id": "worker_a", "p95_latency_ms": 51619, "pass_rate": 0.8335, "provider_group": "provider_y", "quality": 0.9144, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.401, "id": "worker_b", "p95_latency_ms": 53457, "pass_rate": 0.9302, "provider_group": "provider_x", "quality": 0.9767, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.207, "id": "worker_c", "p95_latency_ms": 29728, "pass_rate": 0.4593, "provider_group": "provider_z", "quality": 0.7026, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.923, "max_latency_ms": 37831, "minimum_quality": 0.6988}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000185", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.148, "estimated_latency_ms": 31852, "estimated_pass_rate": 0.388, "estimated_quality": 0.6038, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.923, "max_latency_ms": 37831, "policy": "lean", "reason_codes": ["best_feasible_below_target", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.388, "quality": 0.6038, "quality_target_met": false, "utility": 34.401915}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.923,\"max_latency_ms\":37831,\"minimum_quality\":0.6988},\"task\":{\"ambiguity\":0.396,\"category\":\"analysis\",\"complexity\":0.5065,\"consequence\":0.5008,\"evidence_need\":0.6228,\"family\":\"performance_tuning\",\"input_tokens\":3520,\"reversibility\":0.879,\"risk\":0.5455,\"summary\":\"Choose between throughput and tail-latency changes for usage metering.\"},\"workers\":[{\"credits_per_call\":10.412,\"id\":\"worker_a\",\"p95_latency_ms\":85511,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.801,\"id\":\"worker_b\",\"p95_latency_ms\":67325,\"pass_rate\":0.8652,\"provider_group\":\"provider_y\",\"quality\":0.9076,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.148,\"id\":\"worker_c\",\"p95_latency_ms\":31852,\"pass_rate\":0.5196,\"provider_group\":\"provider_z\",\"quality\":0.6813,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.148,\"estimated_latency_ms\":31852,\"estimated_pass_rate\":0.388,\"estimated_quality\":0.6038,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.923,\"max_latency_ms\":37831,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.396, "category": "analysis", "complexity": 0.5065, "consequence": 0.5008, "evidence_need": 0.6228, "family": "performance_tuning", "input_tokens": 3520, "request_id": "synthetic-performance_tuning-000185", "reversibility": 0.879, "risk": 0.5455, "summary": "Choose between throughput and tail-latency changes for usage metering."}, "workers": [{"credits_per_call": 10.412, "id": "worker_a", "p95_latency_ms": 85511, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.801, "id": "worker_b", "p95_latency_ms": 67325, "pass_rate": 0.8652, "provider_group": "provider_y", "quality": 0.9076, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.148, "id": "worker_c", "p95_latency_ms": 31852, "pass_rate": 0.5196, "provider_group": "provider_z", "quality": 0.6813, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.856, "max_latency_ms": 215031, "minimum_quality": 0.9163}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001050", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.861, "estimated_latency_ms": 193985, "estimated_pass_rate": 0.7431, "estimated_quality": 0.9264, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.856, "max_latency_ms": 215031, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7431, "quality": 0.9264, "quality_target_met": true, "utility": 75.102242}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.856,\"max_latency_ms\":215031,\"minimum_quality\":0.9163},\"task\":{\"ambiguity\":0.7474,\"category\":\"review\",\"complexity\":0.562,\"consequence\":0.9148,\"evidence_need\":0.6059,\"family\":\"policy_compliance\",\"input_tokens\":5640,\"reversibility\":0.5336,\"risk\":0.5576,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":11.831,\"id\":\"worker_a\",\"p95_latency_ms\":82764,\"pass_rate\":0.9185,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.856,\"id\":\"worker_b\",\"p95_latency_ms\":81069,\"pass_rate\":0.7898,\"provider_group\":\"provider_x\",\"quality\":0.922,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.174,\"id\":\"worker_c\",\"p95_latency_ms\":30152,\"pass_rate\":0.5587,\"provider_group\":\"provider_y\",\"quality\":0.7796,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.861,\"estimated_latency_ms\":193985,\"estimated_pass_rate\":0.7431,\"estimated_quality\":0.9264,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.856,\"max_latency_ms\":215031,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7474, "category": "review", "complexity": 0.562, "consequence": 0.9148, "evidence_need": 0.6059, "family": "policy_compliance", "input_tokens": 5640, "request_id": "synthetic-policy_compliance-001050", "reversibility": 0.5336, "risk": 0.5576, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 11.831, "id": "worker_a", "p95_latency_ms": 82764, "pass_rate": 0.9185, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.856, "id": "worker_b", "p95_latency_ms": 81069, "pass_rate": 0.7898, "provider_group": "provider_x", "quality": 0.922, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.174, "id": "worker_c", "p95_latency_ms": 30152, "pass_rate": 0.5587, "provider_group": "provider_y", "quality": 0.7796, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.289, "max_latency_ms": 46646, "minimum_quality": 0.6337}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000723", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.88, "estimated_latency_ms": 26020, "estimated_pass_rate": 0.3203, "estimated_quality": 0.6661, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.289, "max_latency_ms": 46646, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3203, "quality": 0.6661, "quality_target_met": true, "utility": 43.463785}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.289,\"max_latency_ms\":46646,\"minimum_quality\":0.6337},\"task\":{\"ambiguity\":0.4412,\"category\":\"review\",\"complexity\":0.6346,\"consequence\":0.5085,\"evidence_need\":0.8965,\"family\":\"policy_compliance\",\"input_tokens\":7040,\"reversibility\":0.5791,\"risk\":0.7922,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.948,\"id\":\"worker_a\",\"p95_latency_ms\":58152,\"pass_rate\":0.752,\"provider_group\":\"provider_z\",\"quality\":0.9032,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.88,\"id\":\"worker_b\",\"p95_latency_ms\":26020,\"pass_rate\":0.4898,\"provider_group\":\"provider_x\",\"quality\":0.7875,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.131,\"id\":\"worker_c\",\"p95_latency_ms\":64531,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9366,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.88,\"estimated_latency_ms\":26020,\"estimated_pass_rate\":0.3203,\"estimated_quality\":0.6661,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.289,\"max_latency_ms\":46646,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4412, "category": "review", "complexity": 0.6346, "consequence": 0.5085, "evidence_need": 0.8965, "family": "policy_compliance", "input_tokens": 7040, "request_id": "synthetic-policy_compliance-000723", "reversibility": 0.5791, "risk": 0.7922, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.948, "id": "worker_a", "p95_latency_ms": 58152, "pass_rate": 0.752, "provider_group": "provider_z", "quality": 0.9032, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.88, "id": "worker_b", "p95_latency_ms": 26020, "pass_rate": 0.4898, "provider_group": "provider_x", "quality": 0.7875, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.131, "id": "worker_c", "p95_latency_ms": 64531, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9366, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.566, "max_latency_ms": 116729, "minimum_quality": 0.8157}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000711", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.961, "estimated_latency_ms": 103107, "estimated_pass_rate": 0.7238, "estimated_quality": 0.8855, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.566, "max_latency_ms": 116729, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7238, "quality": 0.8855, "quality_target_met": true, "utility": 68.100801}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.566,\"max_latency_ms\":116729,\"minimum_quality\":0.8157},\"task\":{\"ambiguity\":0.5489,\"category\":\"review\",\"complexity\":0.7355,\"consequence\":0.8692,\"evidence_need\":0.7229,\"family\":\"policy_compliance\",\"input_tokens\":2960,\"reversibility\":0.4417,\"risk\":0.7895,\"summary\":\"Identify evidence required before approving background jobs for production.\"},\"workers\":[{\"credits_per_call\":14.855,\"id\":\"worker_a\",\"p95_latency_ms\":85234,\"pass_rate\":0.9433,\"provider_group\":\"provider_z\",\"quality\":0.974,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.174,\"id\":\"worker_b\",\"p95_latency_ms\":53564,\"pass_rate\":0.8269,\"provider_group\":\"provider_x\",\"quality\":0.8586,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.106,\"id\":\"worker_c\",\"p95_latency_ms\":17873,\"pass_rate\":0.5354,\"provider_group\":\"provider_y\",\"quality\":0.7809,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.961,\"estimated_latency_ms\":103107,\"estimated_pass_rate\":0.7238,\"estimated_quality\":0.8855,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.566,\"max_latency_ms\":116729,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5489, "category": "review", "complexity": 0.7355, "consequence": 0.8692, "evidence_need": 0.7229, "family": "policy_compliance", "input_tokens": 2960, "request_id": "synthetic-policy_compliance-000711", "reversibility": 0.4417, "risk": 0.7895, "summary": "Identify evidence required before approving background jobs for production."}, "workers": [{"credits_per_call": 14.855, "id": "worker_a", "p95_latency_ms": 85234, "pass_rate": 0.9433, "provider_group": "provider_z", "quality": 0.974, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.174, "id": "worker_b", "p95_latency_ms": 53564, "pass_rate": 0.8269, "provider_group": "provider_x", "quality": 0.8586, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.106, "id": "worker_c", "p95_latency_ms": 17873, "pass_rate": 0.5354, "provider_group": "provider_y", "quality": 0.7809, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 6.803, "max_latency_ms": 194041, "minimum_quality": 0.7371}, "family": "migration_planning", "id": "synthetic-migration_planning-001117", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.052, "estimated_latency_ms": 86242, "estimated_pass_rate": 0.7644, "estimated_quality": 0.7753, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.803, "max_latency_ms": 194041, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7644, "quality": 0.7753, "quality_target_met": true, "utility": 56.928874}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":6.803,\"max_latency_ms\":194041,\"minimum_quality\":0.7371},\"task\":{\"ambiguity\":0.5203,\"category\":\"coding\",\"complexity\":0.8533,\"consequence\":0.6371,\"evidence_need\":0.4432,\"family\":\"migration_planning\",\"input_tokens\":5840,\"reversibility\":0.3972,\"risk\":0.4904,\"summary\":\"Review a staged migration for schema evolution and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.077,\"id\":\"worker_a\",\"p95_latency_ms\":24290,\"pass_rate\":0.5523,\"provider_group\":\"provider_y\",\"quality\":0.7274,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.015,\"id\":\"worker_b\",\"p95_latency_ms\":58231,\"pass_rate\":0.9759,\"provider_group\":\"provider_x\",\"quality\":0.913,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.975,\"id\":\"worker_c\",\"p95_latency_ms\":61952,\"pass_rate\":0.8836,\"provider_group\":\"provider_z\",\"quality\":0.8654,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.052,\"estimated_latency_ms\":86242,\"estimated_pass_rate\":0.7644,\"estimated_quality\":0.7753,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.803,\"max_latency_ms\":194041,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5203, "category": "coding", "complexity": 0.8533, "consequence": 0.6371, "evidence_need": 0.4432, "family": "migration_planning", "input_tokens": 5840, "request_id": "synthetic-migration_planning-001117", "reversibility": 0.3972, "risk": 0.4904, "summary": "Review a staged migration for schema evolution and identify irreversible steps."}, "workers": [{"credits_per_call": 1.077, "id": "worker_a", "p95_latency_ms": 24290, "pass_rate": 0.5523, "provider_group": "provider_y", "quality": 0.7274, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.015, "id": "worker_b", "p95_latency_ms": 58231, "pass_rate": 0.9759, "provider_group": "provider_x", "quality": 0.913, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.975, "id": "worker_c", "p95_latency_ms": 61952, "pass_rate": 0.8836, "provider_group": "provider_z", "quality": 0.8654, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 7.289, "max_latency_ms": 195049, "minimum_quality": 0.8815}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000924", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.138, "estimated_latency_ms": 74655, "estimated_pass_rate": 0.6536, "estimated_quality": 0.8186, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.289, "max_latency_ms": 195049, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6536, "quality": 0.8186, "quality_target_met": false, "utility": 55.475553}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":7.289,\"max_latency_ms\":195049,\"minimum_quality\":0.8815},\"task\":{\"ambiguity\":0.445,\"category\":\"review\",\"complexity\":0.7786,\"consequence\":0.8548,\"evidence_need\":0.574,\"family\":\"policy_compliance\",\"input_tokens\":2320,\"reversibility\":0.3922,\"risk\":0.6511,\"summary\":\"Audit a workflow for background jobs against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.054,\"id\":\"worker_a\",\"p95_latency_ms\":19780,\"pass_rate\":0.5052,\"provider_group\":\"provider_z\",\"quality\":0.7192,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.084,\"id\":\"worker_b\",\"p95_latency_ms\":54875,\"pass_rate\":0.7413,\"provider_group\":\"provider_y\",\"quality\":0.915,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.17,\"id\":\"worker_c\",\"p95_latency_ms\":66368,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9214,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.138,\"estimated_latency_ms\":74655,\"estimated_pass_rate\":0.6536,\"estimated_quality\":0.8186,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.289,\"max_latency_ms\":195049,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.445, "category": "review", "complexity": 0.7786, "consequence": 0.8548, "evidence_need": 0.574, "family": "policy_compliance", "input_tokens": 2320, "request_id": "synthetic-policy_compliance-000924", "reversibility": 0.3922, "risk": 0.6511, "summary": "Audit a workflow for background jobs against explicit access requirements."}, "workers": [{"credits_per_call": 1.054, "id": "worker_a", "p95_latency_ms": 19780, "pass_rate": 0.5052, "provider_group": "provider_z", "quality": 0.7192, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.084, "id": "worker_b", "p95_latency_ms": 54875, "pass_rate": 0.7413, "provider_group": "provider_y", "quality": 0.915, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.17, "id": "worker_c", "p95_latency_ms": 66368, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9214, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.445, "max_latency_ms": 198052, "minimum_quality": 0.8097}, "family": "migration_planning", "id": "synthetic-migration_planning-000970", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.481, "estimated_latency_ms": 173641, "estimated_pass_rate": 0.7164, "estimated_quality": 0.8985, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.445, "max_latency_ms": 198052, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7164, "quality": 0.8985, "quality_target_met": true, "utility": 68.5197}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.445,\"max_latency_ms\":198052,\"minimum_quality\":0.8097},\"task\":{\"ambiguity\":0.4729,\"category\":\"coding\",\"complexity\":0.7579,\"consequence\":0.4571,\"evidence_need\":0.4338,\"family\":\"migration_planning\",\"input_tokens\":4560,\"reversibility\":0.5477,\"risk\":0.575,\"summary\":\"Plan a zero-downtime migration of prompt caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.842,\"id\":\"worker_a\",\"p95_latency_ms\":91551,\"pass_rate\":0.891,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.848,\"id\":\"worker_b\",\"p95_latency_ms\":28381,\"pass_rate\":0.5199,\"provider_group\":\"provider_y\",\"quality\":0.6965,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.791,\"id\":\"worker_c\",\"p95_latency_ms\":53709,\"pass_rate\":0.8552,\"provider_group\":\"provider_x\",\"quality\":0.8386,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.481,\"estimated_latency_ms\":173641,\"estimated_pass_rate\":0.7164,\"estimated_quality\":0.8985,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.445,\"max_latency_ms\":198052,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4729, "category": "coding", "complexity": 0.7579, "consequence": 0.4571, "evidence_need": 0.4338, "family": "migration_planning", "input_tokens": 4560, "request_id": "synthetic-migration_planning-000970", "reversibility": 0.5477, "risk": 0.575, "summary": "Plan a zero-downtime migration of prompt caching with rollback checkpoints."}, "workers": [{"credits_per_call": 12.842, "id": "worker_a", "p95_latency_ms": 91551, "pass_rate": 0.891, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.848, "id": "worker_b", "p95_latency_ms": 28381, "pass_rate": 0.5199, "provider_group": "provider_y", "quality": 0.6965, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.791, "id": "worker_c", "p95_latency_ms": 53709, "pass_rate": 0.8552, "provider_group": "provider_x", "quality": 0.8386, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 4.968, "max_latency_ms": 175381, "minimum_quality": 0.8035}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000366", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.106, "estimated_latency_ms": 64280, "estimated_pass_rate": 0.5633, "estimated_quality": 0.7813, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.968, "max_latency_ms": 175381, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5633, "quality": 0.7813, "quality_target_met": false, "utility": 56.361174}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":4.968,\"max_latency_ms\":175381,\"minimum_quality\":0.8035},\"task\":{\"ambiguity\":0.4959,\"category\":\"review\",\"complexity\":0.7615,\"consequence\":0.8763,\"evidence_need\":0.7878,\"family\":\"policy_compliance\",\"input_tokens\":3360,\"reversibility\":0.3072,\"risk\":0.508,\"summary\":\"Identify evidence required before approving artifact storage for production.\"},\"workers\":[{\"credits_per_call\":15.061,\"id\":\"worker_a\",\"p95_latency_ms\":49734,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9543,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.154,\"id\":\"worker_b\",\"p95_latency_ms\":45992,\"pass_rate\":0.7312,\"provider_group\":\"provider_x\",\"quality\":0.8473,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.952,\"id\":\"worker_c\",\"p95_latency_ms\":18288,\"pass_rate\":0.5144,\"provider_group\":\"provider_z\",\"quality\":0.74,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.106,\"estimated_latency_ms\":64280,\"estimated_pass_rate\":0.5633,\"estimated_quality\":0.7813,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.968,\"max_latency_ms\":175381,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4959, "category": "review", "complexity": 0.7615, "consequence": 0.8763, "evidence_need": 0.7878, "family": "policy_compliance", "input_tokens": 3360, "request_id": "synthetic-policy_compliance-000366", "reversibility": 0.3072, "risk": 0.508, "summary": "Identify evidence required before approving artifact storage for production."}, "workers": [{"credits_per_call": 15.061, "id": "worker_a", "p95_latency_ms": 49734, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9543, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.154, "id": "worker_b", "p95_latency_ms": 45992, "pass_rate": 0.7312, "provider_group": "provider_x", "quality": 0.8473, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.952, "id": "worker_c", "p95_latency_ms": 18288, "pass_rate": 0.5144, "provider_group": "provider_z", "quality": 0.74, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.295, "max_latency_ms": 91242, "minimum_quality": 0.8148}, "family": "migration_planning", "id": "synthetic-migration_planning-000457", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.046, "estimated_latency_ms": 75567, "estimated_pass_rate": 0.6891, "estimated_quality": 0.8405, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.295, "max_latency_ms": 91242, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6891, "quality": 0.8405, "quality_target_met": true, "utility": 59.118638}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.295,\"max_latency_ms\":91242,\"minimum_quality\":0.8148},\"task\":{\"ambiguity\":0.3039,\"category\":\"coding\",\"complexity\":0.5617,\"consequence\":0.8286,\"evidence_need\":0.5138,\"family\":\"migration_planning\",\"input_tokens\":4600,\"reversibility\":0.3927,\"risk\":0.3563,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.833,\"id\":\"worker_a\",\"p95_latency_ms\":55159,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9375,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.086,\"id\":\"worker_b\",\"p95_latency_ms\":21823,\"pass_rate\":0.6106,\"provider_group\":\"provider_z\",\"quality\":0.7187,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.96,\"id\":\"worker_c\",\"p95_latency_ms\":53744,\"pass_rate\":0.7259,\"provider_group\":\"provider_y\",\"quality\":0.8658,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.046,\"estimated_latency_ms\":75567,\"estimated_pass_rate\":0.6891,\"estimated_quality\":0.8405,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.295,\"max_latency_ms\":91242,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3039, "category": "coding", "complexity": 0.5617, "consequence": 0.8286, "evidence_need": 0.5138, "family": "migration_planning", "input_tokens": 4600, "request_id": "synthetic-migration_planning-000457", "reversibility": 0.3927, "risk": 0.3563, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 14.833, "id": "worker_a", "p95_latency_ms": 55159, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9375, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.086, "id": "worker_b", "p95_latency_ms": 21823, "pass_rate": 0.6106, "provider_group": "provider_z", "quality": 0.7187, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.96, "id": "worker_c", "p95_latency_ms": 53744, "pass_rate": 0.7259, "provider_group": "provider_y", "quality": 0.8658, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.242, "max_latency_ms": 142790, "minimum_quality": 0.7813}, "family": "migration_planning", "id": "synthetic-migration_planning-000232", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.833, "estimated_latency_ms": 139559, "estimated_pass_rate": 0.7225, "estimated_quality": 0.8992, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.242, "max_latency_ms": 142790, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7225, "quality": 0.8992, "quality_target_met": true, "utility": 72.379517}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.242,\"max_latency_ms\":142790,\"minimum_quality\":0.7813},\"task\":{\"ambiguity\":0.7295,\"category\":\"coding\",\"complexity\":0.766,\"consequence\":0.79,\"evidence_need\":0.5676,\"family\":\"migration_planning\",\"input_tokens\":5560,\"reversibility\":0.5082,\"risk\":0.4393,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":12.594,\"id\":\"worker_a\",\"p95_latency_ms\":57076,\"pass_rate\":0.9515,\"provider_group\":\"provider_z\",\"quality\":0.9612,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.865,\"id\":\"worker_b\",\"p95_latency_ms\":22847,\"pass_rate\":0.5684,\"provider_group\":\"provider_x\",\"quality\":0.786,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.374,\"id\":\"worker_c\",\"p95_latency_ms\":59636,\"pass_rate\":0.751,\"provider_group\":\"provider_y\",\"quality\":0.8258,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.833,\"estimated_latency_ms\":139559,\"estimated_pass_rate\":0.7225,\"estimated_quality\":0.8992,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.242,\"max_latency_ms\":142790,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7295, "category": "coding", "complexity": 0.766, "consequence": 0.79, "evidence_need": 0.5676, "family": "migration_planning", "input_tokens": 5560, "request_id": "synthetic-migration_planning-000232", "reversibility": 0.5082, "risk": 0.4393, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 12.594, "id": "worker_a", "p95_latency_ms": 57076, "pass_rate": 0.9515, "provider_group": "provider_z", "quality": 0.9612, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.865, "id": "worker_b", "p95_latency_ms": 22847, "pass_rate": 0.5684, "provider_group": "provider_x", "quality": 0.786, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.374, "id": "worker_c", "p95_latency_ms": 59636, "pass_rate": 0.751, "provider_group": "provider_y", "quality": 0.8258, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.27, "max_latency_ms": 147316, "minimum_quality": 0.7711}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000230", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.506, "estimated_latency_ms": 143851, "estimated_pass_rate": 0.8955, "estimated_quality": 0.921, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.27, "max_latency_ms": 147316, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8955, "quality": 0.921, "quality_target_met": true, "utility": 67.620121}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.27,\"max_latency_ms\":147316,\"minimum_quality\":0.7711},\"task\":{\"ambiguity\":0.5148,\"category\":\"analysis\",\"complexity\":0.662,\"consequence\":0.3105,\"evidence_need\":0.3522,\"family\":\"performance_tuning\",\"input_tokens\":5640,\"reversibility\":0.6843,\"risk\":0.5974,\"summary\":\"Diagnose latency variance in private credentials and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":14.299,\"id\":\"worker_a\",\"p95_latency_ms\":77473,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9931,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.207,\"id\":\"worker_b\",\"p95_latency_ms\":66378,\"pass_rate\":0.7871,\"provider_group\":\"provider_x\",\"quality\":0.867,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.925,\"id\":\"worker_c\",\"p95_latency_ms\":17845,\"pass_rate\":0.4351,\"provider_group\":\"provider_y\",\"quality\":0.6893,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.506,\"estimated_latency_ms\":143851,\"estimated_pass_rate\":0.8955,\"estimated_quality\":0.921,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.27,\"max_latency_ms\":147316,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5148, "category": "analysis", "complexity": 0.662, "consequence": 0.3105, "evidence_need": 0.3522, "family": "performance_tuning", "input_tokens": 5640, "request_id": "synthetic-performance_tuning-000230", "reversibility": 0.6843, "risk": 0.5974, "summary": "Diagnose latency variance in private credentials and rank the next experiments."}, "workers": [{"credits_per_call": 14.299, "id": "worker_a", "p95_latency_ms": 77473, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9931, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.207, "id": "worker_b", "p95_latency_ms": 66378, "pass_rate": 0.7871, "provider_group": "provider_x", "quality": 0.867, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.925, "id": "worker_c", "p95_latency_ms": 17845, "pass_rate": 0.4351, "provider_group": "provider_y", "quality": 0.6893, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.644, "max_latency_ms": 207479, "minimum_quality": 0.9189}, "family": "migration_planning", "id": "synthetic-migration_planning-000088", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.793, "estimated_latency_ms": 147817, "estimated_pass_rate": 0.7794, "estimated_quality": 0.904, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.644, "max_latency_ms": 207479, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7794, "quality": 0.904, "quality_target_met": false, "utility": 71.403772}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.644,\"max_latency_ms\":207479,\"minimum_quality\":0.9189},\"task\":{\"ambiguity\":0.3909,\"category\":\"coding\",\"complexity\":0.7438,\"consequence\":0.498,\"evidence_need\":0.4081,\"family\":\"migration_planning\",\"input_tokens\":2840,\"reversibility\":0.3731,\"risk\":0.7022,\"summary\":\"Review a staged migration for tenant-scoped caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":15.034,\"id\":\"worker_a\",\"p95_latency_ms\":71741,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.663,\"id\":\"worker_b\",\"p95_latency_ms\":53473,\"pass_rate\":0.8132,\"provider_group\":\"provider_x\",\"quality\":0.8141,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.096,\"id\":\"worker_c\",\"p95_latency_ms\":22603,\"pass_rate\":0.5876,\"provider_group\":\"provider_z\",\"quality\":0.723,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.793,\"estimated_latency_ms\":147817,\"estimated_pass_rate\":0.7794,\"estimated_quality\":0.904,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.644,\"max_latency_ms\":207479,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3909, "category": "coding", "complexity": 0.7438, "consequence": 0.498, "evidence_need": 0.4081, "family": "migration_planning", "input_tokens": 2840, "request_id": "synthetic-migration_planning-000088", "reversibility": 0.3731, "risk": 0.7022, "summary": "Review a staged migration for tenant-scoped caching and identify irreversible steps."}, "workers": [{"credits_per_call": 15.034, "id": "worker_a", "p95_latency_ms": 71741, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.663, "id": "worker_b", "p95_latency_ms": 53473, "pass_rate": 0.8132, "provider_group": "provider_x", "quality": 0.8141, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.096, "id": "worker_c", "p95_latency_ms": 22603, "pass_rate": 0.5876, "provider_group": "provider_z", "quality": 0.723, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.772, "max_latency_ms": 209826, "minimum_quality": 0.8586}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000054", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.067, "estimated_latency_ms": 170619, "estimated_pass_rate": 0.7704, "estimated_quality": 0.8746, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.772, "max_latency_ms": 209826, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7704, "quality": 0.8746, "quality_target_met": true, "utility": 76.04991}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.772,\"max_latency_ms\":209826,\"minimum_quality\":0.8586},\"task\":{\"ambiguity\":0.749,\"category\":\"review\",\"complexity\":0.8044,\"consequence\":0.5982,\"evidence_need\":0.6604,\"family\":\"policy_compliance\",\"input_tokens\":7280,\"reversibility\":0.4226,\"risk\":0.774,\"summary\":\"Audit a workflow for webhook delivery against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.896,\"id\":\"worker_a\",\"p95_latency_ms\":77504,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9361,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.168,\"id\":\"worker_b\",\"p95_latency_ms\":71504,\"pass_rate\":0.7335,\"provider_group\":\"provider_z\",\"quality\":0.8784,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.003,\"id\":\"worker_c\",\"p95_latency_ms\":21611,\"pass_rate\":0.5985,\"provider_group\":\"provider_x\",\"quality\":0.6995,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.067,\"estimated_latency_ms\":170619,\"estimated_pass_rate\":0.7704,\"estimated_quality\":0.8746,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.772,\"max_latency_ms\":209826,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.749, "category": "review", "complexity": 0.8044, "consequence": 0.5982, "evidence_need": 0.6604, "family": "policy_compliance", "input_tokens": 7280, "request_id": "synthetic-policy_compliance-000054", "reversibility": 0.4226, "risk": 0.774, "summary": "Audit a workflow for webhook delivery against explicit access requirements."}, "workers": [{"credits_per_call": 11.896, "id": "worker_a", "p95_latency_ms": 77504, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9361, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.168, "id": "worker_b", "p95_latency_ms": 71504, "pass_rate": 0.7335, "provider_group": "provider_z", "quality": 0.8784, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.003, "id": "worker_c", "p95_latency_ms": 21611, "pass_rate": 0.5985, "provider_group": "provider_x", "quality": 0.6995, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.314, "max_latency_ms": 86800, "minimum_quality": 0.8209}, "family": "migration_planning", "id": "synthetic-migration_planning-001225", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.83, "estimated_latency_ms": 74862, "estimated_pass_rate": 0.5856, "estimated_quality": 0.7721, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.314, "max_latency_ms": 86800, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5856, "quality": 0.7721, "quality_target_met": false, "utility": 52.790324}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.314,\"max_latency_ms\":86800,\"minimum_quality\":0.8209},\"task\":{\"ambiguity\":0.4856,\"category\":\"coding\",\"complexity\":0.8653,\"consequence\":0.4703,\"evidence_need\":0.841,\"family\":\"migration_planning\",\"input_tokens\":5520,\"reversibility\":0.4821,\"risk\":0.4862,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":11.714,\"id\":\"worker_a\",\"p95_latency_ms\":73623,\"pass_rate\":0.938,\"provider_group\":\"provider_y\",\"quality\":0.9574,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.973,\"id\":\"worker_b\",\"p95_latency_ms\":50970,\"pass_rate\":0.8017,\"provider_group\":\"provider_x\",\"quality\":0.831,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.857,\"id\":\"worker_c\",\"p95_latency_ms\":23892,\"pass_rate\":0.5379,\"provider_group\":\"provider_z\",\"quality\":0.7037,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.83,\"estimated_latency_ms\":74862,\"estimated_pass_rate\":0.5856,\"estimated_quality\":0.7721,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.314,\"max_latency_ms\":86800,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4856, "category": "coding", "complexity": 0.8653, "consequence": 0.4703, "evidence_need": 0.841, "family": "migration_planning", "input_tokens": 5520, "request_id": "synthetic-migration_planning-001225", "reversibility": 0.4821, "risk": 0.4862, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 11.714, "id": "worker_a", "p95_latency_ms": 73623, "pass_rate": 0.938, "provider_group": "provider_y", "quality": 0.9574, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.973, "id": "worker_b", "p95_latency_ms": 50970, "pass_rate": 0.8017, "provider_group": "provider_x", "quality": 0.831, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.857, "id": "worker_c", "p95_latency_ms": 23892, "pass_rate": 0.5379, "provider_group": "provider_z", "quality": 0.7037, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.841, "max_latency_ms": 228923, "minimum_quality": 0.8481}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001220", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.505, "estimated_latency_ms": 91180, "estimated_pass_rate": 0.7919, "estimated_quality": 0.8605, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.841, "max_latency_ms": 228923, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7919, "quality": 0.8605, "quality_target_met": true, "utility": 64.287855}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.841,\"max_latency_ms\":228923,\"minimum_quality\":0.8481},\"task\":{\"ambiguity\":0.5158,\"category\":\"analysis\",\"complexity\":0.8033,\"consequence\":0.6301,\"evidence_need\":0.4437,\"family\":\"performance_tuning\",\"input_tokens\":1320,\"reversibility\":0.5518,\"risk\":0.5343,\"summary\":\"Diagnose latency variance in observability and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.614,\"id\":\"worker_a\",\"p95_latency_ms\":64792,\"pass_rate\":0.9198,\"provider_group\":\"provider_y\",\"quality\":0.9763,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.891,\"id\":\"worker_b\",\"p95_latency_ms\":26388,\"pass_rate\":0.5054,\"provider_group\":\"provider_x\",\"quality\":0.7389,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.908,\"id\":\"worker_c\",\"p95_latency_ms\":76161,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.505,\"estimated_latency_ms\":91180,\"estimated_pass_rate\":0.7919,\"estimated_quality\":0.8605,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.841,\"max_latency_ms\":228923,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5158, "category": "analysis", "complexity": 0.8033, "consequence": 0.6301, "evidence_need": 0.4437, "family": "performance_tuning", "input_tokens": 1320, "request_id": "synthetic-performance_tuning-001220", "reversibility": 0.5518, "risk": 0.5343, "summary": "Diagnose latency variance in observability and rank the next experiments."}, "workers": [{"credits_per_call": 3.614, "id": "worker_a", "p95_latency_ms": 64792, "pass_rate": 0.9198, "provider_group": "provider_y", "quality": 0.9763, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.891, "id": "worker_b", "p95_latency_ms": 26388, "pass_rate": 0.5054, "provider_group": "provider_x", "quality": 0.7389, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.908, "id": "worker_c", "p95_latency_ms": 76161, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.368, "max_latency_ms": 189962, "minimum_quality": 0.807}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000201", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.387, "estimated_latency_ms": 149854, "estimated_pass_rate": 0.7839, "estimated_quality": 0.9273, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.368, "max_latency_ms": 189962, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7839, "quality": 0.9273, "quality_target_met": true, "utility": 75.422537}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.368,\"max_latency_ms\":189962,\"minimum_quality\":0.807},\"task\":{\"ambiguity\":0.6721,\"category\":\"review\",\"complexity\":0.4819,\"consequence\":0.5101,\"evidence_need\":0.7849,\"family\":\"policy_compliance\",\"input_tokens\":5240,\"reversibility\":0.4809,\"risk\":0.6832,\"summary\":\"Review whether OAuth callbacks satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.774,\"id\":\"worker_a\",\"p95_latency_ms\":80767,\"pass_rate\":0.8531,\"provider_group\":\"provider_z\",\"quality\":0.8036,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.842,\"id\":\"worker_b\",\"p95_latency_ms\":18272,\"pass_rate\":0.5804,\"provider_group\":\"provider_y\",\"quality\":0.7445,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.771,\"id\":\"worker_c\",\"p95_latency_ms\":50815,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.387,\"estimated_latency_ms\":149854,\"estimated_pass_rate\":0.7839,\"estimated_quality\":0.9273,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.368,\"max_latency_ms\":189962,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6721, "category": "review", "complexity": 0.4819, "consequence": 0.5101, "evidence_need": 0.7849, "family": "policy_compliance", "input_tokens": 5240, "request_id": "synthetic-policy_compliance-000201", "reversibility": 0.4809, "risk": 0.6832, "summary": "Review whether OAuth callbacks satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.774, "id": "worker_a", "p95_latency_ms": 80767, "pass_rate": 0.8531, "provider_group": "provider_z", "quality": 0.8036, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.842, "id": "worker_b", "p95_latency_ms": 18272, "pass_rate": 0.5804, "provider_group": "provider_y", "quality": 0.7445, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.771, "id": "worker_c", "p95_latency_ms": 50815, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.119, "max_latency_ms": 179864, "minimum_quality": 0.712}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001262", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.686, "estimated_latency_ms": 94070, "estimated_pass_rate": 0.802, "estimated_quality": 0.8307, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.119, "max_latency_ms": 179864, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.802, "quality": 0.8307, "quality_target_met": true, "utility": 63.180934}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.119,\"max_latency_ms\":179864,\"minimum_quality\":0.712},\"task\":{\"ambiguity\":0.3649,\"category\":\"analysis\",\"complexity\":0.7662,\"consequence\":0.5527,\"evidence_need\":0.3487,\"family\":\"performance_tuning\",\"input_tokens\":6800,\"reversibility\":0.6235,\"risk\":0.5074,\"summary\":\"Review a proposed optimization for model routing for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.793,\"id\":\"worker_a\",\"p95_latency_ms\":73704,\"pass_rate\":0.9231,\"provider_group\":\"provider_x\",\"quality\":0.9301,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.893,\"id\":\"worker_b\",\"p95_latency_ms\":20366,\"pass_rate\":0.4956,\"provider_group\":\"provider_y\",\"quality\":0.6741,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":12.762,\"id\":\"worker_c\",\"p95_latency_ms\":84749,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.686,\"estimated_latency_ms\":94070,\"estimated_pass_rate\":0.802,\"estimated_quality\":0.8307,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.119,\"max_latency_ms\":179864,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3649, "category": "analysis", "complexity": 0.7662, "consequence": 0.5527, "evidence_need": 0.3487, "family": "performance_tuning", "input_tokens": 6800, "request_id": "synthetic-performance_tuning-001262", "reversibility": 0.6235, "risk": 0.5074, "summary": "Review a proposed optimization for model routing for hidden regressions."}, "workers": [{"credits_per_call": 3.793, "id": "worker_a", "p95_latency_ms": 73704, "pass_rate": 0.9231, "provider_group": "provider_x", "quality": 0.9301, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.893, "id": "worker_b", "p95_latency_ms": 20366, "pass_rate": 0.4956, "provider_group": "provider_y", "quality": 0.6741, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 12.762, "id": "worker_c", "p95_latency_ms": 84749, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.644, "max_latency_ms": 163746, "minimum_quality": 0.8114}, "family": "migration_planning", "id": "synthetic-migration_planning-000745", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.605, "estimated_latency_ms": 121296, "estimated_pass_rate": 0.9126, "estimated_quality": 0.9254, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.644, "max_latency_ms": 163746, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9126, "quality": 0.9254, "quality_target_met": true, "utility": 69.705443}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.644,\"max_latency_ms\":163746,\"minimum_quality\":0.8114},\"task\":{\"ambiguity\":0.3474,\"category\":\"coding\",\"complexity\":0.5803,\"consequence\":0.5103,\"evidence_need\":0.7848,\"family\":\"migration_planning\",\"input_tokens\":3360,\"reversibility\":0.2922,\"risk\":0.507,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":1.043,\"id\":\"worker_a\",\"p95_latency_ms\":31395,\"pass_rate\":0.4521,\"provider_group\":\"provider_x\",\"quality\":0.7989,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.442,\"id\":\"worker_b\",\"p95_latency_ms\":66044,\"pass_rate\":0.8357,\"provider_group\":\"provider_z\",\"quality\":0.9155,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.163,\"id\":\"worker_c\",\"p95_latency_ms\":55252,\"pass_rate\":0.9868,\"provider_group\":\"provider_y\",\"quality\":0.923,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.605,\"estimated_latency_ms\":121296,\"estimated_pass_rate\":0.9126,\"estimated_quality\":0.9254,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.644,\"max_latency_ms\":163746,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3474, "category": "coding", "complexity": 0.5803, "consequence": 0.5103, "evidence_need": 0.7848, "family": "migration_planning", "input_tokens": 3360, "request_id": "synthetic-migration_planning-000745", "reversibility": 0.2922, "risk": 0.507, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 1.043, "id": "worker_a", "p95_latency_ms": 31395, "pass_rate": 0.4521, "provider_group": "provider_x", "quality": 0.7989, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.442, "id": "worker_b", "p95_latency_ms": 66044, "pass_rate": 0.8357, "provider_group": "provider_z", "quality": 0.9155, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.163, "id": "worker_c", "p95_latency_ms": 55252, "pass_rate": 0.9868, "provider_group": "provider_y", "quality": 0.923, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.305, "max_latency_ms": 258034, "minimum_quality": 0.8409}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000716", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.529, "estimated_latency_ms": 154630, "estimated_pass_rate": 0.9028, "estimated_quality": 0.9284, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.305, "max_latency_ms": 258034, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9028, "quality": 0.9284, "quality_target_met": true, "utility": 69.282323}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.305,\"max_latency_ms\":258034,\"minimum_quality\":0.8409},\"task\":{\"ambiguity\":0.5963,\"category\":\"analysis\",\"complexity\":0.6693,\"consequence\":0.4334,\"evidence_need\":0.3047,\"family\":\"performance_tuning\",\"input_tokens\":2080,\"reversibility\":0.5984,\"risk\":0.4225,\"summary\":\"Review a proposed optimization for webhook delivery for hidden regressions.\"},\"workers\":[{\"credits_per_call\":0.962,\"id\":\"worker_a\",\"p95_latency_ms\":18413,\"pass_rate\":0.4982,\"provider_group\":\"provider_x\",\"quality\":0.6408,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.263,\"id\":\"worker_b\",\"p95_latency_ms\":67558,\"pass_rate\":0.8859,\"provider_group\":\"provider_y\",\"quality\":0.8911,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.266,\"id\":\"worker_c\",\"p95_latency_ms\":87072,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9932,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.529,\"estimated_latency_ms\":154630,\"estimated_pass_rate\":0.9028,\"estimated_quality\":0.9284,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.305,\"max_latency_ms\":258034,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5963, "category": "analysis", "complexity": 0.6693, "consequence": 0.4334, "evidence_need": 0.3047, "family": "performance_tuning", "input_tokens": 2080, "request_id": "synthetic-performance_tuning-000716", "reversibility": 0.5984, "risk": 0.4225, "summary": "Review a proposed optimization for webhook delivery for hidden regressions."}, "workers": [{"credits_per_call": 0.962, "id": "worker_a", "p95_latency_ms": 18413, "pass_rate": 0.4982, "provider_group": "provider_x", "quality": 0.6408, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.263, "id": "worker_b", "p95_latency_ms": 67558, "pass_rate": 0.8859, "provider_group": "provider_y", "quality": 0.8911, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.266, "id": "worker_c", "p95_latency_ms": 87072, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9932, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.405, "max_latency_ms": 153601, "minimum_quality": 0.7851}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000759", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.151, "estimated_latency_ms": 129483, "estimated_pass_rate": 0.7352, "estimated_quality": 0.9144, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.405, "max_latency_ms": 153601, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7352, "quality": 0.9144, "quality_target_met": true, "utility": 79.437636}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.405,\"max_latency_ms\":153601,\"minimum_quality\":0.7851},\"task\":{\"ambiguity\":0.4878,\"category\":\"review\",\"complexity\":0.7605,\"consequence\":0.7727,\"evidence_need\":0.9241,\"family\":\"policy_compliance\",\"input_tokens\":2400,\"reversibility\":0.6404,\"risk\":0.7989,\"summary\":\"Review whether tenant-scoped caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.934,\"id\":\"worker_a\",\"p95_latency_ms\":57888,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.003,\"id\":\"worker_b\",\"p95_latency_ms\":24439,\"pass_rate\":0.4703,\"provider_group\":\"provider_x\",\"quality\":0.6951,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.214,\"id\":\"worker_c\",\"p95_latency_ms\":47156,\"pass_rate\":0.7857,\"provider_group\":\"provider_y\",\"quality\":0.8258,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.151,\"estimated_latency_ms\":129483,\"estimated_pass_rate\":0.7352,\"estimated_quality\":0.9144,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.405,\"max_latency_ms\":153601,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4878, "category": "review", "complexity": 0.7605, "consequence": 0.7727, "evidence_need": 0.9241, "family": "policy_compliance", "input_tokens": 2400, "request_id": "synthetic-policy_compliance-000759", "reversibility": 0.6404, "risk": 0.7989, "summary": "Review whether tenant-scoped caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.934, "id": "worker_a", "p95_latency_ms": 57888, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.003, "id": "worker_b", "p95_latency_ms": 24439, "pass_rate": 0.4703, "provider_group": "provider_x", "quality": 0.6951, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.214, "id": "worker_c", "p95_latency_ms": 47156, "pass_rate": 0.7857, "provider_group": "provider_y", "quality": 0.8258, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.757, "max_latency_ms": 42519, "minimum_quality": 0.6248}, "family": "migration_planning", "id": "synthetic-migration_planning-000466", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.216, "estimated_latency_ms": 24886, "estimated_pass_rate": 0.439, "estimated_quality": 0.6438, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.757, "max_latency_ms": 42519, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.439, "quality": 0.6438, "quality_target_met": true, "utility": 44.291375}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.757,\"max_latency_ms\":42519,\"minimum_quality\":0.6248},\"task\":{\"ambiguity\":0.5377,\"category\":\"coding\",\"complexity\":0.6463,\"consequence\":0.5211,\"evidence_need\":0.4021,\"family\":\"migration_planning\",\"input_tokens\":4880,\"reversibility\":0.5166,\"risk\":0.3836,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":11.021,\"id\":\"worker_a\",\"p95_latency_ms\":89205,\"pass_rate\":0.9783,\"provider_group\":\"provider_y\",\"quality\":0.9287,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.49,\"id\":\"worker_b\",\"p95_latency_ms\":51141,\"pass_rate\":0.8217,\"provider_group\":\"provider_x\",\"quality\":0.8158,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.216,\"id\":\"worker_c\",\"p95_latency_ms\":24886,\"pass_rate\":0.584,\"provider_group\":\"provider_z\",\"quality\":0.7444,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.216,\"estimated_latency_ms\":24886,\"estimated_pass_rate\":0.439,\"estimated_quality\":0.6438,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.757,\"max_latency_ms\":42519,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5377, "category": "coding", "complexity": 0.6463, "consequence": 0.5211, "evidence_need": 0.4021, "family": "migration_planning", "input_tokens": 4880, "request_id": "synthetic-migration_planning-000466", "reversibility": 0.5166, "risk": 0.3836, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 11.021, "id": "worker_a", "p95_latency_ms": 89205, "pass_rate": 0.9783, "provider_group": "provider_y", "quality": 0.9287, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.49, "id": "worker_b", "p95_latency_ms": 51141, "pass_rate": 0.8217, "provider_group": "provider_x", "quality": 0.8158, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.216, "id": "worker_c", "p95_latency_ms": 24886, "pass_rate": 0.584, "provider_group": "provider_z", "quality": 0.7444, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.831, "max_latency_ms": 37701, "minimum_quality": 0.7404}, "family": "migration_planning", "id": "synthetic-migration_planning-000820", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.894, "estimated_latency_ms": 26634, "estimated_pass_rate": 0.3678, "estimated_quality": 0.6151, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.831, "max_latency_ms": 37701, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3678, "quality": 0.6151, "quality_target_met": false, "utility": 35.943433}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.831,\"max_latency_ms\":37701,\"minimum_quality\":0.7404},\"task\":{\"ambiguity\":0.5294,\"category\":\"coding\",\"complexity\":0.8444,\"consequence\":0.7865,\"evidence_need\":0.7331,\"family\":\"migration_planning\",\"input_tokens\":7280,\"reversibility\":0.4098,\"risk\":0.4854,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.894,\"id\":\"worker_a\",\"p95_latency_ms\":26634,\"pass_rate\":0.5429,\"provider_group\":\"provider_z\",\"quality\":0.7583,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.131,\"id\":\"worker_b\",\"p95_latency_ms\":61089,\"pass_rate\":0.9011,\"provider_group\":\"provider_x\",\"quality\":0.9808,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.06,\"id\":\"worker_c\",\"p95_latency_ms\":68968,\"pass_rate\":0.7991,\"provider_group\":\"provider_y\",\"quality\":0.8213,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.894,\"estimated_latency_ms\":26634,\"estimated_pass_rate\":0.3678,\"estimated_quality\":0.6151,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.831,\"max_latency_ms\":37701,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5294, "category": "coding", "complexity": 0.8444, "consequence": 0.7865, "evidence_need": 0.7331, "family": "migration_planning", "input_tokens": 7280, "request_id": "synthetic-migration_planning-000820", "reversibility": 0.4098, "risk": 0.4854, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 0.894, "id": "worker_a", "p95_latency_ms": 26634, "pass_rate": 0.5429, "provider_group": "provider_z", "quality": 0.7583, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.131, "id": "worker_b", "p95_latency_ms": 61089, "pass_rate": 0.9011, "provider_group": "provider_x", "quality": 0.9808, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.06, "id": "worker_c", "p95_latency_ms": 68968, "pass_rate": 0.7991, "provider_group": "provider_y", "quality": 0.8213, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.286, "max_latency_ms": 165075, "minimum_quality": 0.79}, "family": "migration_planning", "id": "synthetic-migration_planning-000100", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.695, "estimated_latency_ms": 151180, "estimated_pass_rate": 0.7463, "estimated_quality": 0.9096, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.286, "max_latency_ms": 165075, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7463, "quality": 0.9096, "quality_target_met": true, "utility": 72.703017}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.286,\"max_latency_ms\":165075,\"minimum_quality\":0.79},\"task\":{\"ambiguity\":0.4559,\"category\":\"coding\",\"complexity\":0.8194,\"consequence\":0.6898,\"evidence_need\":0.4583,\"family\":\"migration_planning\",\"input_tokens\":2400,\"reversibility\":0.4061,\"risk\":0.5716,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.076,\"id\":\"worker_a\",\"p95_latency_ms\":20367,\"pass_rate\":0.5305,\"provider_group\":\"provider_y\",\"quality\":0.7427,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.701,\"id\":\"worker_b\",\"p95_latency_ms\":69791,\"pass_rate\":0.8505,\"provider_group\":\"provider_z\",\"quality\":0.8561,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.918,\"id\":\"worker_c\",\"p95_latency_ms\":61022,\"pass_rate\":0.9589,\"provider_group\":\"provider_x\",\"quality\":0.9899,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.695,\"estimated_latency_ms\":151180,\"estimated_pass_rate\":0.7463,\"estimated_quality\":0.9096,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.286,\"max_latency_ms\":165075,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4559, "category": "coding", "complexity": 0.8194, "consequence": 0.6898, "evidence_need": 0.4583, "family": "migration_planning", "input_tokens": 2400, "request_id": "synthetic-migration_planning-000100", "reversibility": 0.4061, "risk": 0.5716, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 1.076, "id": "worker_a", "p95_latency_ms": 20367, "pass_rate": 0.5305, "provider_group": "provider_y", "quality": 0.7427, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.701, "id": "worker_b", "p95_latency_ms": 69791, "pass_rate": 0.8505, "provider_group": "provider_z", "quality": 0.8561, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.918, "id": "worker_c", "p95_latency_ms": 61022, "pass_rate": 0.9589, "provider_group": "provider_x", "quality": 0.9899, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.373, "max_latency_ms": 144556, "minimum_quality": 0.7401}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000960", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.835, "estimated_latency_ms": 88030, "estimated_pass_rate": 0.6857, "estimated_quality": 0.8329, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.373, "max_latency_ms": 144556, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6857, "quality": 0.8329, "quality_target_met": true, "utility": 65.548055}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.373,\"max_latency_ms\":144556,\"minimum_quality\":0.7401},\"task\":{\"ambiguity\":0.7058,\"category\":\"review\",\"complexity\":0.4829,\"consequence\":0.5598,\"evidence_need\":0.9444,\"family\":\"policy_compliance\",\"input_tokens\":680,\"reversibility\":0.4194,\"risk\":0.7234,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.942,\"id\":\"worker_a\",\"p95_latency_ms\":29795,\"pass_rate\":0.6181,\"provider_group\":\"provider_z\",\"quality\":0.7144,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.893,\"id\":\"worker_b\",\"p95_latency_ms\":58235,\"pass_rate\":0.8,\"provider_group\":\"provider_y\",\"quality\":0.8838,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.971,\"id\":\"worker_c\",\"p95_latency_ms\":89355,\"pass_rate\":0.9493,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.835,\"estimated_latency_ms\":88030,\"estimated_pass_rate\":0.6857,\"estimated_quality\":0.8329,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.373,\"max_latency_ms\":144556,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7058, "category": "review", "complexity": 0.4829, "consequence": 0.5598, "evidence_need": 0.9444, "family": "policy_compliance", "input_tokens": 680, "request_id": "synthetic-policy_compliance-000960", "reversibility": 0.4194, "risk": 0.7234, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 0.942, "id": "worker_a", "p95_latency_ms": 29795, "pass_rate": 0.6181, "provider_group": "provider_z", "quality": 0.7144, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.893, "id": "worker_b", "p95_latency_ms": 58235, "pass_rate": 0.8, "provider_group": "provider_y", "quality": 0.8838, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.971, "id": "worker_c", "p95_latency_ms": 89355, "pass_rate": 0.9493, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.367, "max_latency_ms": 181889, "minimum_quality": 0.8311}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000396", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.937, "estimated_latency_ms": 158523, "estimated_pass_rate": 0.7912, "estimated_quality": 0.9077, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.367, "max_latency_ms": 181889, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7912, "quality": 0.9077, "quality_target_met": true, "utility": 77.064949}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.367,\"max_latency_ms\":181889,\"minimum_quality\":0.8311},\"task\":{\"ambiguity\":0.6681,\"category\":\"review\",\"complexity\":0.5432,\"consequence\":0.739,\"evidence_need\":0.6694,\"family\":\"policy_compliance\",\"input_tokens\":4280,\"reversibility\":0.362,\"risk\":0.7958,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":9.914,\"id\":\"worker_a\",\"p95_latency_ms\":49308,\"pass_rate\":0.9686,\"provider_group\":\"provider_z\",\"quality\":0.9897,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.031,\"id\":\"worker_b\",\"p95_latency_ms\":32211,\"pass_rate\":0.6192,\"provider_group\":\"provider_y\",\"quality\":0.688,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.992,\"id\":\"worker_c\",\"p95_latency_ms\":77004,\"pass_rate\":0.7624,\"provider_group\":\"provider_x\",\"quality\":0.898,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.937,\"estimated_latency_ms\":158523,\"estimated_pass_rate\":0.7912,\"estimated_quality\":0.9077,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.367,\"max_latency_ms\":181889,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6681, "category": "review", "complexity": 0.5432, "consequence": 0.739, "evidence_need": 0.6694, "family": "policy_compliance", "input_tokens": 4280, "request_id": "synthetic-policy_compliance-000396", "reversibility": 0.362, "risk": 0.7958, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 9.914, "id": "worker_a", "p95_latency_ms": 49308, "pass_rate": 0.9686, "provider_group": "provider_z", "quality": 0.9897, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.031, "id": "worker_b", "p95_latency_ms": 32211, "pass_rate": 0.6192, "provider_group": "provider_y", "quality": 0.688, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.992, "id": "worker_c", "p95_latency_ms": 77004, "pass_rate": 0.7624, "provider_group": "provider_x", "quality": 0.898, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.96, "max_latency_ms": 184049, "minimum_quality": 0.8471}, "family": "migration_planning", "id": "synthetic-migration_planning-001171", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.996, "estimated_latency_ms": 137846, "estimated_pass_rate": 0.7505, "estimated_quality": 0.9058, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.96, "max_latency_ms": 184049, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7505, "quality": 0.9058, "quality_target_met": true, "utility": 72.102136}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.96,\"max_latency_ms\":184049,\"minimum_quality\":0.8471},\"task\":{\"ambiguity\":0.5826,\"category\":\"coding\",\"complexity\":0.7023,\"consequence\":0.5776,\"evidence_need\":0.6239,\"family\":\"migration_planning\",\"input_tokens\":3600,\"reversibility\":0.5145,\"risk\":0.4247,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":1.179,\"id\":\"worker_a\",\"p95_latency_ms\":25537,\"pass_rate\":0.5651,\"provider_group\":\"provider_y\",\"quality\":0.7353,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.086,\"id\":\"worker_b\",\"p95_latency_ms\":57477,\"pass_rate\":0.7444,\"provider_group\":\"provider_x\",\"quality\":0.8735,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.731,\"id\":\"worker_c\",\"p95_latency_ms\":54832,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.975,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.996,\"estimated_latency_ms\":137846,\"estimated_pass_rate\":0.7505,\"estimated_quality\":0.9058,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.96,\"max_latency_ms\":184049,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5826, "category": "coding", "complexity": 0.7023, "consequence": 0.5776, "evidence_need": 0.6239, "family": "migration_planning", "input_tokens": 3600, "request_id": "synthetic-migration_planning-001171", "reversibility": 0.5145, "risk": 0.4247, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 1.179, "id": "worker_a", "p95_latency_ms": 25537, "pass_rate": 0.5651, "provider_group": "provider_y", "quality": 0.7353, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.086, "id": "worker_b", "p95_latency_ms": 57477, "pass_rate": 0.7444, "provider_group": "provider_x", "quality": 0.8735, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 13.731, "id": "worker_c", "p95_latency_ms": 54832, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.975, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.129, "max_latency_ms": 40785, "minimum_quality": 0.5342}, "family": "migration_planning", "id": "synthetic-migration_planning-000343", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.13, "estimated_latency_ms": 29062, "estimated_pass_rate": 0.3589, "estimated_quality": 0.5044, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.129, "max_latency_ms": 40785, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3589, "quality": 0.5044, "quality_target_met": false, "utility": 31.353994}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.129,\"max_latency_ms\":40785,\"minimum_quality\":0.5342},\"task\":{\"ambiguity\":0.7101,\"category\":\"coding\",\"complexity\":0.9088,\"consequence\":0.4613,\"evidence_need\":0.8292,\"family\":\"migration_planning\",\"input_tokens\":2120,\"reversibility\":0.6363,\"risk\":0.6398,\"summary\":\"Review a staged migration for private credentials and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":4.186,\"id\":\"worker_a\",\"p95_latency_ms\":70820,\"pass_rate\":0.796,\"provider_group\":\"provider_z\",\"quality\":0.8123,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.13,\"id\":\"worker_b\",\"p95_latency_ms\":29062,\"pass_rate\":0.5675,\"provider_group\":\"provider_x\",\"quality\":0.6825,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.061,\"id\":\"worker_c\",\"p95_latency_ms\":68628,\"pass_rate\":0.9939,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.13,\"estimated_latency_ms\":29062,\"estimated_pass_rate\":0.3589,\"estimated_quality\":0.5044,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.129,\"max_latency_ms\":40785,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7101, "category": "coding", "complexity": 0.9088, "consequence": 0.4613, "evidence_need": 0.8292, "family": "migration_planning", "input_tokens": 2120, "request_id": "synthetic-migration_planning-000343", "reversibility": 0.6363, "risk": 0.6398, "summary": "Review a staged migration for private credentials and identify irreversible steps."}, "workers": [{"credits_per_call": 4.186, "id": "worker_a", "p95_latency_ms": 70820, "pass_rate": 0.796, "provider_group": "provider_z", "quality": 0.8123, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.13, "id": "worker_b", "p95_latency_ms": 29062, "pass_rate": 0.5675, "provider_group": "provider_x", "quality": 0.6825, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.061, "id": "worker_c", "p95_latency_ms": 68628, "pass_rate": 0.9939, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.575, "max_latency_ms": 76197, "minimum_quality": 0.8466}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000468", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.459, "estimated_latency_ms": 69183, "estimated_pass_rate": 0.5912, "estimated_quality": 0.7762, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.575, "max_latency_ms": 76197, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5912, "quality": 0.7762, "quality_target_met": false, "utility": 53.213266}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.575,\"max_latency_ms\":76197,\"minimum_quality\":0.8466},\"task\":{\"ambiguity\":0.6509,\"category\":\"review\",\"complexity\":0.5075,\"consequence\":0.5255,\"evidence_need\":0.6828,\"family\":\"policy_compliance\",\"input_tokens\":6280,\"reversibility\":0.6443,\"risk\":0.6796,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.283,\"id\":\"worker_a\",\"p95_latency_ms\":44505,\"pass_rate\":0.7308,\"provider_group\":\"provider_y\",\"quality\":0.814,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.377,\"id\":\"worker_b\",\"p95_latency_ms\":69719,\"pass_rate\":0.9843,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.176,\"id\":\"worker_c\",\"p95_latency_ms\":24678,\"pass_rate\":0.5332,\"provider_group\":\"provider_z\",\"quality\":0.7238,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.459,\"estimated_latency_ms\":69183,\"estimated_pass_rate\":0.5912,\"estimated_quality\":0.7762,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.575,\"max_latency_ms\":76197,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6509, "category": "review", "complexity": 0.5075, "consequence": 0.5255, "evidence_need": 0.6828, "family": "policy_compliance", "input_tokens": 6280, "request_id": "synthetic-policy_compliance-000468", "reversibility": 0.6443, "risk": 0.6796, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 3.283, "id": "worker_a", "p95_latency_ms": 44505, "pass_rate": 0.7308, "provider_group": "provider_y", "quality": 0.814, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.377, "id": "worker_b", "p95_latency_ms": 69719, "pass_rate": 0.9843, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.176, "id": "worker_c", "p95_latency_ms": 24678, "pass_rate": 0.5332, "provider_group": "provider_z", "quality": 0.7238, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.832, "max_latency_ms": 210848, "minimum_quality": 0.9049}, "family": "migration_planning", "id": "synthetic-migration_planning-000913", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.707, "estimated_latency_ms": 150823, "estimated_pass_rate": 0.6956, "estimated_quality": 0.8837, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.832, "max_latency_ms": 210848, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6956, "quality": 0.8837, "quality_target_met": false, "utility": 68.076556}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.832,\"max_latency_ms\":210848,\"minimum_quality\":0.9049},\"task\":{\"ambiguity\":0.4504,\"category\":\"coding\",\"complexity\":0.8257,\"consequence\":0.7342,\"evidence_need\":0.5798,\"family\":\"migration_planning\",\"input_tokens\":6120,\"reversibility\":0.5812,\"risk\":0.3894,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":9.932,\"id\":\"worker_a\",\"p95_latency_ms\":51863,\"pass_rate\":0.9451,\"provider_group\":\"provider_x\",\"quality\":0.9539,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.933,\"id\":\"worker_b\",\"p95_latency_ms\":78061,\"pass_rate\":0.7484,\"provider_group\":\"provider_z\",\"quality\":0.8214,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.842,\"id\":\"worker_c\",\"p95_latency_ms\":20899,\"pass_rate\":0.5046,\"provider_group\":\"provider_y\",\"quality\":0.7228,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.707,\"estimated_latency_ms\":150823,\"estimated_pass_rate\":0.6956,\"estimated_quality\":0.8837,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.832,\"max_latency_ms\":210848,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4504, "category": "coding", "complexity": 0.8257, "consequence": 0.7342, "evidence_need": 0.5798, "family": "migration_planning", "input_tokens": 6120, "request_id": "synthetic-migration_planning-000913", "reversibility": 0.5812, "risk": 0.3894, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 9.932, "id": "worker_a", "p95_latency_ms": 51863, "pass_rate": 0.9451, "provider_group": "provider_x", "quality": 0.9539, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.933, "id": "worker_b", "p95_latency_ms": 78061, "pass_rate": 0.7484, "provider_group": "provider_z", "quality": 0.8214, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.842, "id": "worker_c", "p95_latency_ms": 20899, "pass_rate": 0.5046, "provider_group": "provider_y", "quality": 0.7228, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.845, "max_latency_ms": 127009, "minimum_quality": 0.7693}, "family": "migration_planning", "id": "synthetic-migration_planning-001153", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.305, "estimated_latency_ms": 99458, "estimated_pass_rate": 0.6088, "estimated_quality": 0.8134, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.845, "max_latency_ms": 127009, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6088, "quality": 0.8134, "quality_target_met": true, "utility": 61.169515}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.845,\"max_latency_ms\":127009,\"minimum_quality\":0.7693},\"task\":{\"ambiguity\":0.3198,\"category\":\"coding\",\"complexity\":0.572,\"consequence\":0.821,\"evidence_need\":0.5717,\"family\":\"migration_planning\",\"input_tokens\":1120,\"reversibility\":0.4542,\"risk\":0.6497,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.157,\"id\":\"worker_a\",\"p95_latency_ms\":80413,\"pass_rate\":0.7391,\"provider_group\":\"provider_x\",\"quality\":0.8855,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.755,\"id\":\"worker_b\",\"p95_latency_ms\":64794,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9279,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.148,\"id\":\"worker_c\",\"p95_latency_ms\":19045,\"pass_rate\":0.5115,\"provider_group\":\"provider_z\",\"quality\":0.766,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.305,\"estimated_latency_ms\":99458,\"estimated_pass_rate\":0.6088,\"estimated_quality\":0.8134,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.845,\"max_latency_ms\":127009,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3198, "category": "coding", "complexity": 0.572, "consequence": 0.821, "evidence_need": 0.5717, "family": "migration_planning", "input_tokens": 1120, "request_id": "synthetic-migration_planning-001153", "reversibility": 0.4542, "risk": 0.6497, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 3.157, "id": "worker_a", "p95_latency_ms": 80413, "pass_rate": 0.7391, "provider_group": "provider_x", "quality": 0.8855, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.755, "id": "worker_b", "p95_latency_ms": 64794, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9279, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.148, "id": "worker_c", "p95_latency_ms": 19045, "pass_rate": 0.5115, "provider_group": "provider_z", "quality": 0.766, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.547, "max_latency_ms": 38996, "minimum_quality": 0.6099}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000173", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.213, "estimated_latency_ms": 32747, "estimated_pass_rate": 0.3508, "estimated_quality": 0.535, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.547, "max_latency_ms": 38996, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3508, "quality": 0.535, "quality_target_met": false, "utility": 31.919858}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.547,\"max_latency_ms\":38996,\"minimum_quality\":0.6099},\"task\":{\"ambiguity\":0.3893,\"category\":\"analysis\",\"complexity\":0.631,\"consequence\":0.5904,\"evidence_need\":0.7265,\"family\":\"performance_tuning\",\"input_tokens\":2040,\"reversibility\":0.8593,\"risk\":0.4153,\"summary\":\"Choose between throughput and tail-latency changes for prompt caching.\"},\"workers\":[{\"credits_per_call\":11.798,\"id\":\"worker_a\",\"p95_latency_ms\":58806,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.213,\"id\":\"worker_b\",\"p95_latency_ms\":32747,\"pass_rate\":0.4854,\"provider_group\":\"provider_x\",\"quality\":0.6251,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.66,\"id\":\"worker_c\",\"p95_latency_ms\":79001,\"pass_rate\":0.896,\"provider_group\":\"provider_z\",\"quality\":0.9764,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.213,\"estimated_latency_ms\":32747,\"estimated_pass_rate\":0.3508,\"estimated_quality\":0.535,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.547,\"max_latency_ms\":38996,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3893, "category": "analysis", "complexity": 0.631, "consequence": 0.5904, "evidence_need": 0.7265, "family": "performance_tuning", "input_tokens": 2040, "request_id": "synthetic-performance_tuning-000173", "reversibility": 0.8593, "risk": 0.4153, "summary": "Choose between throughput and tail-latency changes for prompt caching."}, "workers": [{"credits_per_call": 11.798, "id": "worker_a", "p95_latency_ms": 58806, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.213, "id": "worker_b", "p95_latency_ms": 32747, "pass_rate": 0.4854, "provider_group": "provider_x", "quality": 0.6251, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.66, "id": "worker_c", "p95_latency_ms": 79001, "pass_rate": 0.896, "provider_group": "provider_z", "quality": 0.9764, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 2.387, "max_latency_ms": 30223, "minimum_quality": 0.4941}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000642", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.983, "estimated_latency_ms": 28532, "estimated_pass_rate": 0.2813, "estimated_quality": 0.6372, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.387, "max_latency_ms": 30223, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2813, "quality": 0.6372, "quality_target_met": true, "utility": 24.510914}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":2.387,\"max_latency_ms\":30223,\"minimum_quality\":0.4941},\"task\":{\"ambiguity\":0.7506,\"category\":\"review\",\"complexity\":0.5562,\"consequence\":0.6791,\"evidence_need\":0.7733,\"family\":\"policy_compliance\",\"input_tokens\":3680,\"reversibility\":0.5879,\"risk\":0.8395,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.983,\"id\":\"worker_a\",\"p95_latency_ms\":28532,\"pass_rate\":0.4691,\"provider_group\":\"provider_z\",\"quality\":0.7686,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.132,\"id\":\"worker_b\",\"p95_latency_ms\":76836,\"pass_rate\":0.752,\"provider_group\":\"provider_x\",\"quality\":0.8468,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.741,\"id\":\"worker_c\",\"p95_latency_ms\":51512,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9142,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.983,\"estimated_latency_ms\":28532,\"estimated_pass_rate\":0.2813,\"estimated_quality\":0.6372,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.387,\"max_latency_ms\":30223,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7506, "category": "review", "complexity": 0.5562, "consequence": 0.6791, "evidence_need": 0.7733, "family": "policy_compliance", "input_tokens": 3680, "request_id": "synthetic-policy_compliance-000642", "reversibility": 0.5879, "risk": 0.8395, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 0.983, "id": "worker_a", "p95_latency_ms": 28532, "pass_rate": 0.4691, "provider_group": "provider_z", "quality": 0.7686, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.132, "id": "worker_b", "p95_latency_ms": 76836, "pass_rate": 0.752, "provider_group": "provider_x", "quality": 0.8468, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.741, "id": "worker_c", "p95_latency_ms": 51512, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9142, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.112, "max_latency_ms": 211908, "minimum_quality": 0.8039}, "family": "migration_planning", "id": "synthetic-migration_planning-000943", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.98, "estimated_latency_ms": 72474, "estimated_pass_rate": 0.6213, "estimated_quality": 0.8196, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.112, "max_latency_ms": 211908, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6213, "quality": 0.8196, "quality_target_met": true, "utility": 63.966283}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.112,\"max_latency_ms\":211908,\"minimum_quality\":0.8039},\"task\":{\"ambiguity\":0.3165,\"category\":\"coding\",\"complexity\":0.5757,\"consequence\":0.7404,\"evidence_need\":0.6696,\"family\":\"migration_planning\",\"input_tokens\":4200,\"reversibility\":0.6056,\"risk\":0.5562,\"summary\":\"Review a staged migration for observability and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.917,\"id\":\"worker_a\",\"p95_latency_ms\":51058,\"pass_rate\":0.7619,\"provider_group\":\"provider_y\",\"quality\":0.8056,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.063,\"id\":\"worker_b\",\"p95_latency_ms\":21416,\"pass_rate\":0.5646,\"provider_group\":\"provider_z\",\"quality\":0.7814,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.368,\"id\":\"worker_c\",\"p95_latency_ms\":84882,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.98,\"estimated_latency_ms\":72474,\"estimated_pass_rate\":0.6213,\"estimated_quality\":0.8196,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.112,\"max_latency_ms\":211908,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3165, "category": "coding", "complexity": 0.5757, "consequence": 0.7404, "evidence_need": 0.6696, "family": "migration_planning", "input_tokens": 4200, "request_id": "synthetic-migration_planning-000943", "reversibility": 0.6056, "risk": 0.5562, "summary": "Review a staged migration for observability and identify irreversible steps."}, "workers": [{"credits_per_call": 3.917, "id": "worker_a", "p95_latency_ms": 51058, "pass_rate": 0.7619, "provider_group": "provider_y", "quality": 0.8056, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.063, "id": "worker_b", "p95_latency_ms": 21416, "pass_rate": 0.5646, "provider_group": "provider_z", "quality": 0.7814, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.368, "id": "worker_c", "p95_latency_ms": 84882, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.751, "max_latency_ms": 44166, "minimum_quality": 0.7102}, "family": "migration_planning", "id": "synthetic-migration_planning-000253", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.133, "estimated_latency_ms": 24445, "estimated_pass_rate": 0.4485, "estimated_quality": 0.6349, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.751, "max_latency_ms": 44166, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4485, "quality": 0.6349, "quality_target_met": false, "utility": 41.91167}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.751,\"max_latency_ms\":44166,\"minimum_quality\":0.7102},\"task\":{\"ambiguity\":0.4832,\"category\":\"coding\",\"complexity\":0.7535,\"consequence\":0.7824,\"evidence_need\":0.696,\"family\":\"migration_planning\",\"input_tokens\":3840,\"reversibility\":0.5404,\"risk\":0.5318,\"summary\":\"Review a staged migration for observability and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.245,\"id\":\"worker_a\",\"p95_latency_ms\":81409,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9434,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.567,\"id\":\"worker_b\",\"p95_latency_ms\":65125,\"pass_rate\":0.7946,\"provider_group\":\"provider_y\",\"quality\":0.8521,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.133,\"id\":\"worker_c\",\"p95_latency_ms\":24445,\"pass_rate\":0.6136,\"provider_group\":\"provider_x\",\"quality\":0.7619,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.133,\"estimated_latency_ms\":24445,\"estimated_pass_rate\":0.4485,\"estimated_quality\":0.6349,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.751,\"max_latency_ms\":44166,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4832, "category": "coding", "complexity": 0.7535, "consequence": 0.7824, "evidence_need": 0.696, "family": "migration_planning", "input_tokens": 3840, "request_id": "synthetic-migration_planning-000253", "reversibility": 0.5404, "risk": 0.5318, "summary": "Review a staged migration for observability and identify irreversible steps."}, "workers": [{"credits_per_call": 13.245, "id": "worker_a", "p95_latency_ms": 81409, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9434, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.567, "id": "worker_b", "p95_latency_ms": 65125, "pass_rate": 0.7946, "provider_group": "provider_y", "quality": 0.8521, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.133, "id": "worker_c", "p95_latency_ms": 24445, "pass_rate": 0.6136, "provider_group": "provider_x", "quality": 0.7619, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.767, "max_latency_ms": 173229, "minimum_quality": 0.9239}, "family": "migration_planning", "id": "synthetic-migration_planning-001240", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.675, "estimated_latency_ms": 153143, "estimated_pass_rate": 0.7526, "estimated_quality": 0.9005, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 22.767, "max_latency_ms": 173229, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7526, "quality": 0.9005, "quality_target_met": false, "utility": 74.407178}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.767,\"max_latency_ms\":173229,\"minimum_quality\":0.9239},\"task\":{\"ambiguity\":0.4481,\"category\":\"coding\",\"complexity\":0.8268,\"consequence\":0.5265,\"evidence_need\":0.7461,\"family\":\"migration_planning\",\"input_tokens\":6800,\"reversibility\":0.4647,\"risk\":0.7349,\"summary\":\"Plan a zero-downtime migration of background jobs with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.987,\"id\":\"worker_a\",\"p95_latency_ms\":66040,\"pass_rate\":0.8247,\"provider_group\":\"provider_x\",\"quality\":0.8784,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.644,\"id\":\"worker_b\",\"p95_latency_ms\":61866,\"pass_rate\":0.9157,\"provider_group\":\"provider_z\",\"quality\":0.9725,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.044,\"id\":\"worker_c\",\"p95_latency_ms\":25237,\"pass_rate\":0.5656,\"provider_group\":\"provider_y\",\"quality\":0.6929,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.675,\"estimated_latency_ms\":153143,\"estimated_pass_rate\":0.7526,\"estimated_quality\":0.9005,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.767,\"max_latency_ms\":173229,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4481, "category": "coding", "complexity": 0.8268, "consequence": 0.5265, "evidence_need": 0.7461, "family": "migration_planning", "input_tokens": 6800, "request_id": "synthetic-migration_planning-001240", "reversibility": 0.4647, "risk": 0.7349, "summary": "Plan a zero-downtime migration of background jobs with rollback checkpoints."}, "workers": [{"credits_per_call": 3.987, "id": "worker_a", "p95_latency_ms": 66040, "pass_rate": 0.8247, "provider_group": "provider_x", "quality": 0.8784, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.644, "id": "worker_b", "p95_latency_ms": 61866, "pass_rate": 0.9157, "provider_group": "provider_z", "quality": 0.9725, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.044, "id": "worker_c", "p95_latency_ms": 25237, "pass_rate": 0.5656, "provider_group": "provider_y", "quality": 0.6929, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.968, "max_latency_ms": 34120, "minimum_quality": 0.6617}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000480", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.158, "estimated_latency_ms": 24823, "estimated_pass_rate": 0.4181, "estimated_quality": 0.6951, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.968, "max_latency_ms": 34120, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4181, "quality": 0.6951, "quality_target_met": true, "utility": 35.159975}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.968,\"max_latency_ms\":34120,\"minimum_quality\":0.6617},\"task\":{\"ambiguity\":0.6971,\"category\":\"review\",\"complexity\":0.5138,\"consequence\":0.7238,\"evidence_need\":0.814,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.3603,\"risk\":0.6589,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":14.788,\"id\":\"worker_a\",\"p95_latency_ms\":92373,\"pass_rate\":0.9795,\"provider_group\":\"provider_z\",\"quality\":0.97,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.158,\"id\":\"worker_b\",\"p95_latency_ms\":24823,\"pass_rate\":0.5824,\"provider_group\":\"provider_y\",\"quality\":0.8023,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.793,\"id\":\"worker_c\",\"p95_latency_ms\":46331,\"pass_rate\":0.8253,\"provider_group\":\"provider_x\",\"quality\":0.8969,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.158,\"estimated_latency_ms\":24823,\"estimated_pass_rate\":0.4181,\"estimated_quality\":0.6951,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.968,\"max_latency_ms\":34120,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6971, "category": "review", "complexity": 0.5138, "consequence": 0.7238, "evidence_need": 0.814, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-000480", "reversibility": 0.3603, "risk": 0.6589, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 14.788, "id": "worker_a", "p95_latency_ms": 92373, "pass_rate": 0.9795, "provider_group": "provider_z", "quality": 0.97, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.158, "id": "worker_b", "p95_latency_ms": 24823, "pass_rate": 0.5824, "provider_group": "provider_y", "quality": 0.8023, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.793, "id": "worker_c", "p95_latency_ms": 46331, "pass_rate": 0.8253, "provider_group": "provider_x", "quality": 0.8969, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.642, "max_latency_ms": 30615, "minimum_quality": 0.6564}, "family": "migration_planning", "id": "synthetic-migration_planning-000475", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.819, "estimated_latency_ms": 28204, "estimated_pass_rate": 0.3831, "estimated_quality": 0.6846, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.642, "max_latency_ms": 30615, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3831, "quality": 0.6846, "quality_target_met": true, "utility": 45.53675}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.642,\"max_latency_ms\":30615,\"minimum_quality\":0.6564},\"task\":{\"ambiguity\":0.4697,\"category\":\"coding\",\"complexity\":0.7851,\"consequence\":0.6998,\"evidence_need\":0.4538,\"family\":\"migration_planning\",\"input_tokens\":2720,\"reversibility\":0.6459,\"risk\":0.3722,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":11.878,\"id\":\"worker_a\",\"p95_latency_ms\":55104,\"pass_rate\":0.9158,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.062,\"id\":\"worker_b\",\"p95_latency_ms\":73400,\"pass_rate\":0.8639,\"provider_group\":\"provider_y\",\"quality\":0.82,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.819,\"id\":\"worker_c\",\"p95_latency_ms\":28204,\"pass_rate\":0.5378,\"provider_group\":\"provider_x\",\"quality\":0.8049,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.819,\"estimated_latency_ms\":28204,\"estimated_pass_rate\":0.3831,\"estimated_quality\":0.6846,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.642,\"max_latency_ms\":30615,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4697, "category": "coding", "complexity": 0.7851, "consequence": 0.6998, "evidence_need": 0.4538, "family": "migration_planning", "input_tokens": 2720, "request_id": "synthetic-migration_planning-000475", "reversibility": 0.6459, "risk": 0.3722, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 11.878, "id": "worker_a", "p95_latency_ms": 55104, "pass_rate": 0.9158, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.062, "id": "worker_b", "p95_latency_ms": 73400, "pass_rate": 0.8639, "provider_group": "provider_y", "quality": 0.82, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.819, "id": "worker_c", "p95_latency_ms": 28204, "pass_rate": 0.5378, "provider_group": "provider_x", "quality": 0.8049, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.898, "max_latency_ms": 181609, "minimum_quality": 0.792}, "family": "migration_planning", "id": "synthetic-migration_planning-000715", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.909, "estimated_latency_ms": 172097, "estimated_pass_rate": 0.7407, "estimated_quality": 0.8929, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.898, "max_latency_ms": 181609, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7407, "quality": 0.8929, "quality_target_met": true, "utility": 73.05267}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.898,\"max_latency_ms\":181609,\"minimum_quality\":0.792},\"task\":{\"ambiguity\":0.633,\"category\":\"coding\",\"complexity\":0.7918,\"consequence\":0.8306,\"evidence_need\":0.6497,\"family\":\"migration_planning\",\"input_tokens\":4680,\"reversibility\":0.3716,\"risk\":0.5445,\"summary\":\"Review a staged migration for provider failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.465,\"id\":\"worker_a\",\"p95_latency_ms\":57506,\"pass_rate\":0.8324,\"provider_group\":\"provider_y\",\"quality\":0.8781,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.796,\"id\":\"worker_b\",\"p95_latency_ms\":24605,\"pass_rate\":0.4924,\"provider_group\":\"provider_z\",\"quality\":0.7453,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.648,\"id\":\"worker_c\",\"p95_latency_ms\":89986,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9357,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.909,\"estimated_latency_ms\":172097,\"estimated_pass_rate\":0.7407,\"estimated_quality\":0.8929,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.898,\"max_latency_ms\":181609,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.633, "category": "coding", "complexity": 0.7918, "consequence": 0.8306, "evidence_need": 0.6497, "family": "migration_planning", "input_tokens": 4680, "request_id": "synthetic-migration_planning-000715", "reversibility": 0.3716, "risk": 0.5445, "summary": "Review a staged migration for provider failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.465, "id": "worker_a", "p95_latency_ms": 57506, "pass_rate": 0.8324, "provider_group": "provider_y", "quality": 0.8781, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.796, "id": "worker_b", "p95_latency_ms": 24605, "pass_rate": 0.4924, "provider_group": "provider_z", "quality": 0.7453, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.648, "id": "worker_c", "p95_latency_ms": 89986, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9357, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.667, "max_latency_ms": 255382, "minimum_quality": 0.7674}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000105", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.089, "estimated_latency_ms": 94542, "estimated_pass_rate": 0.7503, "estimated_quality": 0.8149, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.667, "max_latency_ms": 255382, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7503, "quality": 0.8149, "quality_target_met": true, "utility": 65.286341}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.667,\"max_latency_ms\":255382,\"minimum_quality\":0.7674},\"task\":{\"ambiguity\":0.3983,\"category\":\"review\",\"complexity\":0.7327,\"consequence\":0.95,\"evidence_need\":0.5573,\"family\":\"policy_compliance\",\"input_tokens\":7360,\"reversibility\":0.389,\"risk\":0.6364,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":0.983,\"id\":\"worker_a\",\"p95_latency_ms\":28746,\"pass_rate\":0.5646,\"provider_group\":\"provider_y\",\"quality\":0.7316,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.206,\"id\":\"worker_b\",\"p95_latency_ms\":73778,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9288,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.106,\"id\":\"worker_c\",\"p95_latency_ms\":65796,\"pass_rate\":0.8429,\"provider_group\":\"provider_x\",\"quality\":0.885,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.089,\"estimated_latency_ms\":94542,\"estimated_pass_rate\":0.7503,\"estimated_quality\":0.8149,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.667,\"max_latency_ms\":255382,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3983, "category": "review", "complexity": 0.7327, "consequence": 0.95, "evidence_need": 0.5573, "family": "policy_compliance", "input_tokens": 7360, "request_id": "synthetic-policy_compliance-000105", "reversibility": 0.389, "risk": 0.6364, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 0.983, "id": "worker_a", "p95_latency_ms": 28746, "pass_rate": 0.5646, "provider_group": "provider_y", "quality": 0.7316, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.206, "id": "worker_b", "p95_latency_ms": 73778, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9288, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.106, "id": "worker_c", "p95_latency_ms": 65796, "pass_rate": 0.8429, "provider_group": "provider_x", "quality": 0.885, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.574, "max_latency_ms": 160942, "minimum_quality": 0.8154}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000627", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.419, "estimated_latency_ms": 104123, "estimated_pass_rate": 0.6576, "estimated_quality": 0.8231, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.574, "max_latency_ms": 160942, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6576, "quality": 0.8231, "quality_target_met": true, "utility": 62.635577}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.574,\"max_latency_ms\":160942,\"minimum_quality\":0.8154},\"task\":{\"ambiguity\":0.5227,\"category\":\"review\",\"complexity\":0.7033,\"consequence\":0.5608,\"evidence_need\":0.6274,\"family\":\"policy_compliance\",\"input_tokens\":5520,\"reversibility\":0.5526,\"risk\":0.744,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.209,\"id\":\"worker_a\",\"p95_latency_ms\":23145,\"pass_rate\":0.5896,\"provider_group\":\"provider_z\",\"quality\":0.7986,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.463,\"id\":\"worker_b\",\"p95_latency_ms\":88734,\"pass_rate\":0.9161,\"provider_group\":\"provider_y\",\"quality\":0.9336,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.21,\"id\":\"worker_c\",\"p95_latency_ms\":80978,\"pass_rate\":0.858,\"provider_group\":\"provider_x\",\"quality\":0.8002,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.419,\"estimated_latency_ms\":104123,\"estimated_pass_rate\":0.6576,\"estimated_quality\":0.8231,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.574,\"max_latency_ms\":160942,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5227, "category": "review", "complexity": 0.7033, "consequence": 0.5608, "evidence_need": 0.6274, "family": "policy_compliance", "input_tokens": 5520, "request_id": "synthetic-policy_compliance-000627", "reversibility": 0.5526, "risk": 0.744, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 1.209, "id": "worker_a", "p95_latency_ms": 23145, "pass_rate": 0.5896, "provider_group": "provider_z", "quality": 0.7986, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.463, "id": "worker_b", "p95_latency_ms": 88734, "pass_rate": 0.9161, "provider_group": "provider_y", "quality": 0.9336, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.21, "id": "worker_c", "p95_latency_ms": 80978, "pass_rate": 0.858, "provider_group": "provider_x", "quality": 0.8002, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.447, "max_latency_ms": 28402, "minimum_quality": 0.6679}, "family": "migration_planning", "id": "synthetic-migration_planning-000412", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.114, "estimated_latency_ms": 21661, "estimated_pass_rate": 0.3279, "estimated_quality": 0.6468, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.447, "max_latency_ms": 28402, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3279, "quality": 0.6468, "quality_target_met": false, "utility": 40.138382}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.447,\"max_latency_ms\":28402,\"minimum_quality\":0.6679},\"task\":{\"ambiguity\":0.3716,\"category\":\"coding\",\"complexity\":0.9438,\"consequence\":0.4324,\"evidence_need\":0.7594,\"family\":\"migration_planning\",\"input_tokens\":4720,\"reversibility\":0.471,\"risk\":0.4615,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.588,\"id\":\"worker_a\",\"p95_latency_ms\":52707,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9142,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.114,\"id\":\"worker_b\",\"p95_latency_ms\":21661,\"pass_rate\":0.5004,\"provider_group\":\"provider_y\",\"quality\":0.7956,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.782,\"id\":\"worker_c\",\"p95_latency_ms\":60005,\"pass_rate\":0.7817,\"provider_group\":\"provider_z\",\"quality\":0.8439,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.114,\"estimated_latency_ms\":21661,\"estimated_pass_rate\":0.3279,\"estimated_quality\":0.6468,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.447,\"max_latency_ms\":28402,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3716, "category": "coding", "complexity": 0.9438, "consequence": 0.4324, "evidence_need": 0.7594, "family": "migration_planning", "input_tokens": 4720, "request_id": "synthetic-migration_planning-000412", "reversibility": 0.471, "risk": 0.4615, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 14.588, "id": "worker_a", "p95_latency_ms": 52707, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9142, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.114, "id": "worker_b", "p95_latency_ms": 21661, "pass_rate": 0.5004, "provider_group": "provider_y", "quality": 0.7956, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.782, "id": "worker_c", "p95_latency_ms": 60005, "pass_rate": 0.7817, "provider_group": "provider_z", "quality": 0.8439, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.375, "max_latency_ms": 55072, "minimum_quality": 0.6601}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000750", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.9, "estimated_latency_ms": 31709, "estimated_pass_rate": 0.3747, "estimated_quality": 0.608, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.375, "max_latency_ms": 55072, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3747, "quality": 0.608, "quality_target_met": false, "utility": 38.882288}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.375,\"max_latency_ms\":55072,\"minimum_quality\":0.6601},\"task\":{\"ambiguity\":0.7465,\"category\":\"review\",\"complexity\":0.8322,\"consequence\":0.6779,\"evidence_need\":0.7252,\"family\":\"policy_compliance\",\"input_tokens\":440,\"reversibility\":0.3911,\"risk\":0.4763,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":0.9,\"id\":\"worker_a\",\"p95_latency_ms\":31709,\"pass_rate\":0.5641,\"provider_group\":\"provider_z\",\"quality\":0.7629,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.237,\"id\":\"worker_b\",\"p95_latency_ms\":76323,\"pass_rate\":0.7529,\"provider_group\":\"provider_x\",\"quality\":0.8917,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.558,\"id\":\"worker_c\",\"p95_latency_ms\":86685,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9548,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.9,\"estimated_latency_ms\":31709,\"estimated_pass_rate\":0.3747,\"estimated_quality\":0.608,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.375,\"max_latency_ms\":55072,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7465, "category": "review", "complexity": 0.8322, "consequence": 0.6779, "evidence_need": 0.7252, "family": "policy_compliance", "input_tokens": 440, "request_id": "synthetic-policy_compliance-000750", "reversibility": 0.3911, "risk": 0.4763, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 0.9, "id": "worker_a", "p95_latency_ms": 31709, "pass_rate": 0.5641, "provider_group": "provider_z", "quality": 0.7629, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.237, "id": "worker_b", "p95_latency_ms": 76323, "pass_rate": 0.7529, "provider_group": "provider_x", "quality": 0.8917, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 11.558, "id": "worker_c", "p95_latency_ms": 86685, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9548, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.489, "max_latency_ms": 182405, "minimum_quality": 0.7861}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000837", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.844, "estimated_latency_ms": 146361, "estimated_pass_rate": 0.7489, "estimated_quality": 0.8833, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.489, "max_latency_ms": 182405, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7489, "quality": 0.8833, "quality_target_met": true, "utility": 77.091259}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.489,\"max_latency_ms\":182405,\"minimum_quality\":0.7861},\"task\":{\"ambiguity\":0.4609,\"category\":\"review\",\"complexity\":0.7442,\"consequence\":0.8443,\"evidence_need\":0.7325,\"family\":\"policy_compliance\",\"input_tokens\":3120,\"reversibility\":0.484,\"risk\":0.7849,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.846,\"id\":\"worker_a\",\"p95_latency_ms\":69775,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9381,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.208,\"id\":\"worker_b\",\"p95_latency_ms\":56726,\"pass_rate\":0.7316,\"provider_group\":\"provider_x\",\"quality\":0.8864,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.79,\"id\":\"worker_c\",\"p95_latency_ms\":19860,\"pass_rate\":0.5066,\"provider_group\":\"provider_y\",\"quality\":0.7097,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.844,\"estimated_latency_ms\":146361,\"estimated_pass_rate\":0.7489,\"estimated_quality\":0.8833,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.489,\"max_latency_ms\":182405,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4609, "category": "review", "complexity": 0.7442, "consequence": 0.8443, "evidence_need": 0.7325, "family": "policy_compliance", "input_tokens": 3120, "request_id": "synthetic-policy_compliance-000837", "reversibility": 0.484, "risk": 0.7849, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 13.846, "id": "worker_a", "p95_latency_ms": 69775, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9381, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.208, "id": "worker_b", "p95_latency_ms": 56726, "pass_rate": 0.7316, "provider_group": "provider_x", "quality": 0.8864, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.79, "id": "worker_c", "p95_latency_ms": 19860, "pass_rate": 0.5066, "provider_group": "provider_y", "quality": 0.7097, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.985, "max_latency_ms": 36939, "minimum_quality": 0.5581}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000461", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.189, "estimated_latency_ms": 19028, "estimated_pass_rate": 0.3455, "estimated_quality": 0.523, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.985, "max_latency_ms": 36939, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3455, "quality": 0.523, "quality_target_met": false, "utility": 32.387315}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.985,\"max_latency_ms\":36939,\"minimum_quality\":0.5581},\"task\":{\"ambiguity\":0.4278,\"category\":\"analysis\",\"complexity\":0.822,\"consequence\":0.5638,\"evidence_need\":0.3682,\"family\":\"performance_tuning\",\"input_tokens\":7640,\"reversibility\":0.5625,\"risk\":0.231,\"summary\":\"Review a proposed optimization for prompt caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":4.245,\"id\":\"worker_a\",\"p95_latency_ms\":72021,\"pass_rate\":0.8378,\"provider_group\":\"provider_z\",\"quality\":0.9616,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.373,\"id\":\"worker_b\",\"p95_latency_ms\":82811,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9883,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.189,\"id\":\"worker_c\",\"p95_latency_ms\":19028,\"pass_rate\":0.4898,\"provider_group\":\"provider_y\",\"quality\":0.637,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.189,\"estimated_latency_ms\":19028,\"estimated_pass_rate\":0.3455,\"estimated_quality\":0.523,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.985,\"max_latency_ms\":36939,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4278, "category": "analysis", "complexity": 0.822, "consequence": 0.5638, "evidence_need": 0.3682, "family": "performance_tuning", "input_tokens": 7640, "request_id": "synthetic-performance_tuning-000461", "reversibility": 0.5625, "risk": 0.231, "summary": "Review a proposed optimization for prompt caching for hidden regressions."}, "workers": [{"credits_per_call": 4.245, "id": "worker_a", "p95_latency_ms": 72021, "pass_rate": 0.8378, "provider_group": "provider_z", "quality": 0.9616, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.373, "id": "worker_b", "p95_latency_ms": 82811, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9883, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.189, "id": "worker_c", "p95_latency_ms": 19028, "pass_rate": 0.4898, "provider_group": "provider_y", "quality": 0.637, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.841, "max_latency_ms": 199439, "minimum_quality": 0.8237}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000371", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.443, "estimated_latency_ms": 164992, "estimated_pass_rate": 0.7475, "estimated_quality": 0.9063, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.841, "max_latency_ms": 199439, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7475, "quality": 0.9063, "quality_target_met": true, "utility": 70.663891}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.841,\"max_latency_ms\":199439,\"minimum_quality\":0.8237},\"task\":{\"ambiguity\":0.6427,\"category\":\"analysis\",\"complexity\":0.7267,\"consequence\":0.6614,\"evidence_need\":0.3304,\"family\":\"performance_tuning\",\"input_tokens\":4120,\"reversibility\":0.7203,\"risk\":0.5106,\"summary\":\"Review a proposed optimization for tenant-scoped caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.078,\"id\":\"worker_a\",\"p95_latency_ms\":18353,\"pass_rate\":0.3947,\"provider_group\":\"provider_x\",\"quality\":0.637,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.07,\"id\":\"worker_b\",\"p95_latency_ms\":57088,\"pass_rate\":0.9275,\"provider_group\":\"provider_z\",\"quality\":0.9704,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.295,\"id\":\"worker_c\",\"p95_latency_ms\":89551,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.443,\"estimated_latency_ms\":164992,\"estimated_pass_rate\":0.7475,\"estimated_quality\":0.9063,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.841,\"max_latency_ms\":199439,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6427, "category": "analysis", "complexity": 0.7267, "consequence": 0.6614, "evidence_need": 0.3304, "family": "performance_tuning", "input_tokens": 4120, "request_id": "synthetic-performance_tuning-000371", "reversibility": 0.7203, "risk": 0.5106, "summary": "Review a proposed optimization for tenant-scoped caching for hidden regressions."}, "workers": [{"credits_per_call": 1.078, "id": "worker_a", "p95_latency_ms": 18353, "pass_rate": 0.3947, "provider_group": "provider_x", "quality": 0.637, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.07, "id": "worker_b", "p95_latency_ms": 57088, "pass_rate": 0.9275, "provider_group": "provider_z", "quality": 0.9704, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.295, "id": "worker_c", "p95_latency_ms": 89551, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.995, "max_latency_ms": 37084, "minimum_quality": 0.735}, "family": "migration_planning", "id": "synthetic-migration_planning-000727", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.091, "estimated_latency_ms": 20674, "estimated_pass_rate": 0.4286, "estimated_quality": 0.5784, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.995, "max_latency_ms": 37084, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4286, "quality": 0.5784, "quality_target_met": false, "utility": 34.082774}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.995,\"max_latency_ms\":37084,\"minimum_quality\":0.735},\"task\":{\"ambiguity\":0.449,\"category\":\"coding\",\"complexity\":0.8463,\"consequence\":0.575,\"evidence_need\":0.7559,\"family\":\"migration_planning\",\"input_tokens\":3800,\"reversibility\":0.4705,\"risk\":0.5742,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":1.091,\"id\":\"worker_a\",\"p95_latency_ms\":20674,\"pass_rate\":0.605,\"provider_group\":\"provider_y\",\"quality\":0.7229,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.639,\"id\":\"worker_b\",\"p95_latency_ms\":47929,\"pass_rate\":0.8667,\"provider_group\":\"provider_z\",\"quality\":0.8055,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.054,\"id\":\"worker_c\",\"p95_latency_ms\":80074,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9389,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.091,\"estimated_latency_ms\":20674,\"estimated_pass_rate\":0.4286,\"estimated_quality\":0.5784,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.995,\"max_latency_ms\":37084,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.449, "category": "coding", "complexity": 0.8463, "consequence": 0.575, "evidence_need": 0.7559, "family": "migration_planning", "input_tokens": 3800, "request_id": "synthetic-migration_planning-000727", "reversibility": 0.4705, "risk": 0.5742, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 1.091, "id": "worker_a", "p95_latency_ms": 20674, "pass_rate": 0.605, "provider_group": "provider_y", "quality": 0.7229, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.639, "id": "worker_b", "p95_latency_ms": 47929, "pass_rate": 0.8667, "provider_group": "provider_z", "quality": 0.8055, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.054, "id": "worker_c", "p95_latency_ms": 80074, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9389, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.896, "max_latency_ms": 165054, "minimum_quality": 0.7934}, "family": "migration_planning", "id": "synthetic-migration_planning-000781", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.068, "estimated_latency_ms": 141187, "estimated_pass_rate": 0.7802, "estimated_quality": 0.917, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.896, "max_latency_ms": 165054, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7802, "quality": 0.917, "quality_target_met": true, "utility": 74.689581}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.896,\"max_latency_ms\":165054,\"minimum_quality\":0.7934},\"task\":{\"ambiguity\":0.5394,\"category\":\"coding\",\"complexity\":0.8562,\"consequence\":0.4093,\"evidence_need\":0.4336,\"family\":\"migration_planning\",\"input_tokens\":240,\"reversibility\":0.5705,\"risk\":0.6983,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.902,\"id\":\"worker_a\",\"p95_latency_ms\":24404,\"pass_rate\":0.5862,\"provider_group\":\"provider_z\",\"quality\":0.7868,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.903,\"id\":\"worker_b\",\"p95_latency_ms\":65609,\"pass_rate\":0.8448,\"provider_group\":\"provider_y\",\"quality\":0.8386,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.263,\"id\":\"worker_c\",\"p95_latency_ms\":51174,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.068,\"estimated_latency_ms\":141187,\"estimated_pass_rate\":0.7802,\"estimated_quality\":0.917,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.896,\"max_latency_ms\":165054,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5394, "category": "coding", "complexity": 0.8562, "consequence": 0.4093, "evidence_need": 0.4336, "family": "migration_planning", "input_tokens": 240, "request_id": "synthetic-migration_planning-000781", "reversibility": 0.5705, "risk": 0.6983, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 0.902, "id": "worker_a", "p95_latency_ms": 24404, "pass_rate": 0.5862, "provider_group": "provider_z", "quality": 0.7868, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.903, "id": "worker_b", "p95_latency_ms": 65609, "pass_rate": 0.8448, "provider_group": "provider_y", "quality": 0.8386, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.263, "id": "worker_c", "p95_latency_ms": 51174, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.001, "max_latency_ms": 48682, "minimum_quality": 0.7049}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000191", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.172, "estimated_latency_ms": 24985, "estimated_pass_rate": 0.3234, "estimated_quality": 0.6271, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.001, "max_latency_ms": 48682, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3234, "quality": 0.6271, "quality_target_met": false, "utility": 38.007515}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.001,\"max_latency_ms\":48682,\"minimum_quality\":0.7049},\"task\":{\"ambiguity\":0.7394,\"category\":\"analysis\",\"complexity\":0.494,\"consequence\":0.3265,\"evidence_need\":0.7481,\"family\":\"performance_tuning\",\"input_tokens\":6000,\"reversibility\":0.6173,\"risk\":0.5742,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":12.872,\"id\":\"worker_a\",\"p95_latency_ms\":76842,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.266,\"id\":\"worker_b\",\"p95_latency_ms\":74996,\"pass_rate\":0.8743,\"provider_group\":\"provider_y\",\"quality\":0.9179,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.172,\"id\":\"worker_c\",\"p95_latency_ms\":24985,\"pass_rate\":0.4818,\"provider_group\":\"provider_x\",\"quality\":0.7274,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.172,\"estimated_latency_ms\":24985,\"estimated_pass_rate\":0.3234,\"estimated_quality\":0.6271,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.001,\"max_latency_ms\":48682,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7394, "category": "analysis", "complexity": 0.494, "consequence": 0.3265, "evidence_need": 0.7481, "family": "performance_tuning", "input_tokens": 6000, "request_id": "synthetic-performance_tuning-000191", "reversibility": 0.6173, "risk": 0.5742, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 12.872, "id": "worker_a", "p95_latency_ms": 76842, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.266, "id": "worker_b", "p95_latency_ms": 74996, "pass_rate": 0.8743, "provider_group": "provider_y", "quality": 0.9179, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.172, "id": "worker_c", "p95_latency_ms": 24985, "pass_rate": 0.4818, "provider_group": "provider_x", "quality": 0.7274, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.593, "max_latency_ms": 215712, "minimum_quality": 0.8074}, "family": "migration_planning", "id": "synthetic-migration_planning-000073", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.505, "estimated_latency_ms": 170495, "estimated_pass_rate": 0.737, "estimated_quality": 0.8831, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.593, "max_latency_ms": 215712, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.737, "quality": 0.8831, "quality_target_met": true, "utility": 72.791098}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.593,\"max_latency_ms\":215712,\"minimum_quality\":0.8074},\"task\":{\"ambiguity\":0.6331,\"category\":\"coding\",\"complexity\":0.9217,\"consequence\":0.5821,\"evidence_need\":0.4551,\"family\":\"migration_planning\",\"input_tokens\":5760,\"reversibility\":0.5641,\"risk\":0.5768,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.863,\"id\":\"worker_a\",\"p95_latency_ms\":27319,\"pass_rate\":0.5965,\"provider_group\":\"provider_x\",\"quality\":0.7207,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.962,\"id\":\"worker_b\",\"p95_latency_ms\":83361,\"pass_rate\":0.8899,\"provider_group\":\"provider_z\",\"quality\":0.9592,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.68,\"id\":\"worker_c\",\"p95_latency_ms\":59815,\"pass_rate\":0.8019,\"provider_group\":\"provider_y\",\"quality\":0.9164,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.505,\"estimated_latency_ms\":170495,\"estimated_pass_rate\":0.737,\"estimated_quality\":0.8831,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.593,\"max_latency_ms\":215712,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6331, "category": "coding", "complexity": 0.9217, "consequence": 0.5821, "evidence_need": 0.4551, "family": "migration_planning", "input_tokens": 5760, "request_id": "synthetic-migration_planning-000073", "reversibility": 0.5641, "risk": 0.5768, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 0.863, "id": "worker_a", "p95_latency_ms": 27319, "pass_rate": 0.5965, "provider_group": "provider_x", "quality": 0.7207, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.962, "id": "worker_b", "p95_latency_ms": 83361, "pass_rate": 0.8899, "provider_group": "provider_z", "quality": 0.9592, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.68, "id": "worker_c", "p95_latency_ms": 59815, "pass_rate": 0.8019, "provider_group": "provider_y", "quality": 0.9164, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.846, "max_latency_ms": 218050, "minimum_quality": 0.9249}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000663", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.201, "estimated_latency_ms": 175250, "estimated_pass_rate": 0.7173, "estimated_quality": 0.9102, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 26.846, "max_latency_ms": 218050, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7173, "quality": 0.9102, "quality_target_met": false, "utility": 75.727037}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.846,\"max_latency_ms\":218050,\"minimum_quality\":0.9249},\"task\":{\"ambiguity\":0.4931,\"category\":\"review\",\"complexity\":0.8199,\"consequence\":0.5516,\"evidence_need\":0.8602,\"family\":\"policy_compliance\",\"input_tokens\":3160,\"reversibility\":0.3799,\"risk\":0.6662,\"summary\":\"Audit a workflow for background jobs against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.97,\"id\":\"worker_a\",\"p95_latency_ms\":71055,\"pass_rate\":0.8954,\"provider_group\":\"provider_x\",\"quality\":0.9715,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.011,\"id\":\"worker_b\",\"p95_latency_ms\":71566,\"pass_rate\":0.7581,\"provider_group\":\"provider_z\",\"quality\":0.8865,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.22,\"id\":\"worker_c\",\"p95_latency_ms\":32629,\"pass_rate\":0.5385,\"provider_group\":\"provider_y\",\"quality\":0.7706,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.201,\"estimated_latency_ms\":175250,\"estimated_pass_rate\":0.7173,\"estimated_quality\":0.9102,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.846,\"max_latency_ms\":218050,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4931, "category": "review", "complexity": 0.8199, "consequence": 0.5516, "evidence_need": 0.8602, "family": "policy_compliance", "input_tokens": 3160, "request_id": "synthetic-policy_compliance-000663", "reversibility": 0.3799, "risk": 0.6662, "summary": "Audit a workflow for background jobs against explicit access requirements."}, "workers": [{"credits_per_call": 13.97, "id": "worker_a", "p95_latency_ms": 71055, "pass_rate": 0.8954, "provider_group": "provider_x", "quality": 0.9715, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.011, "id": "worker_b", "p95_latency_ms": 71566, "pass_rate": 0.7581, "provider_group": "provider_z", "quality": 0.8865, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.22, "id": "worker_c", "p95_latency_ms": 32629, "pass_rate": 0.5385, "provider_group": "provider_y", "quality": 0.7706, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.736, "max_latency_ms": 27789, "minimum_quality": 0.5312}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000275", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.134, "estimated_latency_ms": 22125, "estimated_pass_rate": 0.3104, "estimated_quality": 0.5407, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.736, "max_latency_ms": 27789, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3104, "quality": 0.5407, "quality_target_met": true, "utility": 32.507217}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.736,\"max_latency_ms\":27789,\"minimum_quality\":0.5312},\"task\":{\"ambiguity\":0.3472,\"category\":\"analysis\",\"complexity\":0.6235,\"consequence\":0.6374,\"evidence_need\":0.5729,\"family\":\"performance_tuning\",\"input_tokens\":2040,\"reversibility\":0.7765,\"risk\":0.2486,\"summary\":\"Review a proposed optimization for rate limiting for hidden regressions.\"},\"workers\":[{\"credits_per_call\":10.944,\"id\":\"worker_a\",\"p95_latency_ms\":77224,\"pass_rate\":0.9942,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.732,\"id\":\"worker_b\",\"p95_latency_ms\":64339,\"pass_rate\":0.8693,\"provider_group\":\"provider_x\",\"quality\":0.8843,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.134,\"id\":\"worker_c\",\"p95_latency_ms\":22125,\"pass_rate\":0.4275,\"provider_group\":\"provider_y\",\"quality\":0.6147,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.134,\"estimated_latency_ms\":22125,\"estimated_pass_rate\":0.3104,\"estimated_quality\":0.5407,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.736,\"max_latency_ms\":27789,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3472, "category": "analysis", "complexity": 0.6235, "consequence": 0.6374, "evidence_need": 0.5729, "family": "performance_tuning", "input_tokens": 2040, "request_id": "synthetic-performance_tuning-000275", "reversibility": 0.7765, "risk": 0.2486, "summary": "Review a proposed optimization for rate limiting for hidden regressions."}, "workers": [{"credits_per_call": 10.944, "id": "worker_a", "p95_latency_ms": 77224, "pass_rate": 0.9942, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.732, "id": "worker_b", "p95_latency_ms": 64339, "pass_rate": 0.8693, "provider_group": "provider_x", "quality": 0.8843, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.134, "id": "worker_c", "p95_latency_ms": 22125, "pass_rate": 0.4275, "provider_group": "provider_y", "quality": 0.6147, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.751, "max_latency_ms": 235176, "minimum_quality": 0.8811}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001154", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.171, "estimated_latency_ms": 171505, "estimated_pass_rate": 0.7953, "estimated_quality": 0.9203, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.751, "max_latency_ms": 235176, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7953, "quality": 0.9203, "quality_target_met": true, "utility": 76.29779}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.751,\"max_latency_ms\":235176,\"minimum_quality\":0.8811},\"task\":{\"ambiguity\":0.4156,\"category\":\"analysis\",\"complexity\":0.8355,\"consequence\":0.5771,\"evidence_need\":0.6264,\"family\":\"performance_tuning\",\"input_tokens\":4280,\"reversibility\":0.6941,\"risk\":0.5854,\"summary\":\"Review a proposed optimization for credit reservation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.138,\"id\":\"worker_a\",\"p95_latency_ms\":28743,\"pass_rate\":0.5515,\"provider_group\":\"provider_x\",\"quality\":0.6688,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.26,\"id\":\"worker_b\",\"p95_latency_ms\":60703,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.773,\"id\":\"worker_c\",\"p95_latency_ms\":82059,\"pass_rate\":0.8909,\"provider_group\":\"provider_y\",\"quality\":0.9523,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.171,\"estimated_latency_ms\":171505,\"estimated_pass_rate\":0.7953,\"estimated_quality\":0.9203,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.751,\"max_latency_ms\":235176,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4156, "category": "analysis", "complexity": 0.8355, "consequence": 0.5771, "evidence_need": 0.6264, "family": "performance_tuning", "input_tokens": 4280, "request_id": "synthetic-performance_tuning-001154", "reversibility": 0.6941, "risk": 0.5854, "summary": "Review a proposed optimization for credit reservation for hidden regressions."}, "workers": [{"credits_per_call": 1.138, "id": "worker_a", "p95_latency_ms": 28743, "pass_rate": 0.5515, "provider_group": "provider_x", "quality": 0.6688, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.26, "id": "worker_b", "p95_latency_ms": 60703, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.773, "id": "worker_c", "p95_latency_ms": 82059, "pass_rate": 0.8909, "provider_group": "provider_y", "quality": 0.9523, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 4.98, "max_latency_ms": 201465, "minimum_quality": 0.722}, "family": "migration_planning", "id": "synthetic-migration_planning-000622", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.052, "estimated_latency_ms": 65284, "estimated_pass_rate": 0.541, "estimated_quality": 0.7757, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.98, "max_latency_ms": 201465, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.541, "quality": 0.7757, "quality_target_met": true, "utility": 54.949173}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":4.98,\"max_latency_ms\":201465,\"minimum_quality\":0.722},\"task\":{\"ambiguity\":0.6985,\"category\":\"coding\",\"complexity\":0.7086,\"consequence\":0.7047,\"evidence_need\":0.843,\"family\":\"migration_planning\",\"input_tokens\":1440,\"reversibility\":0.5898,\"risk\":0.3902,\"summary\":\"Review a staged migration for private credentials and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.835,\"id\":\"worker_a\",\"p95_latency_ms\":47290,\"pass_rate\":0.8425,\"provider_group\":\"provider_x\",\"quality\":0.7977,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.217,\"id\":\"worker_b\",\"p95_latency_ms\":17994,\"pass_rate\":0.4796,\"provider_group\":\"provider_z\",\"quality\":0.6971,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.291,\"id\":\"worker_c\",\"p95_latency_ms\":87477,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9532,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.052,\"estimated_latency_ms\":65284,\"estimated_pass_rate\":0.541,\"estimated_quality\":0.7757,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.98,\"max_latency_ms\":201465,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6985, "category": "coding", "complexity": 0.7086, "consequence": 0.7047, "evidence_need": 0.843, "family": "migration_planning", "input_tokens": 1440, "request_id": "synthetic-migration_planning-000622", "reversibility": 0.5898, "risk": 0.3902, "summary": "Review a staged migration for private credentials and identify irreversible steps."}, "workers": [{"credits_per_call": 2.835, "id": "worker_a", "p95_latency_ms": 47290, "pass_rate": 0.8425, "provider_group": "provider_x", "quality": 0.7977, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.217, "id": "worker_b", "p95_latency_ms": 17994, "pass_rate": 0.4796, "provider_group": "provider_z", "quality": 0.6971, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.291, "id": "worker_c", "p95_latency_ms": 87477, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9532, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.248, "max_latency_ms": 180786, "minimum_quality": 0.8838}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001253", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.38, "estimated_latency_ms": 147633, "estimated_pass_rate": 0.7497, "estimated_quality": 0.9229, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.248, "max_latency_ms": 180786, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7497, "quality": 0.9229, "quality_target_met": true, "utility": 71.054912}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.248,\"max_latency_ms\":180786,\"minimum_quality\":0.8838},\"task\":{\"ambiguity\":0.6363,\"category\":\"analysis\",\"complexity\":0.7825,\"consequence\":0.4501,\"evidence_need\":0.6145,\"family\":\"performance_tuning\",\"input_tokens\":2600,\"reversibility\":0.8997,\"risk\":0.2326,\"summary\":\"Review a proposed optimization for prompt caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":0.843,\"id\":\"worker_a\",\"p95_latency_ms\":28943,\"pass_rate\":0.547,\"provider_group\":\"provider_y\",\"quality\":0.7219,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.356,\"id\":\"worker_b\",\"p95_latency_ms\":66978,\"pass_rate\":0.793,\"provider_group\":\"provider_z\",\"quality\":0.9819,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.181,\"id\":\"worker_c\",\"p95_latency_ms\":51712,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.38,\"estimated_latency_ms\":147633,\"estimated_pass_rate\":0.7497,\"estimated_quality\":0.9229,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.248,\"max_latency_ms\":180786,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6363, "category": "analysis", "complexity": 0.7825, "consequence": 0.4501, "evidence_need": 0.6145, "family": "performance_tuning", "input_tokens": 2600, "request_id": "synthetic-performance_tuning-001253", "reversibility": 0.8997, "risk": 0.2326, "summary": "Review a proposed optimization for prompt caching for hidden regressions."}, "workers": [{"credits_per_call": 0.843, "id": "worker_a", "p95_latency_ms": 28943, "pass_rate": 0.547, "provider_group": "provider_y", "quality": 0.7219, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.356, "id": "worker_b", "p95_latency_ms": 66978, "pass_rate": 0.793, "provider_group": "provider_z", "quality": 0.9819, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.181, "id": "worker_c", "p95_latency_ms": 51712, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.693, "max_latency_ms": 195162, "minimum_quality": 0.8416}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000129", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.119, "estimated_latency_ms": 153630, "estimated_pass_rate": 0.7118, "estimated_quality": 0.906, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.693, "max_latency_ms": 195162, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7118, "quality": 0.906, "quality_target_met": true, "utility": 72.536761}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.693,\"max_latency_ms\":195162,\"minimum_quality\":0.8416},\"task\":{\"ambiguity\":0.5902,\"category\":\"review\",\"complexity\":0.4332,\"consequence\":0.5838,\"evidence_need\":0.8043,\"family\":\"policy_compliance\",\"input_tokens\":1120,\"reversibility\":0.6743,\"risk\":0.7143,\"summary\":\"Audit a workflow for credit reservation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.841,\"id\":\"worker_a\",\"p95_latency_ms\":26966,\"pass_rate\":0.4902,\"provider_group\":\"provider_x\",\"quality\":0.7685,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.982,\"id\":\"worker_b\",\"p95_latency_ms\":80896,\"pass_rate\":0.9142,\"provider_group\":\"provider_y\",\"quality\":0.953,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.296,\"id\":\"worker_c\",\"p95_latency_ms\":45768,\"pass_rate\":0.7397,\"provider_group\":\"provider_z\",\"quality\":0.8475,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.119,\"estimated_latency_ms\":153630,\"estimated_pass_rate\":0.7118,\"estimated_quality\":0.906,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.693,\"max_latency_ms\":195162,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5902, "category": "review", "complexity": 0.4332, "consequence": 0.5838, "evidence_need": 0.8043, "family": "policy_compliance", "input_tokens": 1120, "request_id": "synthetic-policy_compliance-000129", "reversibility": 0.6743, "risk": 0.7143, "summary": "Audit a workflow for credit reservation against explicit access requirements."}, "workers": [{"credits_per_call": 0.841, "id": "worker_a", "p95_latency_ms": 26966, "pass_rate": 0.4902, "provider_group": "provider_x", "quality": 0.7685, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.982, "id": "worker_b", "p95_latency_ms": 80896, "pass_rate": 0.9142, "provider_group": "provider_y", "quality": 0.953, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.296, "id": "worker_c", "p95_latency_ms": 45768, "pass_rate": 0.7397, "provider_group": "provider_z", "quality": 0.8475, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.291, "max_latency_ms": 150487, "minimum_quality": 0.8685}, "family": "migration_planning", "id": "synthetic-migration_planning-001057", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.855, "estimated_latency_ms": 112608, "estimated_pass_rate": 0.8655, "estimated_quality": 0.8445, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.291, "max_latency_ms": 150487, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8655, "quality": 0.8445, "quality_target_met": false, "utility": 59.97998}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.291,\"max_latency_ms\":150487,\"minimum_quality\":0.8685},\"task\":{\"ambiguity\":0.6895,\"category\":\"coding\",\"complexity\":0.6383,\"consequence\":0.6533,\"evidence_need\":0.709,\"family\":\"migration_planning\",\"input_tokens\":6400,\"reversibility\":0.3143,\"risk\":0.3941,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.117,\"id\":\"worker_a\",\"p95_latency_ms\":77013,\"pass_rate\":0.7755,\"provider_group\":\"provider_z\",\"quality\":0.8659,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.788,\"id\":\"worker_b\",\"p95_latency_ms\":87544,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9121,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.067,\"id\":\"worker_c\",\"p95_latency_ms\":25064,\"pass_rate\":0.6028,\"provider_group\":\"provider_x\",\"quality\":0.7,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.855,\"estimated_latency_ms\":112608,\"estimated_pass_rate\":0.8655,\"estimated_quality\":0.8445,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.291,\"max_latency_ms\":150487,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6895, "category": "coding", "complexity": 0.6383, "consequence": 0.6533, "evidence_need": 0.709, "family": "migration_planning", "input_tokens": 6400, "request_id": "synthetic-migration_planning-001057", "reversibility": 0.3143, "risk": 0.3941, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 4.117, "id": "worker_a", "p95_latency_ms": 77013, "pass_rate": 0.7755, "provider_group": "provider_z", "quality": 0.8659, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.788, "id": "worker_b", "p95_latency_ms": 87544, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9121, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.067, "id": "worker_c", "p95_latency_ms": 25064, "pass_rate": 0.6028, "provider_group": "provider_x", "quality": 0.7, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.426, "max_latency_ms": 166368, "minimum_quality": 0.7403}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000045", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.036, "estimated_latency_ms": 118057, "estimated_pass_rate": 0.8922, "estimated_quality": 0.956, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.426, "max_latency_ms": 166368, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8922, "quality": 0.956, "quality_target_met": true, "utility": 74.496104}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.426,\"max_latency_ms\":166368,\"minimum_quality\":0.7403},\"task\":{\"ambiguity\":0.3869,\"category\":\"review\",\"complexity\":0.4712,\"consequence\":0.8762,\"evidence_need\":0.8639,\"family\":\"policy_compliance\",\"input_tokens\":3440,\"reversibility\":0.3622,\"risk\":0.6289,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.378,\"id\":\"worker_a\",\"p95_latency_ms\":56040,\"pass_rate\":0.9558,\"provider_group\":\"provider_y\",\"quality\":0.9873,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.92,\"id\":\"worker_b\",\"p95_latency_ms\":19363,\"pass_rate\":0.5531,\"provider_group\":\"provider_x\",\"quality\":0.7532,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.658,\"id\":\"worker_c\",\"p95_latency_ms\":62017,\"pass_rate\":0.8385,\"provider_group\":\"provider_z\",\"quality\":0.8332,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.036,\"estimated_latency_ms\":118057,\"estimated_pass_rate\":0.8922,\"estimated_quality\":0.956,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.426,\"max_latency_ms\":166368,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3869, "category": "review", "complexity": 0.4712, "consequence": 0.8762, "evidence_need": 0.8639, "family": "policy_compliance", "input_tokens": 3440, "request_id": "synthetic-policy_compliance-000045", "reversibility": 0.3622, "risk": 0.6289, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.378, "id": "worker_a", "p95_latency_ms": 56040, "pass_rate": 0.9558, "provider_group": "provider_y", "quality": 0.9873, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.92, "id": "worker_b", "p95_latency_ms": 19363, "pass_rate": 0.5531, "provider_group": "provider_x", "quality": 0.7532, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.658, "id": "worker_c", "p95_latency_ms": 62017, "pass_rate": 0.8385, "provider_group": "provider_z", "quality": 0.8332, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.789, "max_latency_ms": 33269, "minimum_quality": 0.5889}, "family": "migration_planning", "id": "synthetic-migration_planning-000316", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.827, "estimated_latency_ms": 31602, "estimated_pass_rate": 0.4165, "estimated_quality": 0.6017, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.789, "max_latency_ms": 33269, "policy": "lean", "reason_codes": ["latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4165, "quality": 0.6017, "quality_target_met": true, "utility": 41.082536}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.789,\"max_latency_ms\":33269,\"minimum_quality\":0.5889},\"task\":{\"ambiguity\":0.5259,\"category\":\"coding\",\"complexity\":0.6501,\"consequence\":0.6956,\"evidence_need\":0.5707,\"family\":\"migration_planning\",\"input_tokens\":1120,\"reversibility\":0.304,\"risk\":0.5454,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.086,\"id\":\"worker_a\",\"p95_latency_ms\":64312,\"pass_rate\":0.886,\"provider_group\":\"provider_x\",\"quality\":0.9225,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.827,\"id\":\"worker_b\",\"p95_latency_ms\":31602,\"pass_rate\":0.5742,\"provider_group\":\"provider_y\",\"quality\":0.7138,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.095,\"id\":\"worker_c\",\"p95_latency_ms\":50220,\"pass_rate\":0.92,\"provider_group\":\"provider_z\",\"quality\":0.9931,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.827,\"estimated_latency_ms\":31602,\"estimated_pass_rate\":0.4165,\"estimated_quality\":0.6017,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.789,\"max_latency_ms\":33269,\"policy\":\"lean\",\"reason_codes\":[\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5259, "category": "coding", "complexity": 0.6501, "consequence": 0.6956, "evidence_need": 0.5707, "family": "migration_planning", "input_tokens": 1120, "request_id": "synthetic-migration_planning-000316", "reversibility": 0.304, "risk": 0.5454, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 3.086, "id": "worker_a", "p95_latency_ms": 64312, "pass_rate": 0.886, "provider_group": "provider_x", "quality": 0.9225, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.827, "id": "worker_b", "p95_latency_ms": 31602, "pass_rate": 0.5742, "provider_group": "provider_y", "quality": 0.7138, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.095, "id": "worker_c", "p95_latency_ms": 50220, "pass_rate": 0.92, "provider_group": "provider_z", "quality": 0.9931, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 12.844, "max_latency_ms": 101991, "minimum_quality": 0.819}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000873", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.961, "estimated_latency_ms": 72502, "estimated_pass_rate": 0.6935, "estimated_quality": 0.846, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.844, "max_latency_ms": 101991, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6935, "quality": 0.846, "quality_target_met": true, "utility": 69.567341}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":12.844,\"max_latency_ms\":101991,\"minimum_quality\":0.819},\"task\":{\"ambiguity\":0.6603,\"category\":\"review\",\"complexity\":0.7716,\"consequence\":0.868,\"evidence_need\":0.9457,\"family\":\"policy_compliance\",\"input_tokens\":8000,\"reversibility\":0.4497,\"risk\":0.7895,\"summary\":\"Review whether model routing satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.158,\"id\":\"worker_a\",\"p95_latency_ms\":65064,\"pass_rate\":0.9262,\"provider_group\":\"provider_z\",\"quality\":0.9158,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.058,\"id\":\"worker_b\",\"p95_latency_ms\":23568,\"pass_rate\":0.617,\"provider_group\":\"provider_y\",\"quality\":0.7468,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.903,\"id\":\"worker_c\",\"p95_latency_ms\":48934,\"pass_rate\":0.8873,\"provider_group\":\"provider_x\",\"quality\":0.854,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.961,\"estimated_latency_ms\":72502,\"estimated_pass_rate\":0.6935,\"estimated_quality\":0.846,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.844,\"max_latency_ms\":101991,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6603, "category": "review", "complexity": 0.7716, "consequence": 0.868, "evidence_need": 0.9457, "family": "policy_compliance", "input_tokens": 8000, "request_id": "synthetic-policy_compliance-000873", "reversibility": 0.4497, "risk": 0.7895, "summary": "Review whether model routing satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.158, "id": "worker_a", "p95_latency_ms": 65064, "pass_rate": 0.9262, "provider_group": "provider_z", "quality": 0.9158, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.058, "id": "worker_b", "p95_latency_ms": 23568, "pass_rate": 0.617, "provider_group": "provider_y", "quality": 0.7468, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.903, "id": "worker_c", "p95_latency_ms": 48934, "pass_rate": 0.8873, "provider_group": "provider_x", "quality": 0.854, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.489, "max_latency_ms": 36727, "minimum_quality": 0.5433}, "family": "migration_planning", "id": "synthetic-migration_planning-000754", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.048, "estimated_latency_ms": 31305, "estimated_pass_rate": 0.2969, "estimated_quality": 0.639, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.489, "max_latency_ms": 36727, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2969, "quality": 0.639, "quality_target_met": true, "utility": 39.663666}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.489,\"max_latency_ms\":36727,\"minimum_quality\":0.5433},\"task\":{\"ambiguity\":0.3968,\"category\":\"coding\",\"complexity\":0.7732,\"consequence\":0.5807,\"evidence_need\":0.5272,\"family\":\"migration_planning\",\"input_tokens\":2920,\"reversibility\":0.544,\"risk\":0.6441,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":11.698,\"id\":\"worker_a\",\"p95_latency_ms\":59488,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.968,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.886,\"id\":\"worker_b\",\"p95_latency_ms\":58579,\"pass_rate\":0.7537,\"provider_group\":\"provider_y\",\"quality\":0.9061,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.048,\"id\":\"worker_c\",\"p95_latency_ms\":31305,\"pass_rate\":0.4668,\"provider_group\":\"provider_z\",\"quality\":0.7718,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.048,\"estimated_latency_ms\":31305,\"estimated_pass_rate\":0.2969,\"estimated_quality\":0.639,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.489,\"max_latency_ms\":36727,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3968, "category": "coding", "complexity": 0.7732, "consequence": 0.5807, "evidence_need": 0.5272, "family": "migration_planning", "input_tokens": 2920, "request_id": "synthetic-migration_planning-000754", "reversibility": 0.544, "risk": 0.6441, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 11.698, "id": "worker_a", "p95_latency_ms": 59488, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.968, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.886, "id": "worker_b", "p95_latency_ms": 58579, "pass_rate": 0.7537, "provider_group": "provider_y", "quality": 0.9061, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.048, "id": "worker_c", "p95_latency_ms": 31305, "pass_rate": 0.4668, "provider_group": "provider_z", "quality": 0.7718, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.178, "max_latency_ms": 36292, "minimum_quality": 0.7753}, "family": "migration_planning", "id": "synthetic-migration_planning-000286", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.906, "estimated_latency_ms": 32231, "estimated_pass_rate": 0.4184, "estimated_quality": 0.573, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.178, "max_latency_ms": 36292, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4184, "quality": 0.573, "quality_target_met": false, "utility": 22.244295}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.178,\"max_latency_ms\":36292,\"minimum_quality\":0.7753},\"task\":{\"ambiguity\":0.5798,\"category\":\"coding\",\"complexity\":0.8626,\"consequence\":0.8435,\"evidence_need\":0.4498,\"family\":\"migration_planning\",\"input_tokens\":3440,\"reversibility\":0.285,\"risk\":0.4882,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":0.906,\"id\":\"worker_a\",\"p95_latency_ms\":32231,\"pass_rate\":0.5996,\"provider_group\":\"provider_x\",\"quality\":0.7231,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.146,\"id\":\"worker_b\",\"p95_latency_ms\":63433,\"pass_rate\":0.9673,\"provider_group\":\"provider_y\",\"quality\":0.964,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.144,\"id\":\"worker_c\",\"p95_latency_ms\":78862,\"pass_rate\":0.8377,\"provider_group\":\"provider_z\",\"quality\":0.8164,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.906,\"estimated_latency_ms\":32231,\"estimated_pass_rate\":0.4184,\"estimated_quality\":0.573,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.178,\"max_latency_ms\":36292,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5798, "category": "coding", "complexity": 0.8626, "consequence": 0.8435, "evidence_need": 0.4498, "family": "migration_planning", "input_tokens": 3440, "request_id": "synthetic-migration_planning-000286", "reversibility": 0.285, "risk": 0.4882, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 0.906, "id": "worker_a", "p95_latency_ms": 32231, "pass_rate": 0.5996, "provider_group": "provider_x", "quality": 0.7231, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.146, "id": "worker_b", "p95_latency_ms": 63433, "pass_rate": 0.9673, "provider_group": "provider_y", "quality": 0.964, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.144, "id": "worker_c", "p95_latency_ms": 78862, "pass_rate": 0.8377, "provider_group": "provider_z", "quality": 0.8164, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.748, "max_latency_ms": 210580, "minimum_quality": 0.7915}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000159", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.128, "estimated_latency_ms": 75306, "estimated_pass_rate": 0.6328, "estimated_quality": 0.814, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.748, "max_latency_ms": 210580, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6328, "quality": 0.814, "quality_target_met": true, "utility": 61.374309}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.748,\"max_latency_ms\":210580,\"minimum_quality\":0.7915},\"task\":{\"ambiguity\":0.5011,\"category\":\"review\",\"complexity\":0.7901,\"consequence\":0.6989,\"evidence_need\":0.7539,\"family\":\"policy_compliance\",\"input_tokens\":3840,\"reversibility\":0.6249,\"risk\":0.7651,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":4.178,\"id\":\"worker_a\",\"p95_latency_ms\":49690,\"pass_rate\":0.8772,\"provider_group\":\"provider_z\",\"quality\":0.9099,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.95,\"id\":\"worker_b\",\"p95_latency_ms\":25616,\"pass_rate\":0.4627,\"provider_group\":\"provider_y\",\"quality\":0.6854,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.507,\"id\":\"worker_c\",\"p95_latency_ms\":64455,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9628,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.128,\"estimated_latency_ms\":75306,\"estimated_pass_rate\":0.6328,\"estimated_quality\":0.814,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.748,\"max_latency_ms\":210580,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5011, "category": "review", "complexity": 0.7901, "consequence": 0.6989, "evidence_need": 0.7539, "family": "policy_compliance", "input_tokens": 3840, "request_id": "synthetic-policy_compliance-000159", "reversibility": 0.6249, "risk": 0.7651, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 4.178, "id": "worker_a", "p95_latency_ms": 49690, "pass_rate": 0.8772, "provider_group": "provider_z", "quality": 0.9099, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.95, "id": "worker_b", "p95_latency_ms": 25616, "pass_rate": 0.4627, "provider_group": "provider_y", "quality": 0.6854, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.507, "id": "worker_c", "p95_latency_ms": 64455, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9628, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.604, "max_latency_ms": 29408, "minimum_quality": 0.6714}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000831", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.186, "estimated_latency_ms": 20632, "estimated_pass_rate": 0.3841, "estimated_quality": 0.6985, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.604, "max_latency_ms": 29408, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3841, "quality": 0.6985, "quality_target_met": true, "utility": 43.808389}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.604,\"max_latency_ms\":29408,\"minimum_quality\":0.6714},\"task\":{\"ambiguity\":0.6069,\"category\":\"review\",\"complexity\":0.4782,\"consequence\":0.5347,\"evidence_need\":0.5857,\"family\":\"policy_compliance\",\"input_tokens\":3720,\"reversibility\":0.3853,\"risk\":0.5345,\"summary\":\"Review whether tenant-scoped caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.186,\"id\":\"worker_a\",\"p95_latency_ms\":20632,\"pass_rate\":0.5275,\"provider_group\":\"provider_y\",\"quality\":0.7842,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.406,\"id\":\"worker_b\",\"p95_latency_ms\":73860,\"pass_rate\":0.9011,\"provider_group\":\"provider_x\",\"quality\":0.9439,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.586,\"id\":\"worker_c\",\"p95_latency_ms\":65322,\"pass_rate\":0.8003,\"provider_group\":\"provider_z\",\"quality\":0.7973,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.186,\"estimated_latency_ms\":20632,\"estimated_pass_rate\":0.3841,\"estimated_quality\":0.6985,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.604,\"max_latency_ms\":29408,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6069, "category": "review", "complexity": 0.4782, "consequence": 0.5347, "evidence_need": 0.5857, "family": "policy_compliance", "input_tokens": 3720, "request_id": "synthetic-policy_compliance-000831", "reversibility": 0.3853, "risk": 0.5345, "summary": "Review whether tenant-scoped caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.186, "id": "worker_a", "p95_latency_ms": 20632, "pass_rate": 0.5275, "provider_group": "provider_y", "quality": 0.7842, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.406, "id": "worker_b", "p95_latency_ms": 73860, "pass_rate": 0.9011, "provider_group": "provider_x", "quality": 0.9439, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.586, "id": "worker_c", "p95_latency_ms": 65322, "pass_rate": 0.8003, "provider_group": "provider_z", "quality": 0.7973, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.319, "max_latency_ms": 178576, "minimum_quality": 0.9064}, "family": "migration_planning", "id": "synthetic-migration_planning-000769", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.773, "estimated_latency_ms": 159978, "estimated_pass_rate": 0.7212, "estimated_quality": 0.9106, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.319, "max_latency_ms": 178576, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7212, "quality": 0.9106, "quality_target_met": true, "utility": 72.463172}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.319,\"max_latency_ms\":178576,\"minimum_quality\":0.9064},\"task\":{\"ambiguity\":0.3446,\"category\":\"coding\",\"complexity\":0.8641,\"consequence\":0.4552,\"evidence_need\":0.763,\"family\":\"migration_planning\",\"input_tokens\":2680,\"reversibility\":0.5856,\"risk\":0.577,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":1.068,\"id\":\"worker_a\",\"p95_latency_ms\":26430,\"pass_rate\":0.5151,\"provider_group\":\"provider_y\",\"quality\":0.6854,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.704,\"id\":\"worker_b\",\"p95_latency_ms\":51828,\"pass_rate\":0.9051,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.001,\"id\":\"worker_c\",\"p95_latency_ms\":81720,\"pass_rate\":0.7945,\"provider_group\":\"provider_z\",\"quality\":0.9128,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.773,\"estimated_latency_ms\":159978,\"estimated_pass_rate\":0.7212,\"estimated_quality\":0.9106,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.319,\"max_latency_ms\":178576,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3446, "category": "coding", "complexity": 0.8641, "consequence": 0.4552, "evidence_need": 0.763, "family": "migration_planning", "input_tokens": 2680, "request_id": "synthetic-migration_planning-000769", "reversibility": 0.5856, "risk": 0.577, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 1.068, "id": "worker_a", "p95_latency_ms": 26430, "pass_rate": 0.5151, "provider_group": "provider_y", "quality": 0.6854, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.704, "id": "worker_b", "p95_latency_ms": 51828, "pass_rate": 0.9051, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.001, "id": "worker_c", "p95_latency_ms": 81720, "pass_rate": 0.7945, "provider_group": "provider_z", "quality": 0.9128, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.012, "max_latency_ms": 49973, "minimum_quality": 0.5212}, "family": "migration_planning", "id": "synthetic-migration_planning-000277", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.932, "estimated_latency_ms": 29267, "estimated_pass_rate": 0.4139, "estimated_quality": 0.6048, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.012, "max_latency_ms": 49973, "policy": "lean", "reason_codes": ["credit_constrained", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4139, "quality": 0.6048, "quality_target_met": true, "utility": 37.935207}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.012,\"max_latency_ms\":49973,\"minimum_quality\":0.5212},\"task\":{\"ambiguity\":0.513,\"category\":\"coding\",\"complexity\":0.6851,\"consequence\":0.4317,\"evidence_need\":0.7079,\"family\":\"migration_planning\",\"input_tokens\":7240,\"reversibility\":0.3877,\"risk\":0.513,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.753,\"id\":\"worker_a\",\"p95_latency_ms\":60098,\"pass_rate\":0.7231,\"provider_group\":\"provider_z\",\"quality\":0.8295,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.932,\"id\":\"worker_b\",\"p95_latency_ms\":29267,\"pass_rate\":0.5719,\"provider_group\":\"provider_y\",\"quality\":0.7201,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.834,\"id\":\"worker_c\",\"p95_latency_ms\":60031,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9688,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.932,\"estimated_latency_ms\":29267,\"estimated_pass_rate\":0.4139,\"estimated_quality\":0.6048,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.012,\"max_latency_ms\":49973,\"policy\":\"lean\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.513, "category": "coding", "complexity": 0.6851, "consequence": 0.4317, "evidence_need": 0.7079, "family": "migration_planning", "input_tokens": 7240, "request_id": "synthetic-migration_planning-000277", "reversibility": 0.3877, "risk": 0.513, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 2.753, "id": "worker_a", "p95_latency_ms": 60098, "pass_rate": 0.7231, "provider_group": "provider_z", "quality": 0.8295, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.932, "id": "worker_b", "p95_latency_ms": 29267, "pass_rate": 0.5719, "provider_group": "provider_y", "quality": 0.7201, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.834, "id": "worker_c", "p95_latency_ms": 60031, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9688, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.906, "max_latency_ms": 52314, "minimum_quality": 0.6299}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000398", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.063, "estimated_latency_ms": 31374, "estimated_pass_rate": 0.2697, "estimated_quality": 0.5785, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.906, "max_latency_ms": 52314, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2697, "quality": 0.5785, "quality_target_met": false, "utility": 35.307521}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.906,\"max_latency_ms\":52314,\"minimum_quality\":0.6299},\"task\":{\"ambiguity\":0.5887,\"category\":\"analysis\",\"complexity\":0.7956,\"consequence\":0.2967,\"evidence_need\":0.6141,\"family\":\"performance_tuning\",\"input_tokens\":3200,\"reversibility\":0.6048,\"risk\":0.456,\"summary\":\"Review a proposed optimization for background jobs for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.638,\"id\":\"worker_a\",\"p95_latency_ms\":76497,\"pass_rate\":0.849,\"provider_group\":\"provider_y\",\"quality\":0.8719,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.633,\"id\":\"worker_b\",\"p95_latency_ms\":86950,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.063,\"id\":\"worker_c\",\"p95_latency_ms\":31374,\"pass_rate\":0.4414,\"provider_group\":\"provider_z\",\"quality\":0.7147,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.063,\"estimated_latency_ms\":31374,\"estimated_pass_rate\":0.2697,\"estimated_quality\":0.5785,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.906,\"max_latency_ms\":52314,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5887, "category": "analysis", "complexity": 0.7956, "consequence": 0.2967, "evidence_need": 0.6141, "family": "performance_tuning", "input_tokens": 3200, "request_id": "synthetic-performance_tuning-000398", "reversibility": 0.6048, "risk": 0.456, "summary": "Review a proposed optimization for background jobs for hidden regressions."}, "workers": [{"credits_per_call": 3.638, "id": "worker_a", "p95_latency_ms": 76497, "pass_rate": 0.849, "provider_group": "provider_y", "quality": 0.8719, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.633, "id": "worker_b", "p95_latency_ms": 86950, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.063, "id": "worker_c", "p95_latency_ms": 31374, "pass_rate": 0.4414, "provider_group": "provider_z", "quality": 0.7147, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.801, "max_latency_ms": 252631, "minimum_quality": 0.9057}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001248", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.512, "estimated_latency_ms": 176046, "estimated_pass_rate": 0.7605, "estimated_quality": 0.9231, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.801, "max_latency_ms": 252631, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7605, "quality": 0.9231, "quality_target_met": true, "utility": 75.661378}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.801,\"max_latency_ms\":252631,\"minimum_quality\":0.9057},\"task\":{\"ambiguity\":0.5375,\"category\":\"review\",\"complexity\":0.7042,\"consequence\":0.6185,\"evidence_need\":0.6542,\"family\":\"policy_compliance\",\"input_tokens\":8000,\"reversibility\":0.692,\"risk\":0.6403,\"summary\":\"Review whether model routing satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.155,\"id\":\"worker_a\",\"p95_latency_ms\":74964,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9876,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.831,\"id\":\"worker_b\",\"p95_latency_ms\":30751,\"pass_rate\":0.5034,\"provider_group\":\"provider_y\",\"quality\":0.7604,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.526,\"id\":\"worker_c\",\"p95_latency_ms\":70331,\"pass_rate\":0.8193,\"provider_group\":\"provider_z\",\"quality\":0.9037,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.512,\"estimated_latency_ms\":176046,\"estimated_pass_rate\":0.7605,\"estimated_quality\":0.9231,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.801,\"max_latency_ms\":252631,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5375, "category": "review", "complexity": 0.7042, "consequence": 0.6185, "evidence_need": 0.6542, "family": "policy_compliance", "input_tokens": 8000, "request_id": "synthetic-policy_compliance-001248", "reversibility": 0.692, "risk": 0.6403, "summary": "Review whether model routing satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.155, "id": "worker_a", "p95_latency_ms": 74964, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9876, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.831, "id": "worker_b", "p95_latency_ms": 30751, "pass_rate": 0.5034, "provider_group": "provider_y", "quality": 0.7604, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.526, "id": "worker_c", "p95_latency_ms": 70331, "pass_rate": 0.8193, "provider_group": "provider_z", "quality": 0.9037, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.846, "max_latency_ms": 195187, "minimum_quality": 0.8123}, "family": "migration_planning", "id": "synthetic-migration_planning-000574", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.566, "estimated_latency_ms": 160482, "estimated_pass_rate": 0.7625, "estimated_quality": 0.9112, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.846, "max_latency_ms": 195187, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7625, "quality": 0.9112, "quality_target_met": true, "utility": 73.867899}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.846,\"max_latency_ms\":195187,\"minimum_quality\":0.8123},\"task\":{\"ambiguity\":0.5839,\"category\":\"coding\",\"complexity\":0.747,\"consequence\":0.6639,\"evidence_need\":0.6397,\"family\":\"migration_planning\",\"input_tokens\":6520,\"reversibility\":0.3331,\"risk\":0.4942,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.791,\"id\":\"worker_a\",\"p95_latency_ms\":22469,\"pass_rate\":0.5429,\"provider_group\":\"provider_y\",\"quality\":0.6951,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.636,\"id\":\"worker_b\",\"p95_latency_ms\":77576,\"pass_rate\":0.826,\"provider_group\":\"provider_z\",\"quality\":0.9024,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.139,\"id\":\"worker_c\",\"p95_latency_ms\":60437,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9854,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.566,\"estimated_latency_ms\":160482,\"estimated_pass_rate\":0.7625,\"estimated_quality\":0.9112,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.846,\"max_latency_ms\":195187,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5839, "category": "coding", "complexity": 0.747, "consequence": 0.6639, "evidence_need": 0.6397, "family": "migration_planning", "input_tokens": 6520, "request_id": "synthetic-migration_planning-000574", "reversibility": 0.3331, "risk": 0.4942, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 0.791, "id": "worker_a", "p95_latency_ms": 22469, "pass_rate": 0.5429, "provider_group": "provider_y", "quality": 0.6951, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.636, "id": "worker_b", "p95_latency_ms": 77576, "pass_rate": 0.826, "provider_group": "provider_z", "quality": 0.9024, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.139, "id": "worker_c", "p95_latency_ms": 60437, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9854, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.002, "max_latency_ms": 34179, "minimum_quality": 0.6307}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000290", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.174, "estimated_latency_ms": 19872, "estimated_pass_rate": 0.3628, "estimated_quality": 0.5782, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.002, "max_latency_ms": 34179, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3628, "quality": 0.5782, "quality_target_met": false, "utility": 36.633948}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.002,\"max_latency_ms\":34179,\"minimum_quality\":0.6307},\"task\":{\"ambiguity\":0.7029,\"category\":\"analysis\",\"complexity\":0.7299,\"consequence\":0.4859,\"evidence_need\":0.6489,\"family\":\"performance_tuning\",\"input_tokens\":7200,\"reversibility\":0.7976,\"risk\":0.5359,\"summary\":\"Diagnose latency variance in private credentials and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":9.89,\"id\":\"worker_a\",\"p95_latency_ms\":85625,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.277,\"id\":\"worker_b\",\"p95_latency_ms\":50355,\"pass_rate\":0.897,\"provider_group\":\"provider_y\",\"quality\":0.8717,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.174,\"id\":\"worker_c\",\"p95_latency_ms\":19872,\"pass_rate\":0.5421,\"provider_group\":\"provider_z\",\"quality\":0.716,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.174,\"estimated_latency_ms\":19872,\"estimated_pass_rate\":0.3628,\"estimated_quality\":0.5782,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.002,\"max_latency_ms\":34179,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7029, "category": "analysis", "complexity": 0.7299, "consequence": 0.4859, "evidence_need": 0.6489, "family": "performance_tuning", "input_tokens": 7200, "request_id": "synthetic-performance_tuning-000290", "reversibility": 0.7976, "risk": 0.5359, "summary": "Diagnose latency variance in private credentials and rank the next experiments."}, "workers": [{"credits_per_call": 9.89, "id": "worker_a", "p95_latency_ms": 85625, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.277, "id": "worker_b", "p95_latency_ms": 50355, "pass_rate": 0.897, "provider_group": "provider_y", "quality": 0.8717, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.174, "id": "worker_c", "p95_latency_ms": 19872, "pass_rate": 0.5421, "provider_group": "provider_z", "quality": 0.716, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.914, "max_latency_ms": 46204, "minimum_quality": 0.7032}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000780", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.212, "estimated_latency_ms": 24026, "estimated_pass_rate": 0.2936, "estimated_quality": 0.6083, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.914, "max_latency_ms": 46204, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2936, "quality": 0.6083, "quality_target_met": false, "utility": 21.760475}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.914,\"max_latency_ms\":46204,\"minimum_quality\":0.7032},\"task\":{\"ambiguity\":0.7952,\"category\":\"review\",\"complexity\":0.7741,\"consequence\":0.6659,\"evidence_need\":0.7211,\"family\":\"policy_compliance\",\"input_tokens\":1720,\"reversibility\":0.368,\"risk\":0.8261,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.144,\"id\":\"worker_a\",\"p95_latency_ms\":53718,\"pass_rate\":0.8622,\"provider_group\":\"provider_x\",\"quality\":0.8005,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.212,\"id\":\"worker_b\",\"p95_latency_ms\":24026,\"pass_rate\":0.5084,\"provider_group\":\"provider_z\",\"quality\":0.7812,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.125,\"id\":\"worker_c\",\"p95_latency_ms\":87673,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9277,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.212,\"estimated_latency_ms\":24026,\"estimated_pass_rate\":0.2936,\"estimated_quality\":0.6083,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.914,\"max_latency_ms\":46204,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7952, "category": "review", "complexity": 0.7741, "consequence": 0.6659, "evidence_need": 0.7211, "family": "policy_compliance", "input_tokens": 1720, "request_id": "synthetic-policy_compliance-000780", "reversibility": 0.368, "risk": 0.8261, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.144, "id": "worker_a", "p95_latency_ms": 53718, "pass_rate": 0.8622, "provider_group": "provider_x", "quality": 0.8005, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.212, "id": "worker_b", "p95_latency_ms": 24026, "pass_rate": 0.5084, "provider_group": "provider_z", "quality": 0.7812, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.125, "id": "worker_c", "p95_latency_ms": 87673, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9277, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.642, "max_latency_ms": 120353, "minimum_quality": 0.8541}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001214", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.293, "estimated_latency_ms": 107801, "estimated_pass_rate": 0.9067, "estimated_quality": 0.9499, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.642, "max_latency_ms": 120353, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9067, "quality": 0.9499, "quality_target_met": true, "utility": 68.808229}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.642,\"max_latency_ms\":120353,\"minimum_quality\":0.8541},\"task\":{\"ambiguity\":0.6777,\"category\":\"analysis\",\"complexity\":0.6829,\"consequence\":0.3288,\"evidence_need\":0.6945,\"family\":\"performance_tuning\",\"input_tokens\":3680,\"reversibility\":0.7366,\"risk\":0.278,\"summary\":\"Review a proposed optimization for data retention for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.158,\"id\":\"worker_a\",\"p95_latency_ms\":56904,\"pass_rate\":0.9265,\"provider_group\":\"provider_y\",\"quality\":0.9592,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.135,\"id\":\"worker_b\",\"p95_latency_ms\":50897,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.088,\"id\":\"worker_c\",\"p95_latency_ms\":27917,\"pass_rate\":0.4075,\"provider_group\":\"provider_z\",\"quality\":0.7048,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.293,\"estimated_latency_ms\":107801,\"estimated_pass_rate\":0.9067,\"estimated_quality\":0.9499,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.642,\"max_latency_ms\":120353,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6777, "category": "analysis", "complexity": 0.6829, "consequence": 0.3288, "evidence_need": 0.6945, "family": "performance_tuning", "input_tokens": 3680, "request_id": "synthetic-performance_tuning-001214", "reversibility": 0.7366, "risk": 0.278, "summary": "Review a proposed optimization for data retention for hidden regressions."}, "workers": [{"credits_per_call": 3.158, "id": "worker_a", "p95_latency_ms": 56904, "pass_rate": 0.9265, "provider_group": "provider_y", "quality": 0.9592, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.135, "id": "worker_b", "p95_latency_ms": 50897, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.088, "id": "worker_c", "p95_latency_ms": 27917, "pass_rate": 0.4075, "provider_group": "provider_z", "quality": 0.7048, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.623, "max_latency_ms": 79059, "minimum_quality": 0.6842}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000894", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.0, "estimated_latency_ms": 71605, "estimated_pass_rate": 0.635, "estimated_quality": 0.7842, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.623, "max_latency_ms": 79059, "policy": "balanced", "reason_codes": ["independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.635, "quality": 0.7842, "quality_target_met": true, "utility": 53.921355}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.623,\"max_latency_ms\":79059,\"minimum_quality\":0.6842},\"task\":{\"ambiguity\":0.5621,\"category\":\"review\",\"complexity\":0.6957,\"consequence\":0.5025,\"evidence_need\":0.6363,\"family\":\"policy_compliance\",\"input_tokens\":1720,\"reversibility\":0.3119,\"risk\":0.6658,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.429,\"id\":\"worker_a\",\"p95_latency_ms\":82160,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9628,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.882,\"id\":\"worker_b\",\"p95_latency_ms\":46998,\"pass_rate\":0.7247,\"provider_group\":\"provider_x\",\"quality\":0.8673,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.118,\"id\":\"worker_c\",\"p95_latency_ms\":24607,\"pass_rate\":0.5028,\"provider_group\":\"provider_z\",\"quality\":0.6848,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.0,\"estimated_latency_ms\":71605,\"estimated_pass_rate\":0.635,\"estimated_quality\":0.7842,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.623,\"max_latency_ms\":79059,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5621, "category": "review", "complexity": 0.6957, "consequence": 0.5025, "evidence_need": 0.6363, "family": "policy_compliance", "input_tokens": 1720, "request_id": "synthetic-policy_compliance-000894", "reversibility": 0.3119, "risk": 0.6658, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 11.429, "id": "worker_a", "p95_latency_ms": 82160, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9628, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.882, "id": "worker_b", "p95_latency_ms": 46998, "pass_rate": 0.7247, "provider_group": "provider_x", "quality": 0.8673, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.118, "id": "worker_c", "p95_latency_ms": 24607, "pass_rate": 0.5028, "provider_group": "provider_z", "quality": 0.6848, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.548, "max_latency_ms": 205267, "minimum_quality": 0.8151}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000888", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.307, "estimated_latency_ms": 152314, "estimated_pass_rate": 0.7707, "estimated_quality": 0.9006, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.548, "max_latency_ms": 205267, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7707, "quality": 0.9006, "quality_target_met": true, "utility": 79.157866}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.548,\"max_latency_ms\":205267,\"minimum_quality\":0.8151},\"task\":{\"ambiguity\":0.7886,\"category\":\"review\",\"complexity\":0.8035,\"consequence\":0.7722,\"evidence_need\":0.8999,\"family\":\"policy_compliance\",\"input_tokens\":6040,\"reversibility\":0.6426,\"risk\":0.5416,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.818,\"id\":\"worker_a\",\"p95_latency_ms\":25499,\"pass_rate\":0.5892,\"provider_group\":\"provider_x\",\"quality\":0.6899,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.546,\"id\":\"worker_b\",\"p95_latency_ms\":79769,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9527,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.943,\"id\":\"worker_c\",\"p95_latency_ms\":47046,\"pass_rate\":0.8397,\"provider_group\":\"provider_y\",\"quality\":0.8648,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.307,\"estimated_latency_ms\":152314,\"estimated_pass_rate\":0.7707,\"estimated_quality\":0.9006,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.548,\"max_latency_ms\":205267,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7886, "category": "review", "complexity": 0.8035, "consequence": 0.7722, "evidence_need": 0.8999, "family": "policy_compliance", "input_tokens": 6040, "request_id": "synthetic-policy_compliance-000888", "reversibility": 0.6426, "risk": 0.5416, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 0.818, "id": "worker_a", "p95_latency_ms": 25499, "pass_rate": 0.5892, "provider_group": "provider_x", "quality": 0.6899, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.546, "id": "worker_b", "p95_latency_ms": 79769, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9527, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.943, "id": "worker_c", "p95_latency_ms": 47046, "pass_rate": 0.8397, "provider_group": "provider_y", "quality": 0.8648, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.695, "max_latency_ms": 49274, "minimum_quality": 0.6171}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000402", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.161, "estimated_latency_ms": 27001, "estimated_pass_rate": 0.2723, "estimated_quality": 0.5995, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.695, "max_latency_ms": 49274, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2723, "quality": 0.5995, "quality_target_met": false, "utility": 22.829577}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.695,\"max_latency_ms\":49274,\"minimum_quality\":0.6171},\"task\":{\"ambiguity\":0.3951,\"category\":\"review\",\"complexity\":0.8154,\"consequence\":0.7216,\"evidence_need\":0.8544,\"family\":\"policy_compliance\",\"input_tokens\":4280,\"reversibility\":0.4477,\"risk\":0.7617,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.632,\"id\":\"worker_a\",\"p95_latency_ms\":88293,\"pass_rate\":0.9635,\"provider_group\":\"provider_x\",\"quality\":0.9708,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.161,\"id\":\"worker_b\",\"p95_latency_ms\":27001,\"pass_rate\":0.4563,\"provider_group\":\"provider_z\",\"quality\":0.7483,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.166,\"id\":\"worker_c\",\"p95_latency_ms\":51751,\"pass_rate\":0.7482,\"provider_group\":\"provider_y\",\"quality\":0.8064,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.161,\"estimated_latency_ms\":27001,\"estimated_pass_rate\":0.2723,\"estimated_quality\":0.5995,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.695,\"max_latency_ms\":49274,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3951, "category": "review", "complexity": 0.8154, "consequence": 0.7216, "evidence_need": 0.8544, "family": "policy_compliance", "input_tokens": 4280, "request_id": "synthetic-policy_compliance-000402", "reversibility": 0.4477, "risk": 0.7617, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 12.632, "id": "worker_a", "p95_latency_ms": 88293, "pass_rate": 0.9635, "provider_group": "provider_x", "quality": 0.9708, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.161, "id": "worker_b", "p95_latency_ms": 27001, "pass_rate": 0.4563, "provider_group": "provider_z", "quality": 0.7483, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.166, "id": "worker_c", "p95_latency_ms": 51751, "pass_rate": 0.7482, "provider_group": "provider_y", "quality": 0.8064, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.897, "max_latency_ms": 174506, "minimum_quality": 0.7806}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000675", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.771, "estimated_latency_ms": 144638, "estimated_pass_rate": 0.7715, "estimated_quality": 0.9116, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.897, "max_latency_ms": 174506, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7715, "quality": 0.9116, "quality_target_met": true, "utility": 79.601469}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.897,\"max_latency_ms\":174506,\"minimum_quality\":0.7806},\"task\":{\"ambiguity\":0.7614,\"category\":\"review\",\"complexity\":0.5524,\"consequence\":0.8665,\"evidence_need\":0.7461,\"family\":\"policy_compliance\",\"input_tokens\":1080,\"reversibility\":0.4211,\"risk\":0.8172,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.1,\"id\":\"worker_a\",\"p95_latency_ms\":23115,\"pass_rate\":0.6124,\"provider_group\":\"provider_z\",\"quality\":0.7381,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.909,\"id\":\"worker_b\",\"p95_latency_ms\":59980,\"pass_rate\":0.7513,\"provider_group\":\"provider_x\",\"quality\":0.7983,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.762,\"id\":\"worker_c\",\"p95_latency_ms\":61543,\"pass_rate\":0.9793,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.771,\"estimated_latency_ms\":144638,\"estimated_pass_rate\":0.7715,\"estimated_quality\":0.9116,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.897,\"max_latency_ms\":174506,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7614, "category": "review", "complexity": 0.5524, "consequence": 0.8665, "evidence_need": 0.7461, "family": "policy_compliance", "input_tokens": 1080, "request_id": "synthetic-policy_compliance-000675", "reversibility": 0.4211, "risk": 0.8172, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 1.1, "id": "worker_a", "p95_latency_ms": 23115, "pass_rate": 0.6124, "provider_group": "provider_z", "quality": 0.7381, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.909, "id": "worker_b", "p95_latency_ms": 59980, "pass_rate": 0.7513, "provider_group": "provider_x", "quality": 0.7983, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.762, "id": "worker_c", "p95_latency_ms": 61543, "pass_rate": 0.9793, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.227, "max_latency_ms": 26828, "minimum_quality": 0.5209}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000636", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.161, "estimated_latency_ms": 22150, "estimated_pass_rate": 0.3987, "estimated_quality": 0.6966, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.227, "max_latency_ms": 26828, "policy": "lean", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3987, "quality": 0.6966, "quality_target_met": true, "utility": 42.020493}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.227,\"max_latency_ms\":26828,\"minimum_quality\":0.5209},\"task\":{\"ambiguity\":0.5479,\"category\":\"review\",\"complexity\":0.4757,\"consequence\":0.5246,\"evidence_need\":0.6754,\"family\":\"policy_compliance\",\"input_tokens\":3800,\"reversibility\":0.3552,\"risk\":0.7702,\"summary\":\"Audit a workflow for tenant-scoped caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.626,\"id\":\"worker_a\",\"p95_latency_ms\":78752,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9236,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.161,\"id\":\"worker_b\",\"p95_latency_ms\":22150,\"pass_rate\":0.5565,\"provider_group\":\"provider_z\",\"quality\":0.7949,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.955,\"id\":\"worker_c\",\"p95_latency_ms\":58716,\"pass_rate\":0.8308,\"provider_group\":\"provider_y\",\"quality\":0.8395,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.161,\"estimated_latency_ms\":22150,\"estimated_pass_rate\":0.3987,\"estimated_quality\":0.6966,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.227,\"max_latency_ms\":26828,\"policy\":\"lean\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5479, "category": "review", "complexity": 0.4757, "consequence": 0.5246, "evidence_need": 0.6754, "family": "policy_compliance", "input_tokens": 3800, "request_id": "synthetic-policy_compliance-000636", "reversibility": 0.3552, "risk": 0.7702, "summary": "Audit a workflow for tenant-scoped caching against explicit access requirements."}, "workers": [{"credits_per_call": 10.626, "id": "worker_a", "p95_latency_ms": 78752, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9236, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.161, "id": "worker_b", "p95_latency_ms": 22150, "pass_rate": 0.5565, "provider_group": "provider_z", "quality": 0.7949, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.955, "id": "worker_c", "p95_latency_ms": 58716, "pass_rate": 0.8308, "provider_group": "provider_y", "quality": 0.8395, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.082, "max_latency_ms": 227476, "minimum_quality": 0.9224}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000483", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.639, "estimated_latency_ms": 163547, "estimated_pass_rate": 0.809, "estimated_quality": 0.9227, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.082, "max_latency_ms": 227476, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.809, "quality": 0.9227, "quality_target_met": true, "utility": 76.951866}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.082,\"max_latency_ms\":227476,\"minimum_quality\":0.9224},\"task\":{\"ambiguity\":0.5168,\"category\":\"review\",\"complexity\":0.8136,\"consequence\":0.9415,\"evidence_need\":0.8366,\"family\":\"policy_compliance\",\"input_tokens\":7800,\"reversibility\":0.4497,\"risk\":0.6977,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.953,\"id\":\"worker_a\",\"p95_latency_ms\":30118,\"pass_rate\":0.5829,\"provider_group\":\"provider_x\",\"quality\":0.7431,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.765,\"id\":\"worker_b\",\"p95_latency_ms\":45103,\"pass_rate\":0.7817,\"provider_group\":\"provider_z\",\"quality\":0.8767,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.921,\"id\":\"worker_c\",\"p95_latency_ms\":88326,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.639,\"estimated_latency_ms\":163547,\"estimated_pass_rate\":0.809,\"estimated_quality\":0.9227,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.082,\"max_latency_ms\":227476,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5168, "category": "review", "complexity": 0.8136, "consequence": 0.9415, "evidence_need": 0.8366, "family": "policy_compliance", "input_tokens": 7800, "request_id": "synthetic-policy_compliance-000483", "reversibility": 0.4497, "risk": 0.6977, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.953, "id": "worker_a", "p95_latency_ms": 30118, "pass_rate": 0.5829, "provider_group": "provider_x", "quality": 0.7431, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.765, "id": "worker_b", "p95_latency_ms": 45103, "pass_rate": 0.7817, "provider_group": "provider_z", "quality": 0.8767, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.921, "id": "worker_c", "p95_latency_ms": 88326, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.747, "max_latency_ms": 163531, "minimum_quality": 0.8527}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000903", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.133, "estimated_latency_ms": 122117, "estimated_pass_rate": 0.7079, "estimated_quality": 0.8895, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.747, "max_latency_ms": 163531, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7079, "quality": 0.8895, "quality_target_met": true, "utility": 76.68148}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.747,\"max_latency_ms\":163531,\"minimum_quality\":0.8527},\"task\":{\"ambiguity\":0.7807,\"category\":\"review\",\"complexity\":0.669,\"consequence\":0.8064,\"evidence_need\":0.7548,\"family\":\"policy_compliance\",\"input_tokens\":2800,\"reversibility\":0.6977,\"risk\":0.6818,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.929,\"id\":\"worker_a\",\"p95_latency_ms\":49146,\"pass_rate\":0.7823,\"provider_group\":\"provider_x\",\"quality\":0.8009,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.196,\"id\":\"worker_b\",\"p95_latency_ms\":51658,\"pass_rate\":0.9549,\"provider_group\":\"provider_y\",\"quality\":0.9503,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.008,\"id\":\"worker_c\",\"p95_latency_ms\":21313,\"pass_rate\":0.4868,\"provider_group\":\"provider_z\",\"quality\":0.7328,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.133,\"estimated_latency_ms\":122117,\"estimated_pass_rate\":0.7079,\"estimated_quality\":0.8895,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.747,\"max_latency_ms\":163531,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7807, "category": "review", "complexity": 0.669, "consequence": 0.8064, "evidence_need": 0.7548, "family": "policy_compliance", "input_tokens": 2800, "request_id": "synthetic-policy_compliance-000903", "reversibility": 0.6977, "risk": 0.6818, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.929, "id": "worker_a", "p95_latency_ms": 49146, "pass_rate": 0.7823, "provider_group": "provider_x", "quality": 0.8009, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.196, "id": "worker_b", "p95_latency_ms": 51658, "pass_rate": 0.9549, "provider_group": "provider_y", "quality": 0.9503, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.008, "id": "worker_c", "p95_latency_ms": 21313, "pass_rate": 0.4868, "provider_group": "provider_z", "quality": 0.7328, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.275, "max_latency_ms": 230235, "minimum_quality": 0.8498}, "family": "migration_planning", "id": "synthetic-migration_planning-000871", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.338, "estimated_latency_ms": 95346, "estimated_pass_rate": 0.5848, "estimated_quality": 0.7812, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.275, "max_latency_ms": 230235, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5848, "quality": 0.7812, "quality_target_met": false, "utility": 53.26031}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.275,\"max_latency_ms\":230235,\"minimum_quality\":0.8498},\"task\":{\"ambiguity\":0.468,\"category\":\"coding\",\"complexity\":0.5782,\"consequence\":0.4418,\"evidence_need\":0.6359,\"family\":\"migration_planning\",\"input_tokens\":1240,\"reversibility\":0.3876,\"risk\":0.4818,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":12.904,\"id\":\"worker_a\",\"p95_latency_ms\":70908,\"pass_rate\":0.9062,\"provider_group\":\"provider_z\",\"quality\":0.9648,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.242,\"id\":\"worker_b\",\"p95_latency_ms\":64015,\"pass_rate\":0.8557,\"provider_group\":\"provider_y\",\"quality\":0.8146,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.096,\"id\":\"worker_c\",\"p95_latency_ms\":31331,\"pass_rate\":0.4869,\"provider_group\":\"provider_x\",\"quality\":0.6926,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.338,\"estimated_latency_ms\":95346,\"estimated_pass_rate\":0.5848,\"estimated_quality\":0.7812,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.275,\"max_latency_ms\":230235,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.468, "category": "coding", "complexity": 0.5782, "consequence": 0.4418, "evidence_need": 0.6359, "family": "migration_planning", "input_tokens": 1240, "request_id": "synthetic-migration_planning-000871", "reversibility": 0.3876, "risk": 0.4818, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 12.904, "id": "worker_a", "p95_latency_ms": 70908, "pass_rate": 0.9062, "provider_group": "provider_z", "quality": 0.9648, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.242, "id": "worker_b", "p95_latency_ms": 64015, "pass_rate": 0.8557, "provider_group": "provider_y", "quality": 0.8146, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.096, "id": "worker_c", "p95_latency_ms": 31331, "pass_rate": 0.4869, "provider_group": "provider_x", "quality": 0.6926, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.096, "max_latency_ms": 31782, "minimum_quality": 0.5613}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000300", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.951, "estimated_latency_ms": 20102, "estimated_pass_rate": 0.2687, "estimated_quality": 0.6382, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.096, "max_latency_ms": 31782, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2687, "quality": 0.6382, "quality_target_met": true, "utility": 39.726217}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.096,\"max_latency_ms\":31782,\"minimum_quality\":0.5613},\"task\":{\"ambiguity\":0.7054,\"category\":\"review\",\"complexity\":0.7977,\"consequence\":0.5543,\"evidence_need\":0.7762,\"family\":\"policy_compliance\",\"input_tokens\":7520,\"reversibility\":0.6024,\"risk\":0.5858,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":2.788,\"id\":\"worker_a\",\"p95_latency_ms\":51825,\"pass_rate\":0.7685,\"provider_group\":\"provider_z\",\"quality\":0.8288,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.951,\"id\":\"worker_b\",\"p95_latency_ms\":20102,\"pass_rate\":0.4599,\"provider_group\":\"provider_y\",\"quality\":0.792,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.407,\"id\":\"worker_c\",\"p95_latency_ms\":75895,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9312,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.951,\"estimated_latency_ms\":20102,\"estimated_pass_rate\":0.2687,\"estimated_quality\":0.6382,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.096,\"max_latency_ms\":31782,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7054, "category": "review", "complexity": 0.7977, "consequence": 0.5543, "evidence_need": 0.7762, "family": "policy_compliance", "input_tokens": 7520, "request_id": "synthetic-policy_compliance-000300", "reversibility": 0.6024, "risk": 0.5858, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 2.788, "id": "worker_a", "p95_latency_ms": 51825, "pass_rate": 0.7685, "provider_group": "provider_z", "quality": 0.8288, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.951, "id": "worker_b", "p95_latency_ms": 20102, "pass_rate": 0.4599, "provider_group": "provider_y", "quality": 0.792, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.407, "id": "worker_c", "p95_latency_ms": 75895, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9312, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.904, "max_latency_ms": 188973, "minimum_quality": 0.8081}, "family": "migration_planning", "id": "synthetic-migration_planning-001150", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.524, "estimated_latency_ms": 82963, "estimated_pass_rate": 0.5459, "estimated_quality": 0.7934, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.904, "max_latency_ms": 188973, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5459, "quality": 0.7934, "quality_target_met": false, "utility": 55.523637}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.904,\"max_latency_ms\":188973,\"minimum_quality\":0.8081},\"task\":{\"ambiguity\":0.6924,\"category\":\"coding\",\"complexity\":0.8499,\"consequence\":0.5607,\"evidence_need\":0.6363,\"family\":\"migration_planning\",\"input_tokens\":2160,\"reversibility\":0.2767,\"risk\":0.5747,\"summary\":\"Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.783,\"id\":\"worker_a\",\"p95_latency_ms\":25150,\"pass_rate\":0.4514,\"provider_group\":\"provider_y\",\"quality\":0.7743,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.804,\"id\":\"worker_b\",\"p95_latency_ms\":58335,\"pass_rate\":0.9173,\"provider_group\":\"provider_x\",\"quality\":0.9906,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.741,\"id\":\"worker_c\",\"p95_latency_ms\":57813,\"pass_rate\":0.8034,\"provider_group\":\"provider_z\",\"quality\":0.8742,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.524,\"estimated_latency_ms\":82963,\"estimated_pass_rate\":0.5459,\"estimated_quality\":0.7934,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.904,\"max_latency_ms\":188973,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6924, "category": "coding", "complexity": 0.8499, "consequence": 0.5607, "evidence_need": 0.6363, "family": "migration_planning", "input_tokens": 2160, "request_id": "synthetic-migration_planning-001150", "reversibility": 0.2767, "risk": 0.5747, "summary": "Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints."}, "workers": [{"credits_per_call": 0.783, "id": "worker_a", "p95_latency_ms": 25150, "pass_rate": 0.4514, "provider_group": "provider_y", "quality": 0.7743, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.804, "id": "worker_b", "p95_latency_ms": 58335, "pass_rate": 0.9173, "provider_group": "provider_x", "quality": 0.9906, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.741, "id": "worker_c", "p95_latency_ms": 57813, "pass_rate": 0.8034, "provider_group": "provider_z", "quality": 0.8742, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.173, "max_latency_ms": 40790, "minimum_quality": 0.7099}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000144", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.957, "estimated_latency_ms": 22294, "estimated_pass_rate": 0.3668, "estimated_quality": 0.6332, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.173, "max_latency_ms": 40790, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3668, "quality": 0.6332, "quality_target_met": false, "utility": 27.559822}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.173,\"max_latency_ms\":40790,\"minimum_quality\":0.7099},\"task\":{\"ambiguity\":0.5122,\"category\":\"review\",\"complexity\":0.4052,\"consequence\":0.9173,\"evidence_need\":0.5924,\"family\":\"policy_compliance\",\"input_tokens\":2960,\"reversibility\":0.4369,\"risk\":0.5885,\"summary\":\"Review whether database failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.957,\"id\":\"worker_a\",\"p95_latency_ms\":22294,\"pass_rate\":0.4991,\"provider_group\":\"provider_y\",\"quality\":0.7033,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.206,\"id\":\"worker_b\",\"p95_latency_ms\":44438,\"pass_rate\":0.7362,\"provider_group\":\"provider_z\",\"quality\":0.8349,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.976,\"id\":\"worker_c\",\"p95_latency_ms\":61785,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.959,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.957,\"estimated_latency_ms\":22294,\"estimated_pass_rate\":0.3668,\"estimated_quality\":0.6332,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.173,\"max_latency_ms\":40790,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5122, "category": "review", "complexity": 0.4052, "consequence": 0.9173, "evidence_need": 0.5924, "family": "policy_compliance", "input_tokens": 2960, "request_id": "synthetic-policy_compliance-000144", "reversibility": 0.4369, "risk": 0.5885, "summary": "Review whether database failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.957, "id": "worker_a", "p95_latency_ms": 22294, "pass_rate": 0.4991, "provider_group": "provider_y", "quality": 0.7033, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.206, "id": "worker_b", "p95_latency_ms": 44438, "pass_rate": 0.7362, "provider_group": "provider_z", "quality": 0.8349, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.976, "id": "worker_c", "p95_latency_ms": 61785, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.959, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.11, "max_latency_ms": 174475, "minimum_quality": 0.8939}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000044", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.334, "estimated_latency_ms": 140561, "estimated_pass_rate": 0.7368, "estimated_quality": 0.9181, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.11, "max_latency_ms": 174475, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7368, "quality": 0.9181, "quality_target_met": true, "utility": 72.770042}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.11,\"max_latency_ms\":174475,\"minimum_quality\":0.8939},\"task\":{\"ambiguity\":0.6057,\"category\":\"analysis\",\"complexity\":0.8791,\"consequence\":0.6478,\"evidence_need\":0.6666,\"family\":\"performance_tuning\",\"input_tokens\":7520,\"reversibility\":0.8173,\"risk\":0.2767,\"summary\":\"Diagnose latency variance in usage metering and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":4.027,\"id\":\"worker_a\",\"p95_latency_ms\":51077,\"pass_rate\":0.9457,\"provider_group\":\"provider_y\",\"quality\":0.8601,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.107,\"id\":\"worker_b\",\"p95_latency_ms\":27111,\"pass_rate\":0.4999,\"provider_group\":\"provider_x\",\"quality\":0.6723,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.2,\"id\":\"worker_c\",\"p95_latency_ms\":62373,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.334,\"estimated_latency_ms\":140561,\"estimated_pass_rate\":0.7368,\"estimated_quality\":0.9181,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.11,\"max_latency_ms\":174475,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6057, "category": "analysis", "complexity": 0.8791, "consequence": 0.6478, "evidence_need": 0.6666, "family": "performance_tuning", "input_tokens": 7520, "request_id": "synthetic-performance_tuning-000044", "reversibility": 0.8173, "risk": 0.2767, "summary": "Diagnose latency variance in usage metering and rank the next experiments."}, "workers": [{"credits_per_call": 4.027, "id": "worker_a", "p95_latency_ms": 51077, "pass_rate": 0.9457, "provider_group": "provider_y", "quality": 0.8601, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.107, "id": "worker_b", "p95_latency_ms": 27111, "pass_rate": 0.4999, "provider_group": "provider_x", "quality": 0.6723, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.2, "id": "worker_c", "p95_latency_ms": 62373, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.882, "max_latency_ms": 215688, "minimum_quality": 0.8478}, "family": "migration_planning", "id": "synthetic-migration_planning-001039", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.213, "estimated_latency_ms": 86631, "estimated_pass_rate": 0.6352, "estimated_quality": 0.7942, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.882, "max_latency_ms": 215688, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6352, "quality": 0.7942, "quality_target_met": false, "utility": 61.772193}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.882,\"max_latency_ms\":215688,\"minimum_quality\":0.8478},\"task\":{\"ambiguity\":0.5057,\"category\":\"coding\",\"complexity\":0.7254,\"consequence\":0.7328,\"evidence_need\":0.7813,\"family\":\"migration_planning\",\"input_tokens\":760,\"reversibility\":0.5947,\"risk\":0.707,\"summary\":\"Design compatibility testing for a credit reservation schema migration.\"},\"workers\":[{\"credits_per_call\":3.142,\"id\":\"worker_a\",\"p95_latency_ms\":68989,\"pass_rate\":0.7423,\"provider_group\":\"provider_y\",\"quality\":0.8364,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.801,\"id\":\"worker_b\",\"p95_latency_ms\":90065,\"pass_rate\":0.9007,\"provider_group\":\"provider_z\",\"quality\":0.9282,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.071,\"id\":\"worker_c\",\"p95_latency_ms\":17642,\"pass_rate\":0.5978,\"provider_group\":\"provider_x\",\"quality\":0.7551,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.213,\"estimated_latency_ms\":86631,\"estimated_pass_rate\":0.6352,\"estimated_quality\":0.7942,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.882,\"max_latency_ms\":215688,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5057, "category": "coding", "complexity": 0.7254, "consequence": 0.7328, "evidence_need": 0.7813, "family": "migration_planning", "input_tokens": 760, "request_id": "synthetic-migration_planning-001039", "reversibility": 0.5947, "risk": 0.707, "summary": "Design compatibility testing for a credit reservation schema migration."}, "workers": [{"credits_per_call": 3.142, "id": "worker_a", "p95_latency_ms": 68989, "pass_rate": 0.7423, "provider_group": "provider_y", "quality": 0.8364, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 13.801, "id": "worker_b", "p95_latency_ms": 90065, "pass_rate": 0.9007, "provider_group": "provider_z", "quality": 0.9282, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.071, "id": "worker_c", "p95_latency_ms": 17642, "pass_rate": 0.5978, "provider_group": "provider_x", "quality": 0.7551, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.517, "max_latency_ms": 147882, "minimum_quality": 0.6885}, "family": "migration_planning", "id": "synthetic-migration_planning-000322", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.891, "estimated_latency_ms": 92696, "estimated_pass_rate": 0.5796, "estimated_quality": 0.8329, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.517, "max_latency_ms": 147882, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.5796, "quality": 0.8329, "quality_target_met": true, "utility": 60.744713}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.517,\"max_latency_ms\":147882,\"minimum_quality\":0.6885},\"task\":{\"ambiguity\":0.6383,\"category\":\"coding\",\"complexity\":0.7436,\"consequence\":0.4201,\"evidence_need\":0.619,\"family\":\"migration_planning\",\"input_tokens\":1600,\"reversibility\":0.388,\"risk\":0.4175,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.033,\"id\":\"worker_a\",\"p95_latency_ms\":69371,\"pass_rate\":0.8548,\"provider_group\":\"provider_x\",\"quality\":0.9057,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.742,\"id\":\"worker_b\",\"p95_latency_ms\":64230,\"pass_rate\":0.9287,\"provider_group\":\"provider_y\",\"quality\":0.9226,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.858,\"id\":\"worker_c\",\"p95_latency_ms\":23325,\"pass_rate\":0.4793,\"provider_group\":\"provider_z\",\"quality\":0.8018,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.891,\"estimated_latency_ms\":92696,\"estimated_pass_rate\":0.5796,\"estimated_quality\":0.8329,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.517,\"max_latency_ms\":147882,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6383, "category": "coding", "complexity": 0.7436, "consequence": 0.4201, "evidence_need": 0.619, "family": "migration_planning", "input_tokens": 1600, "request_id": "synthetic-migration_planning-000322", "reversibility": 0.388, "risk": 0.4175, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 3.033, "id": "worker_a", "p95_latency_ms": 69371, "pass_rate": 0.8548, "provider_group": "provider_x", "quality": 0.9057, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.742, "id": "worker_b", "p95_latency_ms": 64230, "pass_rate": 0.9287, "provider_group": "provider_y", "quality": 0.9226, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.858, "id": "worker_c", "p95_latency_ms": 23325, "pass_rate": 0.4793, "provider_group": "provider_z", "quality": 0.8018, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.709, "max_latency_ms": 35691, "minimum_quality": 0.5079}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000120", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.917, "estimated_latency_ms": 21473, "estimated_pass_rate": 0.4525, "estimated_quality": 0.5988, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.709, "max_latency_ms": 35691, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4525, "quality": 0.5988, "quality_target_met": true, "utility": 42.456228}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.709,\"max_latency_ms\":35691,\"minimum_quality\":0.5079},\"task\":{\"ambiguity\":0.6457,\"category\":\"review\",\"complexity\":0.6182,\"consequence\":0.663,\"evidence_need\":0.6393,\"family\":\"policy_compliance\",\"input_tokens\":2440,\"reversibility\":0.5286,\"risk\":0.5702,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.883,\"id\":\"worker_a\",\"p95_latency_ms\":52076,\"pass_rate\":0.8958,\"provider_group\":\"provider_y\",\"quality\":0.9204,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.917,\"id\":\"worker_b\",\"p95_latency_ms\":21473,\"pass_rate\":0.6176,\"provider_group\":\"provider_x\",\"quality\":0.715,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.018,\"id\":\"worker_c\",\"p95_latency_ms\":69535,\"pass_rate\":0.8639,\"provider_group\":\"provider_z\",\"quality\":0.8534,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.917,\"estimated_latency_ms\":21473,\"estimated_pass_rate\":0.4525,\"estimated_quality\":0.5988,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.709,\"max_latency_ms\":35691,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6457, "category": "review", "complexity": 0.6182, "consequence": 0.663, "evidence_need": 0.6393, "family": "policy_compliance", "input_tokens": 2440, "request_id": "synthetic-policy_compliance-000120", "reversibility": 0.5286, "risk": 0.5702, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 12.883, "id": "worker_a", "p95_latency_ms": 52076, "pass_rate": 0.8958, "provider_group": "provider_y", "quality": 0.9204, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.917, "id": "worker_b", "p95_latency_ms": 21473, "pass_rate": 0.6176, "provider_group": "provider_x", "quality": 0.715, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.018, "id": "worker_c", "p95_latency_ms": 69535, "pass_rate": 0.8639, "provider_group": "provider_z", "quality": 0.8534, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.821, "max_latency_ms": 25419, "minimum_quality": 0.5417}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000020", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.05, "estimated_latency_ms": 20723, "estimated_pass_rate": 0.2723, "estimated_quality": 0.6228, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.821, "max_latency_ms": 25419, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2723, "quality": 0.6228, "quality_target_met": true, "utility": 37.171603}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.821,\"max_latency_ms\":25419,\"minimum_quality\":0.5417},\"task\":{\"ambiguity\":0.6089,\"category\":\"analysis\",\"complexity\":0.4617,\"consequence\":0.2529,\"evidence_need\":0.4329,\"family\":\"performance_tuning\",\"input_tokens\":640,\"reversibility\":0.8646,\"risk\":0.4923,\"summary\":\"Review a proposed optimization for access control for hidden regressions.\"},\"workers\":[{\"credits_per_call\":12.248,\"id\":\"worker_a\",\"p95_latency_ms\":68470,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9779,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.05,\"id\":\"worker_b\",\"p95_latency_ms\":20723,\"pass_rate\":0.4106,\"provider_group\":\"provider_z\",\"quality\":0.7026,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.847,\"id\":\"worker_c\",\"p95_latency_ms\":50147,\"pass_rate\":0.9094,\"provider_group\":\"provider_y\",\"quality\":0.9004,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.05,\"estimated_latency_ms\":20723,\"estimated_pass_rate\":0.2723,\"estimated_quality\":0.6228,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.821,\"max_latency_ms\":25419,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6089, "category": "analysis", "complexity": 0.4617, "consequence": 0.2529, "evidence_need": 0.4329, "family": "performance_tuning", "input_tokens": 640, "request_id": "synthetic-performance_tuning-000020", "reversibility": 0.8646, "risk": 0.4923, "summary": "Review a proposed optimization for access control for hidden regressions."}, "workers": [{"credits_per_call": 12.248, "id": "worker_a", "p95_latency_ms": 68470, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9779, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.05, "id": "worker_b", "p95_latency_ms": 20723, "pass_rate": 0.4106, "provider_group": "provider_z", "quality": 0.7026, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.847, "id": "worker_c", "p95_latency_ms": 50147, "pass_rate": 0.9094, "provider_group": "provider_y", "quality": 0.9004, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.801, "max_latency_ms": 222158, "minimum_quality": 0.8831}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000900", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.379, "estimated_latency_ms": 158006, "estimated_pass_rate": 0.7564, "estimated_quality": 0.8945, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.801, "max_latency_ms": 222158, "policy": "strict", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7564, "quality": 0.8945, "quality_target_met": true, "utility": 77.248838}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.801,\"max_latency_ms\":222158,\"minimum_quality\":0.8831},\"task\":{\"ambiguity\":0.6476,\"category\":\"review\",\"complexity\":0.6102,\"consequence\":0.763,\"evidence_need\":0.6281,\"family\":\"policy_compliance\",\"input_tokens\":4320,\"reversibility\":0.4586,\"risk\":0.8267,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":2.84,\"id\":\"worker_a\",\"p95_latency_ms\":63576,\"pass_rate\":0.7755,\"provider_group\":\"provider_z\",\"quality\":0.8015,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.903,\"id\":\"worker_b\",\"p95_latency_ms\":31952,\"pass_rate\":0.5387,\"provider_group\":\"provider_y\",\"quality\":0.7361,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.636,\"id\":\"worker_c\",\"p95_latency_ms\":62478,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9624,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.379,\"estimated_latency_ms\":158006,\"estimated_pass_rate\":0.7564,\"estimated_quality\":0.8945,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.801,\"max_latency_ms\":222158,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6476, "category": "review", "complexity": 0.6102, "consequence": 0.763, "evidence_need": 0.6281, "family": "policy_compliance", "input_tokens": 4320, "request_id": "synthetic-policy_compliance-000900", "reversibility": 0.4586, "risk": 0.8267, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 2.84, "id": "worker_a", "p95_latency_ms": 63576, "pass_rate": 0.7755, "provider_group": "provider_z", "quality": 0.8015, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.903, "id": "worker_b", "p95_latency_ms": 31952, "pass_rate": 0.5387, "provider_group": "provider_y", "quality": 0.7361, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.636, "id": "worker_c", "p95_latency_ms": 62478, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9624, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 4.607, "max_latency_ms": 137638, "minimum_quality": 0.7378}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000393", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.21, "estimated_latency_ms": 73432, "estimated_pass_rate": 0.5413, "estimated_quality": 0.7881, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.607, "max_latency_ms": 137638, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5413, "quality": 0.7881, "quality_target_met": true, "utility": 57.354409}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":4.607,\"max_latency_ms\":137638,\"minimum_quality\":0.7378},\"task\":{\"ambiguity\":0.4771,\"category\":\"review\",\"complexity\":0.7872,\"consequence\":0.8249,\"evidence_need\":0.9336,\"family\":\"policy_compliance\",\"input_tokens\":3080,\"reversibility\":0.619,\"risk\":0.6602,\"summary\":\"Identify evidence required before approving usage metering for production.\"},\"workers\":[{\"credits_per_call\":3.2,\"id\":\"worker_a\",\"p95_latency_ms\":46231,\"pass_rate\":0.7227,\"provider_group\":\"provider_x\",\"quality\":0.8351,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.021,\"id\":\"worker_b\",\"p95_latency_ms\":88633,\"pass_rate\":0.9722,\"provider_group\":\"provider_z\",\"quality\":0.9794,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.01,\"id\":\"worker_c\",\"p95_latency_ms\":27201,\"pass_rate\":0.4713,\"provider_group\":\"provider_y\",\"quality\":0.7374,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.21,\"estimated_latency_ms\":73432,\"estimated_pass_rate\":0.5413,\"estimated_quality\":0.7881,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.607,\"max_latency_ms\":137638,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4771, "category": "review", "complexity": 0.7872, "consequence": 0.8249, "evidence_need": 0.9336, "family": "policy_compliance", "input_tokens": 3080, "request_id": "synthetic-policy_compliance-000393", "reversibility": 0.619, "risk": 0.6602, "summary": "Identify evidence required before approving usage metering for production."}, "workers": [{"credits_per_call": 3.2, "id": "worker_a", "p95_latency_ms": 46231, "pass_rate": 0.7227, "provider_group": "provider_x", "quality": 0.8351, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.021, "id": "worker_b", "p95_latency_ms": 88633, "pass_rate": 0.9722, "provider_group": "provider_z", "quality": 0.9794, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.01, "id": "worker_c", "p95_latency_ms": 27201, "pass_rate": 0.4713, "provider_group": "provider_y", "quality": 0.7374, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.288, "max_latency_ms": 214976, "minimum_quality": 0.8745}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000168", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.676, "estimated_latency_ms": 153197, "estimated_pass_rate": 0.6818, "estimated_quality": 0.8823, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.288, "max_latency_ms": 214976, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6818, "quality": 0.8823, "quality_target_met": true, "utility": 72.833264}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.288,\"max_latency_ms\":214976,\"minimum_quality\":0.8745},\"task\":{\"ambiguity\":0.6574,\"category\":\"review\",\"complexity\":0.7569,\"consequence\":0.5401,\"evidence_need\":0.641,\"family\":\"policy_compliance\",\"input_tokens\":2200,\"reversibility\":0.3175,\"risk\":0.7867,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.287,\"id\":\"worker_a\",\"p95_latency_ms\":55770,\"pass_rate\":0.742,\"provider_group\":\"provider_x\",\"quality\":0.7938,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.881,\"id\":\"worker_b\",\"p95_latency_ms\":30880,\"pass_rate\":0.4845,\"provider_group\":\"provider_y\",\"quality\":0.7789,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.508,\"id\":\"worker_c\",\"p95_latency_ms\":66547,\"pass_rate\":0.8995,\"provider_group\":\"provider_z\",\"quality\":0.9447,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.676,\"estimated_latency_ms\":153197,\"estimated_pass_rate\":0.6818,\"estimated_quality\":0.8823,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.288,\"max_latency_ms\":214976,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6574, "category": "review", "complexity": 0.7569, "consequence": 0.5401, "evidence_need": 0.641, "family": "policy_compliance", "input_tokens": 2200, "request_id": "synthetic-policy_compliance-000168", "reversibility": 0.3175, "risk": 0.7867, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.287, "id": "worker_a", "p95_latency_ms": 55770, "pass_rate": 0.742, "provider_group": "provider_x", "quality": 0.7938, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.881, "id": "worker_b", "p95_latency_ms": 30880, "pass_rate": 0.4845, "provider_group": "provider_y", "quality": 0.7789, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.508, "id": "worker_c", "p95_latency_ms": 66547, "pass_rate": 0.8995, "provider_group": "provider_z", "quality": 0.9447, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.916, "max_latency_ms": 242674, "minimum_quality": 0.8833}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000771", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.778, "estimated_latency_ms": 177712, "estimated_pass_rate": 0.7205, "estimated_quality": 0.8926, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.916, "max_latency_ms": 242674, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7205, "quality": 0.8926, "quality_target_met": true, "utility": 71.034604}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.916,\"max_latency_ms\":242674,\"minimum_quality\":0.8833},\"task\":{\"ambiguity\":0.3787,\"category\":\"review\",\"complexity\":0.743,\"consequence\":0.5899,\"evidence_need\":0.8715,\"family\":\"policy_compliance\",\"input_tokens\":1520,\"reversibility\":0.3475,\"risk\":0.4566,\"summary\":\"Audit a workflow for OAuth callbacks against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.542,\"id\":\"worker_a\",\"p95_latency_ms\":84483,\"pass_rate\":0.9835,\"provider_group\":\"provider_x\",\"quality\":0.9282,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.039,\"id\":\"worker_b\",\"p95_latency_ms\":20306,\"pass_rate\":0.5075,\"provider_group\":\"provider_y\",\"quality\":0.7642,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.197,\"id\":\"worker_c\",\"p95_latency_ms\":72923,\"pass_rate\":0.7476,\"provider_group\":\"provider_z\",\"quality\":0.8186,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.778,\"estimated_latency_ms\":177712,\"estimated_pass_rate\":0.7205,\"estimated_quality\":0.8926,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.916,\"max_latency_ms\":242674,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3787, "category": "review", "complexity": 0.743, "consequence": 0.5899, "evidence_need": 0.8715, "family": "policy_compliance", "input_tokens": 1520, "request_id": "synthetic-policy_compliance-000771", "reversibility": 0.3475, "risk": 0.4566, "summary": "Audit a workflow for OAuth callbacks against explicit access requirements."}, "workers": [{"credits_per_call": 11.542, "id": "worker_a", "p95_latency_ms": 84483, "pass_rate": 0.9835, "provider_group": "provider_x", "quality": 0.9282, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.039, "id": "worker_b", "p95_latency_ms": 20306, "pass_rate": 0.5075, "provider_group": "provider_y", "quality": 0.7642, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.197, "id": "worker_c", "p95_latency_ms": 72923, "pass_rate": 0.7476, "provider_group": "provider_z", "quality": 0.8186, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.503, "max_latency_ms": 113678, "minimum_quality": 0.89}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000270", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.35, "estimated_latency_ms": 96820, "estimated_pass_rate": 0.6667, "estimated_quality": 0.8397, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.503, "max_latency_ms": 113678, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6667, "quality": 0.8397, "quality_target_met": false, "utility": 63.800269}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.503,\"max_latency_ms\":113678,\"minimum_quality\":0.89},\"task\":{\"ambiguity\":0.6228,\"category\":\"review\",\"complexity\":0.808,\"consequence\":0.8078,\"evidence_need\":0.8821,\"family\":\"policy_compliance\",\"input_tokens\":7520,\"reversibility\":0.5681,\"risk\":0.6334,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":3.267,\"id\":\"worker_a\",\"p95_latency_ms\":70101,\"pass_rate\":0.8804,\"provider_group\":\"provider_y\",\"quality\":0.9194,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.083,\"id\":\"worker_b\",\"p95_latency_ms\":26719,\"pass_rate\":0.5667,\"provider_group\":\"provider_x\",\"quality\":0.7232,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.079,\"id\":\"worker_c\",\"p95_latency_ms\":85310,\"pass_rate\":0.9293,\"provider_group\":\"provider_z\",\"quality\":0.9474,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.35,\"estimated_latency_ms\":96820,\"estimated_pass_rate\":0.6667,\"estimated_quality\":0.8397,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.503,\"max_latency_ms\":113678,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6228, "category": "review", "complexity": 0.808, "consequence": 0.8078, "evidence_need": 0.8821, "family": "policy_compliance", "input_tokens": 7520, "request_id": "synthetic-policy_compliance-000270", "reversibility": 0.5681, "risk": 0.6334, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 3.267, "id": "worker_a", "p95_latency_ms": 70101, "pass_rate": 0.8804, "provider_group": "provider_y", "quality": 0.9194, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.083, "id": "worker_b", "p95_latency_ms": 26719, "pass_rate": 0.5667, "provider_group": "provider_x", "quality": 0.7232, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.079, "id": "worker_c", "p95_latency_ms": 85310, "pass_rate": 0.9293, "provider_group": "provider_z", "quality": 0.9474, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.038, "max_latency_ms": 178058, "minimum_quality": 0.806}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000693", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.053, "estimated_latency_ms": 168271, "estimated_pass_rate": 0.7066, "estimated_quality": 0.9266, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.038, "max_latency_ms": 178058, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7066, "quality": 0.9266, "quality_target_met": true, "utility": 74.409945}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.038,\"max_latency_ms\":178058,\"minimum_quality\":0.806},\"task\":{\"ambiguity\":0.786,\"category\":\"review\",\"complexity\":0.4033,\"consequence\":0.7,\"evidence_need\":0.7751,\"family\":\"policy_compliance\",\"input_tokens\":4160,\"reversibility\":0.5856,\"risk\":0.7525,\"summary\":\"Identify evidence required before approving provider failover for production.\"},\"workers\":[{\"credits_per_call\":4.117,\"id\":\"worker_a\",\"p95_latency_ms\":51098,\"pass_rate\":0.7767,\"provider_group\":\"provider_y\",\"quality\":0.8088,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.852,\"id\":\"worker_b\",\"p95_latency_ms\":27234,\"pass_rate\":0.4711,\"provider_group\":\"provider_x\",\"quality\":0.7932,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.084,\"id\":\"worker_c\",\"p95_latency_ms\":89939,\"pass_rate\":0.9247,\"provider_group\":\"provider_z\",\"quality\":0.9931,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.053,\"estimated_latency_ms\":168271,\"estimated_pass_rate\":0.7066,\"estimated_quality\":0.9266,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.038,\"max_latency_ms\":178058,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.786, "category": "review", "complexity": 0.4033, "consequence": 0.7, "evidence_need": 0.7751, "family": "policy_compliance", "input_tokens": 4160, "request_id": "synthetic-policy_compliance-000693", "reversibility": 0.5856, "risk": 0.7525, "summary": "Identify evidence required before approving provider failover for production."}, "workers": [{"credits_per_call": 4.117, "id": "worker_a", "p95_latency_ms": 51098, "pass_rate": 0.7767, "provider_group": "provider_y", "quality": 0.8088, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.852, "id": "worker_b", "p95_latency_ms": 27234, "pass_rate": 0.4711, "provider_group": "provider_x", "quality": 0.7932, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.084, "id": "worker_c", "p95_latency_ms": 89939, "pass_rate": 0.9247, "provider_group": "provider_z", "quality": 0.9931, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.392, "max_latency_ms": 48320, "minimum_quality": 0.7215}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000434", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.211, "estimated_latency_ms": 28309, "estimated_pass_rate": 0.3461, "estimated_quality": 0.5073, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.392, "max_latency_ms": 48320, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3461, "quality": 0.5073, "quality_target_met": false, "utility": 24.236211}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.392,\"max_latency_ms\":48320,\"minimum_quality\":0.7215},\"task\":{\"ambiguity\":0.7377,\"category\":\"analysis\",\"complexity\":0.5618,\"consequence\":0.4425,\"evidence_need\":0.6855,\"family\":\"performance_tuning\",\"input_tokens\":6600,\"reversibility\":0.594,\"risk\":0.5153,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":14.289,\"id\":\"worker_a\",\"p95_latency_ms\":82275,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.014,\"id\":\"worker_b\",\"p95_latency_ms\":46095,\"pass_rate\":0.9241,\"provider_group\":\"provider_x\",\"quality\":0.8871,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.211,\"id\":\"worker_c\",\"p95_latency_ms\":28309,\"pass_rate\":0.5073,\"provider_group\":\"provider_z\",\"quality\":0.6155,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.211,\"estimated_latency_ms\":28309,\"estimated_pass_rate\":0.3461,\"estimated_quality\":0.5073,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.392,\"max_latency_ms\":48320,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7377, "category": "analysis", "complexity": 0.5618, "consequence": 0.4425, "evidence_need": 0.6855, "family": "performance_tuning", "input_tokens": 6600, "request_id": "synthetic-performance_tuning-000434", "reversibility": 0.594, "risk": 0.5153, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 14.289, "id": "worker_a", "p95_latency_ms": 82275, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.014, "id": "worker_b", "p95_latency_ms": 46095, "pass_rate": 0.9241, "provider_group": "provider_x", "quality": 0.8871, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.211, "id": "worker_c", "p95_latency_ms": 28309, "pass_rate": 0.5073, "provider_group": "provider_z", "quality": 0.6155, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.463, "max_latency_ms": 167579, "minimum_quality": 0.8013}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001038", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.736, "estimated_latency_ms": 163970, "estimated_pass_rate": 0.7976, "estimated_quality": 0.8952, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.463, "max_latency_ms": 167579, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7976, "quality": 0.8952, "quality_target_met": true, "utility": 74.659882}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.463,\"max_latency_ms\":167579,\"minimum_quality\":0.8013},\"task\":{\"ambiguity\":0.6936,\"category\":\"review\",\"complexity\":0.4933,\"consequence\":0.7327,\"evidence_need\":0.6458,\"family\":\"policy_compliance\",\"input_tokens\":4440,\"reversibility\":0.6953,\"risk\":0.7861,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":0.881,\"id\":\"worker_a\",\"p95_latency_ms\":20167,\"pass_rate\":0.6121,\"provider_group\":\"provider_z\",\"quality\":0.7257,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.267,\"id\":\"worker_b\",\"p95_latency_ms\":54623,\"pass_rate\":0.7606,\"provider_group\":\"provider_y\",\"quality\":0.8886,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.588,\"id\":\"worker_c\",\"p95_latency_ms\":89180,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.948,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.736,\"estimated_latency_ms\":163970,\"estimated_pass_rate\":0.7976,\"estimated_quality\":0.8952,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.463,\"max_latency_ms\":167579,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6936, "category": "review", "complexity": 0.4933, "consequence": 0.7327, "evidence_need": 0.6458, "family": "policy_compliance", "input_tokens": 4440, "request_id": "synthetic-policy_compliance-001038", "reversibility": 0.6953, "risk": 0.7861, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 0.881, "id": "worker_a", "p95_latency_ms": 20167, "pass_rate": 0.6121, "provider_group": "provider_z", "quality": 0.7257, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.267, "id": "worker_b", "p95_latency_ms": 54623, "pass_rate": 0.7606, "provider_group": "provider_y", "quality": 0.8886, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.588, "id": "worker_c", "p95_latency_ms": 89180, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.948, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.89, "max_latency_ms": 33717, "minimum_quality": 0.5707}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000227", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.886, "estimated_latency_ms": 29630, "estimated_pass_rate": 0.324, "estimated_quality": 0.6153, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.89, "max_latency_ms": 33717, "policy": "lean", "reason_codes": ["latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.324, "quality": 0.6153, "quality_target_met": true, "utility": 39.500046}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.89,\"max_latency_ms\":33717,\"minimum_quality\":0.5707},\"task\":{\"ambiguity\":0.6646,\"category\":\"analysis\",\"complexity\":0.5379,\"consequence\":0.2563,\"evidence_need\":0.4285,\"family\":\"performance_tuning\",\"input_tokens\":3320,\"reversibility\":0.7791,\"risk\":0.5078,\"summary\":\"Diagnose latency variance in webhook delivery and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":4.134,\"id\":\"worker_a\",\"p95_latency_ms\":80674,\"pass_rate\":0.9295,\"provider_group\":\"provider_y\",\"quality\":0.98,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.886,\"id\":\"worker_b\",\"p95_latency_ms\":29630,\"pass_rate\":0.4764,\"provider_group\":\"provider_x\",\"quality\":0.7138,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":14.38,\"id\":\"worker_c\",\"p95_latency_ms\":59520,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.886,\"estimated_latency_ms\":29630,\"estimated_pass_rate\":0.324,\"estimated_quality\":0.6153,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.89,\"max_latency_ms\":33717,\"policy\":\"lean\",\"reason_codes\":[\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6646, "category": "analysis", "complexity": 0.5379, "consequence": 0.2563, "evidence_need": 0.4285, "family": "performance_tuning", "input_tokens": 3320, "request_id": "synthetic-performance_tuning-000227", "reversibility": 0.7791, "risk": 0.5078, "summary": "Diagnose latency variance in webhook delivery and rank the next experiments."}, "workers": [{"credits_per_call": 4.134, "id": "worker_a", "p95_latency_ms": 80674, "pass_rate": 0.9295, "provider_group": "provider_y", "quality": 0.98, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.886, "id": "worker_b", "p95_latency_ms": 29630, "pass_rate": 0.4764, "provider_group": "provider_x", "quality": 0.7138, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 14.38, "id": "worker_c", "p95_latency_ms": 59520, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.643, "max_latency_ms": 216021, "minimum_quality": 0.8895}, "family": "migration_planning", "id": "synthetic-migration_planning-000925", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.222, "estimated_latency_ms": 90611, "estimated_pass_rate": 0.6666, "estimated_quality": 0.841, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.643, "max_latency_ms": 216021, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6666, "quality": 0.841, "quality_target_met": false, "utility": 60.784363}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.643,\"max_latency_ms\":216021,\"minimum_quality\":0.8895},\"task\":{\"ambiguity\":0.5487,\"category\":\"coding\",\"complexity\":0.6882,\"consequence\":0.5386,\"evidence_need\":0.6764,\"family\":\"migration_planning\",\"input_tokens\":3360,\"reversibility\":0.6167,\"risk\":0.5766,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.247,\"id\":\"worker_a\",\"p95_latency_ms\":68212,\"pass_rate\":0.8873,\"provider_group\":\"provider_x\",\"quality\":0.9081,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.975,\"id\":\"worker_b\",\"p95_latency_ms\":22399,\"pass_rate\":0.5627,\"provider_group\":\"provider_y\",\"quality\":0.7682,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.293,\"id\":\"worker_c\",\"p95_latency_ms\":53245,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9136,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.222,\"estimated_latency_ms\":90611,\"estimated_pass_rate\":0.6666,\"estimated_quality\":0.841,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.643,\"max_latency_ms\":216021,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5487, "category": "coding", "complexity": 0.6882, "consequence": 0.5386, "evidence_need": 0.6764, "family": "migration_planning", "input_tokens": 3360, "request_id": "synthetic-migration_planning-000925", "reversibility": 0.6167, "risk": 0.5766, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 3.247, "id": "worker_a", "p95_latency_ms": 68212, "pass_rate": 0.8873, "provider_group": "provider_x", "quality": 0.9081, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.975, "id": "worker_b", "p95_latency_ms": 22399, "pass_rate": 0.5627, "provider_group": "provider_y", "quality": 0.7682, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.293, "id": "worker_c", "p95_latency_ms": 53245, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9136, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.43, "max_latency_ms": 94043, "minimum_quality": 0.8388}, "family": "migration_planning", "id": "synthetic-migration_planning-000490", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.883, "estimated_latency_ms": 81611, "estimated_pass_rate": 0.672, "estimated_quality": 0.82, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.43, "max_latency_ms": 94043, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.672, "quality": 0.82, "quality_target_met": false, "utility": 58.843115}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.43,\"max_latency_ms\":94043,\"minimum_quality\":0.8388},\"task\":{\"ambiguity\":0.5754,\"category\":\"coding\",\"complexity\":0.7467,\"consequence\":0.5123,\"evidence_need\":0.6186,\"family\":\"migration_planning\",\"input_tokens\":3960,\"reversibility\":0.5779,\"risk\":0.7108,\"summary\":\"Design compatibility testing for a webhook delivery schema migration.\"},\"workers\":[{\"credits_per_call\":14.042,\"id\":\"worker_a\",\"p95_latency_ms\":87152,\"pass_rate\":0.9255,\"provider_group\":\"provider_x\",\"quality\":0.9799,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.009,\"id\":\"worker_b\",\"p95_latency_ms\":62978,\"pass_rate\":0.7555,\"provider_group\":\"provider_y\",\"quality\":0.9086,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.874,\"id\":\"worker_c\",\"p95_latency_ms\":18633,\"pass_rate\":0.5505,\"provider_group\":\"provider_z\",\"quality\":0.7463,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.883,\"estimated_latency_ms\":81611,\"estimated_pass_rate\":0.672,\"estimated_quality\":0.82,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.43,\"max_latency_ms\":94043,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5754, "category": "coding", "complexity": 0.7467, "consequence": 0.5123, "evidence_need": 0.6186, "family": "migration_planning", "input_tokens": 3960, "request_id": "synthetic-migration_planning-000490", "reversibility": 0.5779, "risk": 0.7108, "summary": "Design compatibility testing for a webhook delivery schema migration."}, "workers": [{"credits_per_call": 14.042, "id": "worker_a", "p95_latency_ms": 87152, "pass_rate": 0.9255, "provider_group": "provider_x", "quality": 0.9799, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.009, "id": "worker_b", "p95_latency_ms": 62978, "pass_rate": 0.7555, "provider_group": "provider_y", "quality": 0.9086, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.874, "id": "worker_c", "p95_latency_ms": 18633, "pass_rate": 0.5505, "provider_group": "provider_z", "quality": 0.7463, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.959, "max_latency_ms": 31412, "minimum_quality": 0.4924}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000669", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.056, "estimated_latency_ms": 18868, "estimated_pass_rate": 0.3032, "estimated_quality": 0.6618, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.959, "max_latency_ms": 31412, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3032, "quality": 0.6618, "quality_target_met": true, "utility": 41.528866}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.959,\"max_latency_ms\":31412,\"minimum_quality\":0.4924},\"task\":{\"ambiguity\":0.7562,\"category\":\"review\",\"complexity\":0.5543,\"consequence\":0.5816,\"evidence_need\":0.8836,\"family\":\"policy_compliance\",\"input_tokens\":480,\"reversibility\":0.5975,\"risk\":0.6623,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.056,\"id\":\"worker_a\",\"p95_latency_ms\":18868,\"pass_rate\":0.4769,\"provider_group\":\"provider_x\",\"quality\":0.7805,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.185,\"id\":\"worker_b\",\"p95_latency_ms\":51357,\"pass_rate\":0.7521,\"provider_group\":\"provider_y\",\"quality\":0.829,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.579,\"id\":\"worker_c\",\"p95_latency_ms\":58745,\"pass_rate\":0.9452,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.056,\"estimated_latency_ms\":18868,\"estimated_pass_rate\":0.3032,\"estimated_quality\":0.6618,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.959,\"max_latency_ms\":31412,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7562, "category": "review", "complexity": 0.5543, "consequence": 0.5816, "evidence_need": 0.8836, "family": "policy_compliance", "input_tokens": 480, "request_id": "synthetic-policy_compliance-000669", "reversibility": 0.5975, "risk": 0.6623, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 1.056, "id": "worker_a", "p95_latency_ms": 18868, "pass_rate": 0.4769, "provider_group": "provider_x", "quality": 0.7805, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.185, "id": "worker_b", "p95_latency_ms": 51357, "pass_rate": 0.7521, "provider_group": "provider_y", "quality": 0.829, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.579, "id": "worker_c", "p95_latency_ms": 58745, "pass_rate": 0.9452, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.088, "max_latency_ms": 200628, "minimum_quality": 0.8985}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001170", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.134, "estimated_latency_ms": 116141, "estimated_pass_rate": 0.8743, "estimated_quality": 0.9133, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.088, "max_latency_ms": 200628, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8743, "quality": 0.9133, "quality_target_met": true, "utility": 70.751314}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.088,\"max_latency_ms\":200628,\"minimum_quality\":0.8985},\"task\":{\"ambiguity\":0.7794,\"category\":\"review\",\"complexity\":0.5403,\"consequence\":0.5904,\"evidence_need\":0.7188,\"family\":\"policy_compliance\",\"input_tokens\":1520,\"reversibility\":0.6389,\"risk\":0.7563,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.401,\"id\":\"worker_a\",\"p95_latency_ms\":54638,\"pass_rate\":0.7335,\"provider_group\":\"provider_x\",\"quality\":0.8047,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.974,\"id\":\"worker_b\",\"p95_latency_ms\":32199,\"pass_rate\":0.5781,\"provider_group\":\"provider_z\",\"quality\":0.7063,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.733,\"id\":\"worker_c\",\"p95_latency_ms\":61503,\"pass_rate\":0.9884,\"provider_group\":\"provider_y\",\"quality\":0.9819,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.134,\"estimated_latency_ms\":116141,\"estimated_pass_rate\":0.8743,\"estimated_quality\":0.9133,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.088,\"max_latency_ms\":200628,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7794, "category": "review", "complexity": 0.5403, "consequence": 0.5904, "evidence_need": 0.7188, "family": "policy_compliance", "input_tokens": 1520, "request_id": "synthetic-policy_compliance-001170", "reversibility": 0.6389, "risk": 0.7563, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 3.401, "id": "worker_a", "p95_latency_ms": 54638, "pass_rate": 0.7335, "provider_group": "provider_x", "quality": 0.8047, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.974, "id": "worker_b", "p95_latency_ms": 32199, "pass_rate": 0.5781, "provider_group": "provider_z", "quality": 0.7063, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.733, "id": "worker_c", "p95_latency_ms": 61503, "pass_rate": 0.9884, "provider_group": "provider_y", "quality": 0.9819, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.14, "max_latency_ms": 38556, "minimum_quality": 0.7008}, "family": "migration_planning", "id": "synthetic-migration_planning-000250", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.816, "estimated_latency_ms": 31859, "estimated_pass_rate": 0.4281, "estimated_quality": 0.5545, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.14, "max_latency_ms": 38556, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4281, "quality": 0.5545, "quality_target_met": false, "utility": 29.351803}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.14,\"max_latency_ms\":38556,\"minimum_quality\":0.7008},\"task\":{\"ambiguity\":0.7291,\"category\":\"coding\",\"complexity\":0.6222,\"consequence\":0.4917,\"evidence_need\":0.7053,\"family\":\"migration_planning\",\"input_tokens\":4880,\"reversibility\":0.3784,\"risk\":0.625,\"summary\":\"Design compatibility testing for a credit reservation schema migration.\"},\"workers\":[{\"credits_per_call\":13.616,\"id\":\"worker_a\",\"p95_latency_ms\":58033,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.949,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.816,\"id\":\"worker_b\",\"p95_latency_ms\":31859,\"pass_rate\":0.6044,\"provider_group\":\"provider_x\",\"quality\":0.681,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.098,\"id\":\"worker_c\",\"p95_latency_ms\":70639,\"pass_rate\":0.8204,\"provider_group\":\"provider_y\",\"quality\":0.9113,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.816,\"estimated_latency_ms\":31859,\"estimated_pass_rate\":0.4281,\"estimated_quality\":0.5545,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.14,\"max_latency_ms\":38556,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7291, "category": "coding", "complexity": 0.6222, "consequence": 0.4917, "evidence_need": 0.7053, "family": "migration_planning", "input_tokens": 4880, "request_id": "synthetic-migration_planning-000250", "reversibility": 0.3784, "risk": 0.625, "summary": "Design compatibility testing for a credit reservation schema migration."}, "workers": [{"credits_per_call": 13.616, "id": "worker_a", "p95_latency_ms": 58033, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.949, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.816, "id": "worker_b", "p95_latency_ms": 31859, "pass_rate": 0.6044, "provider_group": "provider_x", "quality": 0.681, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.098, "id": "worker_c", "p95_latency_ms": 70639, "pass_rate": 0.8204, "provider_group": "provider_y", "quality": 0.9113, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.67, "max_latency_ms": 30574, "minimum_quality": 0.5452}, "family": "migration_planning", "id": "synthetic-migration_planning-000085", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.816, "estimated_latency_ms": 24414, "estimated_pass_rate": 0.3172, "estimated_quality": 0.5887, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.67, "max_latency_ms": 30574, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3172, "quality": 0.5887, "quality_target_met": true, "utility": 30.562327}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.67,\"max_latency_ms\":30574,\"minimum_quality\":0.5452},\"task\":{\"ambiguity\":0.3349,\"category\":\"coding\",\"complexity\":0.7093,\"consequence\":0.4533,\"evidence_need\":0.5298,\"family\":\"migration_planning\",\"input_tokens\":4680,\"reversibility\":0.2747,\"risk\":0.5671,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":0.816,\"id\":\"worker_a\",\"p95_latency_ms\":24414,\"pass_rate\":0.4689,\"provider_group\":\"provider_y\",\"quality\":0.7003,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.249,\"id\":\"worker_b\",\"p95_latency_ms\":86454,\"pass_rate\":0.8925,\"provider_group\":\"provider_z\",\"quality\":0.9327,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.548,\"id\":\"worker_c\",\"p95_latency_ms\":68226,\"pass_rate\":0.8515,\"provider_group\":\"provider_x\",\"quality\":0.8202,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.816,\"estimated_latency_ms\":24414,\"estimated_pass_rate\":0.3172,\"estimated_quality\":0.5887,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.67,\"max_latency_ms\":30574,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3349, "category": "coding", "complexity": 0.7093, "consequence": 0.4533, "evidence_need": 0.5298, "family": "migration_planning", "input_tokens": 4680, "request_id": "synthetic-migration_planning-000085", "reversibility": 0.2747, "risk": 0.5671, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 0.816, "id": "worker_a", "p95_latency_ms": 24414, "pass_rate": 0.4689, "provider_group": "provider_y", "quality": 0.7003, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.249, "id": "worker_b", "p95_latency_ms": 86454, "pass_rate": 0.8925, "provider_group": "provider_z", "quality": 0.9327, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.548, "id": "worker_c", "p95_latency_ms": 68226, "pass_rate": 0.8515, "provider_group": "provider_x", "quality": 0.8202, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.386, "max_latency_ms": 213263, "minimum_quality": 0.9362}, "family": "migration_planning", "id": "synthetic-migration_planning-000952", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.717, "estimated_latency_ms": 182205, "estimated_pass_rate": 0.7592, "estimated_quality": 0.9201, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.386, "max_latency_ms": 213263, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7592, "quality": 0.9201, "quality_target_met": false, "utility": 77.976629}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.386,\"max_latency_ms\":213263,\"minimum_quality\":0.9362},\"task\":{\"ambiguity\":0.4573,\"category\":\"coding\",\"complexity\":0.884,\"consequence\":0.7592,\"evidence_need\":0.757,\"family\":\"migration_planning\",\"input_tokens\":4800,\"reversibility\":0.4406,\"risk\":0.6743,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":4.26,\"id\":\"worker_a\",\"p95_latency_ms\":72545,\"pass_rate\":0.8426,\"provider_group\":\"provider_y\",\"quality\":0.8405,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.038,\"id\":\"worker_b\",\"p95_latency_ms\":27667,\"pass_rate\":0.5595,\"provider_group\":\"provider_x\",\"quality\":0.8037,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.419,\"id\":\"worker_c\",\"p95_latency_ms\":81993,\"pass_rate\":0.9662,\"provider_group\":\"provider_z\",\"quality\":0.9773,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.717,\"estimated_latency_ms\":182205,\"estimated_pass_rate\":0.7592,\"estimated_quality\":0.9201,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.386,\"max_latency_ms\":213263,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4573, "category": "coding", "complexity": 0.884, "consequence": 0.7592, "evidence_need": 0.757, "family": "migration_planning", "input_tokens": 4800, "request_id": "synthetic-migration_planning-000952", "reversibility": 0.4406, "risk": 0.6743, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 4.26, "id": "worker_a", "p95_latency_ms": 72545, "pass_rate": 0.8426, "provider_group": "provider_y", "quality": 0.8405, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.038, "id": "worker_b", "p95_latency_ms": 27667, "pass_rate": 0.5595, "provider_group": "provider_x", "quality": 0.8037, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.419, "id": "worker_c", "p95_latency_ms": 81993, "pass_rate": 0.9662, "provider_group": "provider_z", "quality": 0.9773, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.031, "max_latency_ms": 187967, "minimum_quality": 0.8205}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000555", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.688, "estimated_latency_ms": 166984, "estimated_pass_rate": 0.8039, "estimated_quality": 0.9072, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.031, "max_latency_ms": 187967, "policy": "strict", "reason_codes": ["high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8039, "quality": 0.9072, "quality_target_met": true, "utility": 74.232747}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.031,\"max_latency_ms\":187967,\"minimum_quality\":0.8205},\"task\":{\"ambiguity\":0.3915,\"category\":\"review\",\"complexity\":0.6247,\"consequence\":0.6486,\"evidence_need\":0.5875,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.3116,\"risk\":0.7554,\"summary\":\"Identify evidence required before approving credit reservation for production.\"},\"workers\":[{\"credits_per_call\":10.389,\"id\":\"worker_a\",\"p95_latency_ms\":90946,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9669,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.481,\"id\":\"worker_b\",\"p95_latency_ms\":56955,\"pass_rate\":0.8703,\"provider_group\":\"provider_z\",\"quality\":0.919,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.818,\"id\":\"worker_c\",\"p95_latency_ms\":19083,\"pass_rate\":0.5437,\"provider_group\":\"provider_x\",\"quality\":0.686,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.688,\"estimated_latency_ms\":166984,\"estimated_pass_rate\":0.8039,\"estimated_quality\":0.9072,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.031,\"max_latency_ms\":187967,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3915, "category": "review", "complexity": 0.6247, "consequence": 0.6486, "evidence_need": 0.5875, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-000555", "reversibility": 0.3116, "risk": 0.7554, "summary": "Identify evidence required before approving credit reservation for production."}, "workers": [{"credits_per_call": 10.389, "id": "worker_a", "p95_latency_ms": 90946, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9669, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.481, "id": "worker_b", "p95_latency_ms": 56955, "pass_rate": 0.8703, "provider_group": "provider_z", "quality": 0.919, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.818, "id": "worker_c", "p95_latency_ms": 19083, "pass_rate": 0.5437, "provider_group": "provider_x", "quality": 0.686, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.502, "max_latency_ms": 202100, "minimum_quality": 0.8548}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000252", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.432, "estimated_latency_ms": 164428, "estimated_pass_rate": 0.7442, "estimated_quality": 0.9032, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 25.502, "max_latency_ms": 202100, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7442, "quality": 0.9032, "quality_target_met": true, "utility": 79.541129}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.502,\"max_latency_ms\":202100,\"minimum_quality\":0.8548},\"task\":{\"ambiguity\":0.5524,\"category\":\"review\",\"complexity\":0.7244,\"consequence\":0.8565,\"evidence_need\":0.8588,\"family\":\"policy_compliance\",\"input_tokens\":1200,\"reversibility\":0.4203,\"risk\":0.7997,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.157,\"id\":\"worker_a\",\"p95_latency_ms\":26786,\"pass_rate\":0.5034,\"provider_group\":\"provider_y\",\"quality\":0.7035,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.235,\"id\":\"worker_b\",\"p95_latency_ms\":91737,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9755,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.04,\"id\":\"worker_c\",\"p95_latency_ms\":45905,\"pass_rate\":0.7542,\"provider_group\":\"provider_x\",\"quality\":0.8467,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.432,\"estimated_latency_ms\":164428,\"estimated_pass_rate\":0.7442,\"estimated_quality\":0.9032,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.502,\"max_latency_ms\":202100,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5524, "category": "review", "complexity": 0.7244, "consequence": 0.8565, "evidence_need": 0.8588, "family": "policy_compliance", "input_tokens": 1200, "request_id": "synthetic-policy_compliance-000252", "reversibility": 0.4203, "risk": 0.7997, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.157, "id": "worker_a", "p95_latency_ms": 26786, "pass_rate": 0.5034, "provider_group": "provider_y", "quality": 0.7035, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.235, "id": "worker_b", "p95_latency_ms": 91737, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9755, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.04, "id": "worker_c", "p95_latency_ms": 45905, "pass_rate": 0.7542, "provider_group": "provider_x", "quality": 0.8467, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.479, "max_latency_ms": 37511, "minimum_quality": 0.5004}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000549", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.078, "estimated_latency_ms": 24833, "estimated_pass_rate": 0.3513, "estimated_quality": 0.6129, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.479, "max_latency_ms": 37511, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3513, "quality": 0.6129, "quality_target_met": true, "utility": 23.163893}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.479,\"max_latency_ms\":37511,\"minimum_quality\":0.5004},\"task\":{\"ambiguity\":0.4915,\"category\":\"review\",\"complexity\":0.591,\"consequence\":0.6905,\"evidence_need\":0.8004,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.6154,\"risk\":0.843,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.072,\"id\":\"worker_a\",\"p95_latency_ms\":56393,\"pass_rate\":0.7875,\"provider_group\":\"provider_y\",\"quality\":0.796,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.481,\"id\":\"worker_b\",\"p95_latency_ms\":83124,\"pass_rate\":0.9683,\"provider_group\":\"provider_z\",\"quality\":0.9487,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.078,\"id\":\"worker_c\",\"p95_latency_ms\":24833,\"pass_rate\":0.5238,\"provider_group\":\"provider_x\",\"quality\":0.7335,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.078,\"estimated_latency_ms\":24833,\"estimated_pass_rate\":0.3513,\"estimated_quality\":0.6129,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.479,\"max_latency_ms\":37511,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4915, "category": "review", "complexity": 0.591, "consequence": 0.6905, "evidence_need": 0.8004, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-000549", "reversibility": 0.6154, "risk": 0.843, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 3.072, "id": "worker_a", "p95_latency_ms": 56393, "pass_rate": 0.7875, "provider_group": "provider_y", "quality": 0.796, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.481, "id": "worker_b", "p95_latency_ms": 83124, "pass_rate": 0.9683, "provider_group": "provider_z", "quality": 0.9487, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.078, "id": "worker_c", "p95_latency_ms": 24833, "pass_rate": 0.5238, "provider_group": "provider_x", "quality": 0.7335, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.58, "max_latency_ms": 151298, "minimum_quality": 0.7778}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000654", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.012, "estimated_latency_ms": 77672, "estimated_pass_rate": 0.5753, "estimated_quality": 0.8218, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.58, "max_latency_ms": 151298, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.5753, "quality": 0.8218, "quality_target_met": true, "utility": 62.53165}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.58,\"max_latency_ms\":151298,\"minimum_quality\":0.7778},\"task\":{\"ambiguity\":0.5285,\"category\":\"review\",\"complexity\":0.7761,\"consequence\":0.6838,\"evidence_need\":0.8394,\"family\":\"policy_compliance\",\"input_tokens\":4720,\"reversibility\":0.4745,\"risk\":0.6021,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.195,\"id\":\"worker_a\",\"p95_latency_ms\":22999,\"pass_rate\":0.4652,\"provider_group\":\"provider_y\",\"quality\":0.7559,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.754,\"id\":\"worker_b\",\"p95_latency_ms\":75895,\"pass_rate\":0.8991,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.817,\"id\":\"worker_c\",\"p95_latency_ms\":54673,\"pass_rate\":0.8171,\"provider_group\":\"provider_x\",\"quality\":0.8749,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.012,\"estimated_latency_ms\":77672,\"estimated_pass_rate\":0.5753,\"estimated_quality\":0.8218,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.58,\"max_latency_ms\":151298,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5285, "category": "review", "complexity": 0.7761, "consequence": 0.6838, "evidence_need": 0.8394, "family": "policy_compliance", "input_tokens": 4720, "request_id": "synthetic-policy_compliance-000654", "reversibility": 0.4745, "risk": 0.6021, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 1.195, "id": "worker_a", "p95_latency_ms": 22999, "pass_rate": 0.4652, "provider_group": "provider_y", "quality": 0.7559, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.754, "id": "worker_b", "p95_latency_ms": 75895, "pass_rate": 0.8991, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.817, "id": "worker_c", "p95_latency_ms": 54673, "pass_rate": 0.8171, "provider_group": "provider_x", "quality": 0.8749, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.922, "max_latency_ms": 38309, "minimum_quality": 0.5191}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000455", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.858, "estimated_latency_ms": 32299, "estimated_pass_rate": 0.3007, "estimated_quality": 0.5923, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.922, "max_latency_ms": 38309, "policy": "lean", "reason_codes": ["latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3007, "quality": 0.5923, "quality_target_met": true, "utility": 38.313308}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.922,\"max_latency_ms\":38309,\"minimum_quality\":0.5191},\"task\":{\"ambiguity\":0.6284,\"category\":\"analysis\",\"complexity\":0.4507,\"consequence\":0.5034,\"evidence_need\":0.3449,\"family\":\"performance_tuning\",\"input_tokens\":3720,\"reversibility\":0.8111,\"risk\":0.3164,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":14.922,\"id\":\"worker_a\",\"p95_latency_ms\":52321,\"pass_rate\":0.9809,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.858,\"id\":\"worker_b\",\"p95_latency_ms\":32299,\"pass_rate\":0.4249,\"provider_group\":\"provider_x\",\"quality\":0.6588,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.065,\"id\":\"worker_c\",\"p95_latency_ms\":56389,\"pass_rate\":0.8789,\"provider_group\":\"provider_y\",\"quality\":0.9517,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.858,\"estimated_latency_ms\":32299,\"estimated_pass_rate\":0.3007,\"estimated_quality\":0.5923,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.922,\"max_latency_ms\":38309,\"policy\":\"lean\",\"reason_codes\":[\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6284, "category": "analysis", "complexity": 0.4507, "consequence": 0.5034, "evidence_need": 0.3449, "family": "performance_tuning", "input_tokens": 3720, "request_id": "synthetic-performance_tuning-000455", "reversibility": 0.8111, "risk": 0.3164, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 14.922, "id": "worker_a", "p95_latency_ms": 52321, "pass_rate": 0.9809, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.858, "id": "worker_b", "p95_latency_ms": 32299, "pass_rate": 0.4249, "provider_group": "provider_x", "quality": 0.6588, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.065, "id": "worker_c", "p95_latency_ms": 56389, "pass_rate": 0.8789, "provider_group": "provider_y", "quality": 0.9517, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 1.397, "max_latency_ms": 32078, "minimum_quality": 0.6764}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000824", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.036, "estimated_latency_ms": 21844, "estimated_pass_rate": 0.2509, "estimated_quality": 0.6011, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.397, "max_latency_ms": 32078, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2509, "quality": 0.6011, "quality_target_met": false, "utility": 31.620833}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":1.397,\"max_latency_ms\":32078,\"minimum_quality\":0.6764},\"task\":{\"ambiguity\":0.3433,\"category\":\"analysis\",\"complexity\":0.8894,\"consequence\":0.3352,\"evidence_need\":0.4065,\"family\":\"performance_tuning\",\"input_tokens\":6120,\"reversibility\":0.7272,\"risk\":0.2833,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":12.73,\"id\":\"worker_a\",\"p95_latency_ms\":55545,\"pass_rate\":0.9778,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.036,\"id\":\"worker_b\",\"p95_latency_ms\":21844,\"pass_rate\":0.4007,\"provider_group\":\"provider_y\",\"quality\":0.7254,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.975,\"id\":\"worker_c\",\"p95_latency_ms\":58860,\"pass_rate\":0.8711,\"provider_group\":\"provider_x\",\"quality\":0.9397,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.036,\"estimated_latency_ms\":21844,\"estimated_pass_rate\":0.2509,\"estimated_quality\":0.6011,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.397,\"max_latency_ms\":32078,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3433, "category": "analysis", "complexity": 0.8894, "consequence": 0.3352, "evidence_need": 0.4065, "family": "performance_tuning", "input_tokens": 6120, "request_id": "synthetic-performance_tuning-000824", "reversibility": 0.7272, "risk": 0.2833, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 12.73, "id": "worker_a", "p95_latency_ms": 55545, "pass_rate": 0.9778, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.036, "id": "worker_b", "p95_latency_ms": 21844, "pass_rate": 0.4007, "provider_group": "provider_y", "quality": 0.7254, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.975, "id": "worker_c", "p95_latency_ms": 58860, "pass_rate": 0.8711, "provider_group": "provider_x", "quality": 0.9397, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.526, "max_latency_ms": 33308, "minimum_quality": 0.5903}, "family": "migration_planning", "id": "synthetic-migration_planning-000364", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.119, "estimated_latency_ms": 22659, "estimated_pass_rate": 0.4024, "estimated_quality": 0.5718, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.526, "max_latency_ms": 33308, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4024, "quality": 0.5718, "quality_target_met": false, "utility": 38.647011}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.526,\"max_latency_ms\":33308,\"minimum_quality\":0.5903},\"task\":{\"ambiguity\":0.6748,\"category\":\"coding\",\"complexity\":0.5545,\"consequence\":0.5444,\"evidence_need\":0.8076,\"family\":\"migration_planning\",\"input_tokens\":7920,\"reversibility\":0.6,\"risk\":0.527,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":2.741,\"id\":\"worker_a\",\"p95_latency_ms\":74715,\"pass_rate\":0.8211,\"provider_group\":\"provider_z\",\"quality\":0.8689,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.05,\"id\":\"worker_b\",\"p95_latency_ms\":53855,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.119,\"id\":\"worker_c\",\"p95_latency_ms\":22659,\"pass_rate\":0.559,\"provider_group\":\"provider_y\",\"quality\":0.6753,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.119,\"estimated_latency_ms\":22659,\"estimated_pass_rate\":0.4024,\"estimated_quality\":0.5718,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.526,\"max_latency_ms\":33308,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6748, "category": "coding", "complexity": 0.5545, "consequence": 0.5444, "evidence_need": 0.8076, "family": "migration_planning", "input_tokens": 7920, "request_id": "synthetic-migration_planning-000364", "reversibility": 0.6, "risk": 0.527, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 2.741, "id": "worker_a", "p95_latency_ms": 74715, "pass_rate": 0.8211, "provider_group": "provider_z", "quality": 0.8689, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.05, "id": "worker_b", "p95_latency_ms": 53855, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.119, "id": "worker_c", "p95_latency_ms": 22659, "pass_rate": 0.559, "provider_group": "provider_y", "quality": 0.6753, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.215, "max_latency_ms": 229640, "minimum_quality": 0.8003}, "family": "migration_planning", "id": "synthetic-migration_planning-000361", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.491, "estimated_latency_ms": 138635, "estimated_pass_rate": 0.9026, "estimated_quality": 0.9235, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.215, "max_latency_ms": 229640, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9026, "quality": 0.9235, "quality_target_met": true, "utility": 71.354427}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.215,\"max_latency_ms\":229640,\"minimum_quality\":0.8003},\"task\":{\"ambiguity\":0.5044,\"category\":\"coding\",\"complexity\":0.6618,\"consequence\":0.6794,\"evidence_need\":0.6158,\"family\":\"migration_planning\",\"input_tokens\":5560,\"reversibility\":0.2825,\"risk\":0.4742,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":1.021,\"id\":\"worker_a\",\"p95_latency_ms\":28406,\"pass_rate\":0.5022,\"provider_group\":\"provider_x\",\"quality\":0.677,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.358,\"id\":\"worker_b\",\"p95_latency_ms\":70525,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9659,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.133,\"id\":\"worker_c\",\"p95_latency_ms\":68110,\"pass_rate\":0.8387,\"provider_group\":\"provider_y\",\"quality\":0.8792,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.491,\"estimated_latency_ms\":138635,\"estimated_pass_rate\":0.9026,\"estimated_quality\":0.9235,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.215,\"max_latency_ms\":229640,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5044, "category": "coding", "complexity": 0.6618, "consequence": 0.6794, "evidence_need": 0.6158, "family": "migration_planning", "input_tokens": 5560, "request_id": "synthetic-migration_planning-000361", "reversibility": 0.2825, "risk": 0.4742, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 1.021, "id": "worker_a", "p95_latency_ms": 28406, "pass_rate": 0.5022, "provider_group": "provider_x", "quality": 0.677, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.358, "id": "worker_b", "p95_latency_ms": 70525, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9659, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.133, "id": "worker_c", "p95_latency_ms": 68110, "pass_rate": 0.8387, "provider_group": "provider_y", "quality": 0.8792, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.214, "max_latency_ms": 31743, "minimum_quality": 0.5588}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000702", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.199, "estimated_latency_ms": 23275, "estimated_pass_rate": 0.3461, "estimated_quality": 0.6516, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.214, "max_latency_ms": 31743, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3461, "quality": 0.6516, "quality_target_met": true, "utility": 28.815687}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.214,\"max_latency_ms\":31743,\"minimum_quality\":0.5588},\"task\":{\"ambiguity\":0.4795,\"category\":\"review\",\"complexity\":0.4595,\"consequence\":0.7205,\"evidence_need\":0.7673,\"family\":\"policy_compliance\",\"input_tokens\":3200,\"reversibility\":0.3062,\"risk\":0.6895,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.199,\"id\":\"worker_a\",\"p95_latency_ms\":23275,\"pass_rate\":0.4903,\"provider_group\":\"provider_y\",\"quality\":0.7365,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.656,\"id\":\"worker_b\",\"p95_latency_ms\":82445,\"pass_rate\":0.8891,\"provider_group\":\"provider_x\",\"quality\":0.9899,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.064,\"id\":\"worker_c\",\"p95_latency_ms\":46475,\"pass_rate\":0.7302,\"provider_group\":\"provider_z\",\"quality\":0.8093,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.199,\"estimated_latency_ms\":23275,\"estimated_pass_rate\":0.3461,\"estimated_quality\":0.6516,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.214,\"max_latency_ms\":31743,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4795, "category": "review", "complexity": 0.4595, "consequence": 0.7205, "evidence_need": 0.7673, "family": "policy_compliance", "input_tokens": 3200, "request_id": "synthetic-policy_compliance-000702", "reversibility": 0.3062, "risk": 0.6895, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 1.199, "id": "worker_a", "p95_latency_ms": 23275, "pass_rate": 0.4903, "provider_group": "provider_y", "quality": 0.7365, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.656, "id": "worker_b", "p95_latency_ms": 82445, "pass_rate": 0.8891, "provider_group": "provider_x", "quality": 0.9899, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.064, "id": "worker_c", "p95_latency_ms": 46475, "pass_rate": 0.7302, "provider_group": "provider_z", "quality": 0.8093, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.216, "max_latency_ms": 230825, "minimum_quality": 0.8728}, "family": "migration_planning", "id": "synthetic-migration_planning-000025", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.569, "estimated_latency_ms": 167970, "estimated_pass_rate": 0.7836, "estimated_quality": 0.9178, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.216, "max_latency_ms": 230825, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7836, "quality": 0.9178, "quality_target_met": true, "utility": 76.477996}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.216,\"max_latency_ms\":230825,\"minimum_quality\":0.8728},\"task\":{\"ambiguity\":0.585,\"category\":\"coding\",\"complexity\":0.8815,\"consequence\":0.5197,\"evidence_need\":0.4503,\"family\":\"migration_planning\",\"input_tokens\":7800,\"reversibility\":0.6396,\"risk\":0.6892,\"summary\":\"Plan a zero-downtime migration of access control with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.797,\"id\":\"worker_a\",\"p95_latency_ms\":28320,\"pass_rate\":0.6151,\"provider_group\":\"provider_x\",\"quality\":0.7887,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.57,\"id\":\"worker_b\",\"p95_latency_ms\":63166,\"pass_rate\":0.9668,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.202,\"id\":\"worker_c\",\"p95_latency_ms\":76484,\"pass_rate\":0.8155,\"provider_group\":\"provider_y\",\"quality\":0.9057,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.569,\"estimated_latency_ms\":167970,\"estimated_pass_rate\":0.7836,\"estimated_quality\":0.9178,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.216,\"max_latency_ms\":230825,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.585, "category": "coding", "complexity": 0.8815, "consequence": 0.5197, "evidence_need": 0.4503, "family": "migration_planning", "input_tokens": 7800, "request_id": "synthetic-migration_planning-000025", "reversibility": 0.6396, "risk": 0.6892, "summary": "Plan a zero-downtime migration of access control with rollback checkpoints."}, "workers": [{"credits_per_call": 0.797, "id": "worker_a", "p95_latency_ms": 28320, "pass_rate": 0.6151, "provider_group": "provider_x", "quality": 0.7887, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.57, "id": "worker_b", "p95_latency_ms": 63166, "pass_rate": 0.9668, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.202, "id": "worker_c", "p95_latency_ms": 76484, "pass_rate": 0.8155, "provider_group": "provider_y", "quality": 0.9057, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.236, "max_latency_ms": 235264, "minimum_quality": 0.869}, "family": "migration_planning", "id": "synthetic-migration_planning-000418", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.629, "estimated_latency_ms": 196306, "estimated_pass_rate": 0.7712, "estimated_quality": 0.8975, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.236, "max_latency_ms": 235264, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7712, "quality": 0.8975, "quality_target_met": true, "utility": 74.984431}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.236,\"max_latency_ms\":235264,\"minimum_quality\":0.869},\"task\":{\"ambiguity\":0.5137,\"category\":\"coding\",\"complexity\":0.7297,\"consequence\":0.7418,\"evidence_need\":0.7157,\"family\":\"migration_planning\",\"input_tokens\":280,\"reversibility\":0.3421,\"risk\":0.7125,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.046,\"id\":\"worker_a\",\"p95_latency_ms\":78501,\"pass_rate\":0.8421,\"provider_group\":\"provider_z\",\"quality\":0.9083,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.561,\"id\":\"worker_b\",\"p95_latency_ms\":88439,\"pass_rate\":0.9465,\"provider_group\":\"provider_x\",\"quality\":0.9232,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.022,\"id\":\"worker_c\",\"p95_latency_ms\":29366,\"pass_rate\":0.5565,\"provider_group\":\"provider_y\",\"quality\":0.7774,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.629,\"estimated_latency_ms\":196306,\"estimated_pass_rate\":0.7712,\"estimated_quality\":0.8975,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.236,\"max_latency_ms\":235264,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5137, "category": "coding", "complexity": 0.7297, "consequence": 0.7418, "evidence_need": 0.7157, "family": "migration_planning", "input_tokens": 280, "request_id": "synthetic-migration_planning-000418", "reversibility": 0.3421, "risk": 0.7125, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 3.046, "id": "worker_a", "p95_latency_ms": 78501, "pass_rate": 0.8421, "provider_group": "provider_z", "quality": 0.9083, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.561, "id": "worker_b", "p95_latency_ms": 88439, "pass_rate": 0.9465, "provider_group": "provider_x", "quality": 0.9232, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.022, "id": "worker_c", "p95_latency_ms": 29366, "pass_rate": 0.5565, "provider_group": "provider_y", "quality": 0.7774, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.332, "max_latency_ms": 25538, "minimum_quality": 0.5658}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000030", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.931, "estimated_latency_ms": 22160, "estimated_pass_rate": 0.29, "estimated_quality": 0.6319, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.332, "max_latency_ms": 25538, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.29, "quality": 0.6319, "quality_target_met": true, "utility": 39.148434}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.332,\"max_latency_ms\":25538,\"minimum_quality\":0.5658},\"task\":{\"ambiguity\":0.4852,\"category\":\"review\",\"complexity\":0.736,\"consequence\":0.648,\"evidence_need\":0.6695,\"family\":\"policy_compliance\",\"input_tokens\":7560,\"reversibility\":0.5382,\"risk\":0.7412,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":14.495,\"id\":\"worker_a\",\"p95_latency_ms\":88217,\"pass_rate\":0.9431,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.173,\"id\":\"worker_b\",\"p95_latency_ms\":64114,\"pass_rate\":0.7216,\"provider_group\":\"provider_z\",\"quality\":0.8867,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.931,\"id\":\"worker_c\",\"p95_latency_ms\":22160,\"pass_rate\":0.4702,\"provider_group\":\"provider_y\",\"quality\":0.7709,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.931,\"estimated_latency_ms\":22160,\"estimated_pass_rate\":0.29,\"estimated_quality\":0.6319,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.332,\"max_latency_ms\":25538,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4852, "category": "review", "complexity": 0.736, "consequence": 0.648, "evidence_need": 0.6695, "family": "policy_compliance", "input_tokens": 7560, "request_id": "synthetic-policy_compliance-000030", "reversibility": 0.5382, "risk": 0.7412, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 14.495, "id": "worker_a", "p95_latency_ms": 88217, "pass_rate": 0.9431, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.173, "id": "worker_b", "p95_latency_ms": 64114, "pass_rate": 0.7216, "provider_group": "provider_z", "quality": 0.8867, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.931, "id": "worker_c", "p95_latency_ms": 22160, "pass_rate": 0.4702, "provider_group": "provider_y", "quality": 0.7709, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.953, "max_latency_ms": 42668, "minimum_quality": 0.6924}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000279", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.044, "estimated_latency_ms": 29351, "estimated_pass_rate": 0.3478, "estimated_quality": 0.6854, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.953, "max_latency_ms": 42668, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3478, "quality": 0.6854, "quality_target_met": false, "utility": 44.629912}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.953,\"max_latency_ms\":42668,\"minimum_quality\":0.6924},\"task\":{\"ambiguity\":0.5397,\"category\":\"review\",\"complexity\":0.4206,\"consequence\":0.5371,\"evidence_need\":0.8019,\"family\":\"policy_compliance\",\"input_tokens\":5720,\"reversibility\":0.3037,\"risk\":0.5414,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.044,\"id\":\"worker_a\",\"p95_latency_ms\":29351,\"pass_rate\":0.4802,\"provider_group\":\"provider_x\",\"quality\":0.7567,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.171,\"id\":\"worker_b\",\"p95_latency_ms\":59532,\"pass_rate\":0.7422,\"provider_group\":\"provider_y\",\"quality\":0.84,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.416,\"id\":\"worker_c\",\"p95_latency_ms\":78796,\"pass_rate\":0.9843,\"provider_group\":\"provider_z\",\"quality\":0.9271,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.044,\"estimated_latency_ms\":29351,\"estimated_pass_rate\":0.3478,\"estimated_quality\":0.6854,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.953,\"max_latency_ms\":42668,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5397, "category": "review", "complexity": 0.4206, "consequence": 0.5371, "evidence_need": 0.8019, "family": "policy_compliance", "input_tokens": 5720, "request_id": "synthetic-policy_compliance-000279", "reversibility": 0.3037, "risk": 0.5414, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 1.044, "id": "worker_a", "p95_latency_ms": 29351, "pass_rate": 0.4802, "provider_group": "provider_x", "quality": 0.7567, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.171, "id": "worker_b", "p95_latency_ms": 59532, "pass_rate": 0.7422, "provider_group": "provider_y", "quality": 0.84, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.416, "id": "worker_c", "p95_latency_ms": 78796, "pass_rate": 0.9843, "provider_group": "provider_z", "quality": 0.9271, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.196, "max_latency_ms": 189126, "minimum_quality": 0.8352}, "family": "migration_planning", "id": "synthetic-migration_planning-001075", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.929, "estimated_latency_ms": 119880, "estimated_pass_rate": 0.8505, "estimated_quality": 0.8519, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.196, "max_latency_ms": 189126, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8505, "quality": 0.8519, "quality_target_met": true, "utility": 65.897784}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.196,\"max_latency_ms\":189126,\"minimum_quality\":0.8352},\"task\":{\"ambiguity\":0.4881,\"category\":\"coding\",\"complexity\":0.8226,\"consequence\":0.7875,\"evidence_need\":0.4516,\"family\":\"migration_planning\",\"input_tokens\":2240,\"reversibility\":0.4519,\"risk\":0.6194,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":14.255,\"id\":\"worker_a\",\"p95_latency_ms\":60768,\"pass_rate\":0.9606,\"provider_group\":\"provider_y\",\"quality\":0.9246,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.94,\"id\":\"worker_b\",\"p95_latency_ms\":23767,\"pass_rate\":0.5781,\"provider_group\":\"provider_x\",\"quality\":0.7218,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.674,\"id\":\"worker_c\",\"p95_latency_ms\":59112,\"pass_rate\":0.7314,\"provider_group\":\"provider_z\",\"quality\":0.7942,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.929,\"estimated_latency_ms\":119880,\"estimated_pass_rate\":0.8505,\"estimated_quality\":0.8519,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.196,\"max_latency_ms\":189126,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4881, "category": "coding", "complexity": 0.8226, "consequence": 0.7875, "evidence_need": 0.4516, "family": "migration_planning", "input_tokens": 2240, "request_id": "synthetic-migration_planning-001075", "reversibility": 0.4519, "risk": 0.6194, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 14.255, "id": "worker_a", "p95_latency_ms": 60768, "pass_rate": 0.9606, "provider_group": "provider_y", "quality": 0.9246, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.94, "id": "worker_b", "p95_latency_ms": 23767, "pass_rate": 0.5781, "provider_group": "provider_x", "quality": 0.7218, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.674, "id": "worker_c", "p95_latency_ms": 59112, "pass_rate": 0.7314, "provider_group": "provider_z", "quality": 0.7942, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.681, "max_latency_ms": 109884, "minimum_quality": 0.6977}, "family": "migration_planning", "id": "synthetic-migration_planning-000040", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.347, "estimated_latency_ms": 101380, "estimated_pass_rate": 0.6041, "estimated_quality": 0.757, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.681, "max_latency_ms": 109884, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6041, "quality": 0.757, "quality_target_met": true, "utility": 56.92729}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.681,\"max_latency_ms\":109884,\"minimum_quality\":0.6977},\"task\":{\"ambiguity\":0.3797,\"category\":\"coding\",\"complexity\":0.8755,\"consequence\":0.6581,\"evidence_need\":0.6289,\"family\":\"migration_planning\",\"input_tokens\":680,\"reversibility\":0.2845,\"risk\":0.6992,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.655,\"id\":\"worker_a\",\"p95_latency_ms\":60458,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9585,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.163,\"id\":\"worker_b\",\"p95_latency_ms\":69755,\"pass_rate\":0.7979,\"provider_group\":\"provider_z\",\"quality\":0.8607,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.184,\"id\":\"worker_c\",\"p95_latency_ms\":31625,\"pass_rate\":0.5139,\"provider_group\":\"provider_y\",\"quality\":0.6931,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.347,\"estimated_latency_ms\":101380,\"estimated_pass_rate\":0.6041,\"estimated_quality\":0.757,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.681,\"max_latency_ms\":109884,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3797, "category": "coding", "complexity": 0.8755, "consequence": 0.6581, "evidence_need": 0.6289, "family": "migration_planning", "input_tokens": 680, "request_id": "synthetic-migration_planning-000040", "reversibility": 0.2845, "risk": 0.6992, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 14.655, "id": "worker_a", "p95_latency_ms": 60458, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9585, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.163, "id": "worker_b", "p95_latency_ms": 69755, "pass_rate": 0.7979, "provider_group": "provider_z", "quality": 0.8607, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.184, "id": "worker_c", "p95_latency_ms": 31625, "pass_rate": 0.5139, "provider_group": "provider_y", "quality": 0.6931, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.537, "max_latency_ms": 41323, "minimum_quality": 0.6836}, "family": "migration_planning", "id": "synthetic-migration_planning-000382", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.963, "estimated_latency_ms": 23797, "estimated_pass_rate": 0.3009, "estimated_quality": 0.5954, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.537, "max_latency_ms": 41323, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3009, "quality": 0.5954, "quality_target_met": false, "utility": 34.882574}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.537,\"max_latency_ms\":41323,\"minimum_quality\":0.6836},\"task\":{\"ambiguity\":0.7447,\"category\":\"coding\",\"complexity\":0.6812,\"consequence\":0.8019,\"evidence_need\":0.6842,\"family\":\"migration_planning\",\"input_tokens\":7520,\"reversibility\":0.3265,\"risk\":0.5363,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.963,\"id\":\"worker_a\",\"p95_latency_ms\":23797,\"pass_rate\":0.4779,\"provider_group\":\"provider_y\",\"quality\":0.7272,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.082,\"id\":\"worker_b\",\"p95_latency_ms\":77393,\"pass_rate\":0.7562,\"provider_group\":\"provider_z\",\"quality\":0.8486,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.85,\"id\":\"worker_c\",\"p95_latency_ms\":79297,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9771,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.963,\"estimated_latency_ms\":23797,\"estimated_pass_rate\":0.3009,\"estimated_quality\":0.5954,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.537,\"max_latency_ms\":41323,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7447, "category": "coding", "complexity": 0.6812, "consequence": 0.8019, "evidence_need": 0.6842, "family": "migration_planning", "input_tokens": 7520, "request_id": "synthetic-migration_planning-000382", "reversibility": 0.3265, "risk": 0.5363, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 0.963, "id": "worker_a", "p95_latency_ms": 23797, "pass_rate": 0.4779, "provider_group": "provider_y", "quality": 0.7272, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.082, "id": "worker_b", "p95_latency_ms": 77393, "pass_rate": 0.7562, "provider_group": "provider_z", "quality": 0.8486, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.85, "id": "worker_c", "p95_latency_ms": 79297, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9771, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.359, "max_latency_ms": 34575, "minimum_quality": 0.562}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000668", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.185, "estimated_latency_ms": 19288, "estimated_pass_rate": 0.2612, "estimated_quality": 0.5975, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.359, "max_latency_ms": 34575, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2612, "quality": 0.5975, "quality_target_met": true, "utility": 37.025911}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.359,\"max_latency_ms\":34575,\"minimum_quality\":0.562},\"task\":{\"ambiguity\":0.452,\"category\":\"analysis\",\"complexity\":0.7001,\"consequence\":0.2568,\"evidence_need\":0.7034,\"family\":\"performance_tuning\",\"input_tokens\":560,\"reversibility\":0.6496,\"risk\":0.2684,\"summary\":\"Review a proposed optimization for credit reservation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.185,\"id\":\"worker_a\",\"p95_latency_ms\":19288,\"pass_rate\":0.3965,\"provider_group\":\"provider_y\",\"quality\":0.6938,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.738,\"id\":\"worker_b\",\"p95_latency_ms\":54143,\"pass_rate\":0.939,\"provider_group\":\"provider_x\",\"quality\":0.9517,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.482,\"id\":\"worker_c\",\"p95_latency_ms\":72443,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.185,\"estimated_latency_ms\":19288,\"estimated_pass_rate\":0.2612,\"estimated_quality\":0.5975,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.359,\"max_latency_ms\":34575,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.452, "category": "analysis", "complexity": 0.7001, "consequence": 0.2568, "evidence_need": 0.7034, "family": "performance_tuning", "input_tokens": 560, "request_id": "synthetic-performance_tuning-000668", "reversibility": 0.6496, "risk": 0.2684, "summary": "Review a proposed optimization for credit reservation for hidden regressions."}, "workers": [{"credits_per_call": 1.185, "id": "worker_a", "p95_latency_ms": 19288, "pass_rate": 0.3965, "provider_group": "provider_y", "quality": 0.6938, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.738, "id": "worker_b", "p95_latency_ms": 54143, "pass_rate": 0.939, "provider_group": "provider_x", "quality": 0.9517, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.482, "id": "worker_c", "p95_latency_ms": 72443, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.207, "max_latency_ms": 216837, "minimum_quality": 0.8787}, "family": "migration_planning", "id": "synthetic-migration_planning-000295", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.553, "estimated_latency_ms": 157601, "estimated_pass_rate": 0.7282, "estimated_quality": 0.9226, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.207, "max_latency_ms": 216837, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7282, "quality": 0.9226, "quality_target_met": true, "utility": 77.898365}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.207,\"max_latency_ms\":216837,\"minimum_quality\":0.8787},\"task\":{\"ambiguity\":0.5281,\"category\":\"coding\",\"complexity\":0.9052,\"consequence\":0.4856,\"evidence_need\":0.8305,\"family\":\"migration_planning\",\"input_tokens\":3600,\"reversibility\":0.4392,\"risk\":0.6548,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.183,\"id\":\"worker_a\",\"p95_latency_ms\":30073,\"pass_rate\":0.4656,\"provider_group\":\"provider_y\",\"quality\":0.7681,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.299,\"id\":\"worker_b\",\"p95_latency_ms\":58801,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9905,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.071,\"id\":\"worker_c\",\"p95_latency_ms\":68727,\"pass_rate\":0.7832,\"provider_group\":\"provider_x\",\"quality\":0.8857,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.553,\"estimated_latency_ms\":157601,\"estimated_pass_rate\":0.7282,\"estimated_quality\":0.9226,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.207,\"max_latency_ms\":216837,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5281, "category": "coding", "complexity": 0.9052, "consequence": 0.4856, "evidence_need": 0.8305, "family": "migration_planning", "input_tokens": 3600, "request_id": "synthetic-migration_planning-000295", "reversibility": 0.4392, "risk": 0.6548, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 1.183, "id": "worker_a", "p95_latency_ms": 30073, "pass_rate": 0.4656, "provider_group": "provider_y", "quality": 0.7681, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.299, "id": "worker_b", "p95_latency_ms": 58801, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9905, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.071, "id": "worker_c", "p95_latency_ms": 68727, "pass_rate": 0.7832, "provider_group": "provider_x", "quality": 0.8857, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.421, "max_latency_ms": 42165, "minimum_quality": 0.6464}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000458", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.908, "estimated_latency_ms": 32082, "estimated_pass_rate": 0.3062, "estimated_quality": 0.5437, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.421, "max_latency_ms": 42165, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3062, "quality": 0.5437, "quality_target_met": false, "utility": 31.156758}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.421,\"max_latency_ms\":42165,\"minimum_quality\":0.6464},\"task\":{\"ambiguity\":0.5296,\"category\":\"analysis\",\"complexity\":0.6365,\"consequence\":0.3893,\"evidence_need\":0.6743,\"family\":\"performance_tuning\",\"input_tokens\":1360,\"reversibility\":0.7141,\"risk\":0.4534,\"summary\":\"Review a proposed optimization for model routing for hidden regressions.\"},\"workers\":[{\"credits_per_call\":13.295,\"id\":\"worker_a\",\"p95_latency_ms\":58292,\"pass_rate\":0.9595,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.908,\"id\":\"worker_b\",\"p95_latency_ms\":32082,\"pass_rate\":0.4551,\"provider_group\":\"provider_z\",\"quality\":0.647,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.323,\"id\":\"worker_c\",\"p95_latency_ms\":72990,\"pass_rate\":0.8073,\"provider_group\":\"provider_y\",\"quality\":0.8848,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.908,\"estimated_latency_ms\":32082,\"estimated_pass_rate\":0.3062,\"estimated_quality\":0.5437,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.421,\"max_latency_ms\":42165,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5296, "category": "analysis", "complexity": 0.6365, "consequence": 0.3893, "evidence_need": 0.6743, "family": "performance_tuning", "input_tokens": 1360, "request_id": "synthetic-performance_tuning-000458", "reversibility": 0.7141, "risk": 0.4534, "summary": "Review a proposed optimization for model routing for hidden regressions."}, "workers": [{"credits_per_call": 13.295, "id": "worker_a", "p95_latency_ms": 58292, "pass_rate": 0.9595, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.908, "id": "worker_b", "p95_latency_ms": 32082, "pass_rate": 0.4551, "provider_group": "provider_z", "quality": 0.647, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.323, "id": "worker_c", "p95_latency_ms": 72990, "pass_rate": 0.8073, "provider_group": "provider_y", "quality": 0.8848, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.398, "max_latency_ms": 236393, "minimum_quality": 0.9462}, "family": "migration_planning", "id": "synthetic-migration_planning-001129", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.519, "estimated_latency_ms": 163604, "estimated_pass_rate": 0.7088, "estimated_quality": 0.8725, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 25.398, "max_latency_ms": 236393, "policy": "strict", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7088, "quality": 0.8725, "quality_target_met": false, "utility": 70.216026}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.398,\"max_latency_ms\":236393,\"minimum_quality\":0.9462},\"task\":{\"ambiguity\":0.6493,\"category\":\"coding\",\"complexity\":0.6999,\"consequence\":0.668,\"evidence_need\":0.7728,\"family\":\"migration_planning\",\"input_tokens\":4360,\"reversibility\":0.2587,\"risk\":0.6375,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.945,\"id\":\"worker_a\",\"p95_latency_ms\":76548,\"pass_rate\":0.831,\"provider_group\":\"provider_y\",\"quality\":0.8086,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.097,\"id\":\"worker_b\",\"p95_latency_ms\":24986,\"pass_rate\":0.4951,\"provider_group\":\"provider_z\",\"quality\":0.7042,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.477,\"id\":\"worker_c\",\"p95_latency_ms\":62070,\"pass_rate\":0.9194,\"provider_group\":\"provider_x\",\"quality\":0.916,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.519,\"estimated_latency_ms\":163604,\"estimated_pass_rate\":0.7088,\"estimated_quality\":0.8725,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.398,\"max_latency_ms\":236393,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6493, "category": "coding", "complexity": 0.6999, "consequence": 0.668, "evidence_need": 0.7728, "family": "migration_planning", "input_tokens": 4360, "request_id": "synthetic-migration_planning-001129", "reversibility": 0.2587, "risk": 0.6375, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 3.945, "id": "worker_a", "p95_latency_ms": 76548, "pass_rate": 0.831, "provider_group": "provider_y", "quality": 0.8086, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.097, "id": "worker_b", "p95_latency_ms": 24986, "pass_rate": 0.4951, "provider_group": "provider_z", "quality": 0.7042, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.477, "id": "worker_c", "p95_latency_ms": 62070, "pass_rate": 0.9194, "provider_group": "provider_x", "quality": 0.916, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.813, "max_latency_ms": 161861, "minimum_quality": 0.8291}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000975", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.661, "estimated_latency_ms": 137197, "estimated_pass_rate": 0.8465, "estimated_quality": 0.8779, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.813, "max_latency_ms": 161861, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8465, "quality": 0.8779, "quality_target_met": true, "utility": 67.859317}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.813,\"max_latency_ms\":161861,\"minimum_quality\":0.8291},\"task\":{\"ambiguity\":0.7544,\"category\":\"review\",\"complexity\":0.6997,\"consequence\":0.888,\"evidence_need\":0.7062,\"family\":\"policy_compliance\",\"input_tokens\":720,\"reversibility\":0.6053,\"risk\":0.5275,\"summary\":\"Identify evidence required before approving credit reservation for production.\"},\"workers\":[{\"credits_per_call\":13.102,\"id\":\"worker_a\",\"p95_latency_ms\":91725,\"pass_rate\":0.9444,\"provider_group\":\"provider_y\",\"quality\":0.9192,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.882,\"id\":\"worker_b\",\"p95_latency_ms\":27681,\"pass_rate\":0.4919,\"provider_group\":\"provider_z\",\"quality\":0.7637,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.559,\"id\":\"worker_c\",\"p95_latency_ms\":45472,\"pass_rate\":0.7647,\"provider_group\":\"provider_x\",\"quality\":0.8576,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.661,\"estimated_latency_ms\":137197,\"estimated_pass_rate\":0.8465,\"estimated_quality\":0.8779,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.813,\"max_latency_ms\":161861,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7544, "category": "review", "complexity": 0.6997, "consequence": 0.888, "evidence_need": 0.7062, "family": "policy_compliance", "input_tokens": 720, "request_id": "synthetic-policy_compliance-000975", "reversibility": 0.6053, "risk": 0.5275, "summary": "Identify evidence required before approving credit reservation for production."}, "workers": [{"credits_per_call": 13.102, "id": "worker_a", "p95_latency_ms": 91725, "pass_rate": 0.9444, "provider_group": "provider_y", "quality": 0.9192, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.882, "id": "worker_b", "p95_latency_ms": 27681, "pass_rate": 0.4919, "provider_group": "provider_z", "quality": 0.7637, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.559, "id": "worker_c", "p95_latency_ms": 45472, "pass_rate": 0.7647, "provider_group": "provider_x", "quality": 0.8576, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.412, "max_latency_ms": 221524, "minimum_quality": 0.8018}, "family": "migration_planning", "id": "synthetic-migration_planning-000550", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.726, "estimated_latency_ms": 169788, "estimated_pass_rate": 0.7094, "estimated_quality": 0.9101, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.412, "max_latency_ms": 221524, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7094, "quality": 0.9101, "quality_target_met": true, "utility": 73.623351}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.412,\"max_latency_ms\":221524,\"minimum_quality\":0.8018},\"task\":{\"ambiguity\":0.4494,\"category\":\"coding\",\"complexity\":0.946,\"consequence\":0.6756,\"evidence_need\":0.8176,\"family\":\"migration_planning\",\"input_tokens\":5360,\"reversibility\":0.3887,\"risk\":0.374,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":14.301,\"id\":\"worker_a\",\"p95_latency_ms\":83349,\"pass_rate\":0.8903,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.29,\"id\":\"worker_b\",\"p95_latency_ms\":59429,\"pass_rate\":0.7682,\"provider_group\":\"provider_z\",\"quality\":0.8313,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.135,\"id\":\"worker_c\",\"p95_latency_ms\":27010,\"pass_rate\":0.5902,\"provider_group\":\"provider_x\",\"quality\":0.7387,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.726,\"estimated_latency_ms\":169788,\"estimated_pass_rate\":0.7094,\"estimated_quality\":0.9101,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.412,\"max_latency_ms\":221524,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4494, "category": "coding", "complexity": 0.946, "consequence": 0.6756, "evidence_need": 0.8176, "family": "migration_planning", "input_tokens": 5360, "request_id": "synthetic-migration_planning-000550", "reversibility": 0.3887, "risk": 0.374, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 14.301, "id": "worker_a", "p95_latency_ms": 83349, "pass_rate": 0.8903, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.29, "id": "worker_b", "p95_latency_ms": 59429, "pass_rate": 0.7682, "provider_group": "provider_z", "quality": 0.8313, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.135, "id": "worker_c", "p95_latency_ms": 27010, "pass_rate": 0.5902, "provider_group": "provider_x", "quality": 0.7387, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.86, "max_latency_ms": 37459, "minimum_quality": 0.557}, "family": "migration_planning", "id": "synthetic-migration_planning-000448", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.092, "estimated_latency_ms": 20630, "estimated_pass_rate": 0.2863, "estimated_quality": 0.5988, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.86, "max_latency_ms": 37459, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2863, "quality": 0.5988, "quality_target_met": true, "utility": 38.559463}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.86,\"max_latency_ms\":37459,\"minimum_quality\":0.557},\"task\":{\"ambiguity\":0.4822,\"category\":\"coding\",\"complexity\":0.8848,\"consequence\":0.6318,\"evidence_need\":0.7369,\"family\":\"migration_planning\",\"input_tokens\":4280,\"reversibility\":0.4062,\"risk\":0.702,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.844,\"id\":\"worker_a\",\"p95_latency_ms\":56533,\"pass_rate\":0.8479,\"provider_group\":\"provider_y\",\"quality\":0.8233,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.092,\"id\":\"worker_b\",\"p95_latency_ms\":20630,\"pass_rate\":0.48,\"provider_group\":\"provider_z\",\"quality\":0.7617,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.085,\"id\":\"worker_c\",\"p95_latency_ms\":78853,\"pass_rate\":0.9909,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.092,\"estimated_latency_ms\":20630,\"estimated_pass_rate\":0.2863,\"estimated_quality\":0.5988,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.86,\"max_latency_ms\":37459,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4822, "category": "coding", "complexity": 0.8848, "consequence": 0.6318, "evidence_need": 0.7369, "family": "migration_planning", "input_tokens": 4280, "request_id": "synthetic-migration_planning-000448", "reversibility": 0.4062, "risk": 0.702, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.844, "id": "worker_a", "p95_latency_ms": 56533, "pass_rate": 0.8479, "provider_group": "provider_y", "quality": 0.8233, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.092, "id": "worker_b", "p95_latency_ms": 20630, "pass_rate": 0.48, "provider_group": "provider_z", "quality": 0.7617, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.085, "id": "worker_c", "p95_latency_ms": 78853, "pass_rate": 0.9909, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.123, "max_latency_ms": 205833, "minimum_quality": 0.9589}, "family": "migration_planning", "id": "synthetic-migration_planning-000007", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.083, "estimated_latency_ms": 154375, "estimated_pass_rate": 0.735, "estimated_quality": 0.9352, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.123, "max_latency_ms": 205833, "policy": "strict", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.735, "quality": 0.9352, "quality_target_met": false, "utility": 75.074318}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.123,\"max_latency_ms\":205833,\"minimum_quality\":0.9589},\"task\":{\"ambiguity\":0.5697,\"category\":\"coding\",\"complexity\":0.7089,\"consequence\":0.8208,\"evidence_need\":0.8415,\"family\":\"migration_planning\",\"input_tokens\":520,\"reversibility\":0.3428,\"risk\":0.4344,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.892,\"id\":\"worker_a\",\"p95_latency_ms\":28936,\"pass_rate\":0.5242,\"provider_group\":\"provider_y\",\"quality\":0.7936,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.097,\"id\":\"worker_b\",\"p95_latency_ms\":75050,\"pass_rate\":0.8261,\"provider_group\":\"provider_x\",\"quality\":0.858,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.094,\"id\":\"worker_c\",\"p95_latency_ms\":50389,\"pass_rate\":0.9663,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.083,\"estimated_latency_ms\":154375,\"estimated_pass_rate\":0.735,\"estimated_quality\":0.9352,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.123,\"max_latency_ms\":205833,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5697, "category": "coding", "complexity": 0.7089, "consequence": 0.8208, "evidence_need": 0.8415, "family": "migration_planning", "input_tokens": 520, "request_id": "synthetic-migration_planning-000007", "reversibility": 0.3428, "risk": 0.4344, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 0.892, "id": "worker_a", "p95_latency_ms": 28936, "pass_rate": 0.5242, "provider_group": "provider_y", "quality": 0.7936, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.097, "id": "worker_b", "p95_latency_ms": 75050, "pass_rate": 0.8261, "provider_group": "provider_x", "quality": 0.858, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.094, "id": "worker_c", "p95_latency_ms": 50389, "pass_rate": 0.9663, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.021, "max_latency_ms": 47640, "minimum_quality": 0.5634}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000183", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.796, "estimated_latency_ms": 27550, "estimated_pass_rate": 0.3167, "estimated_quality": 0.6034, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.021, "max_latency_ms": 47640, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3167, "quality": 0.6034, "quality_target_met": true, "utility": 22.639345}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.021,\"max_latency_ms\":47640,\"minimum_quality\":0.5634},\"task\":{\"ambiguity\":0.399,\"category\":\"review\",\"complexity\":0.5352,\"consequence\":0.9078,\"evidence_need\":0.6554,\"family\":\"policy_compliance\",\"input_tokens\":7640,\"reversibility\":0.318,\"risk\":0.8199,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":4.142,\"id\":\"worker_a\",\"p95_latency_ms\":74036,\"pass_rate\":0.7726,\"provider_group\":\"provider_x\",\"quality\":0.8251,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.765,\"id\":\"worker_b\",\"p95_latency_ms\":67104,\"pass_rate\":0.9643,\"provider_group\":\"provider_y\",\"quality\":0.9933,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.796,\"id\":\"worker_c\",\"p95_latency_ms\":27550,\"pass_rate\":0.474,\"provider_group\":\"provider_z\",\"quality\":0.706,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.796,\"estimated_latency_ms\":27550,\"estimated_pass_rate\":0.3167,\"estimated_quality\":0.6034,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.021,\"max_latency_ms\":47640,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.399, "category": "review", "complexity": 0.5352, "consequence": 0.9078, "evidence_need": 0.6554, "family": "policy_compliance", "input_tokens": 7640, "request_id": "synthetic-policy_compliance-000183", "reversibility": 0.318, "risk": 0.8199, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 4.142, "id": "worker_a", "p95_latency_ms": 74036, "pass_rate": 0.7726, "provider_group": "provider_x", "quality": 0.8251, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.765, "id": "worker_b", "p95_latency_ms": 67104, "pass_rate": 0.9643, "provider_group": "provider_y", "quality": 0.9933, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.796, "id": "worker_c", "p95_latency_ms": 27550, "pass_rate": 0.474, "provider_group": "provider_z", "quality": 0.706, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.997, "max_latency_ms": 160270, "minimum_quality": 0.8766}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000324", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.995, "estimated_latency_ms": 134897, "estimated_pass_rate": 0.7367, "estimated_quality": 0.9076, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.997, "max_latency_ms": 160270, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7367, "quality": 0.9076, "quality_target_met": true, "utility": 71.931911}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.997,\"max_latency_ms\":160270,\"minimum_quality\":0.8766},\"task\":{\"ambiguity\":0.7666,\"category\":\"review\",\"complexity\":0.6416,\"consequence\":0.5253,\"evidence_need\":0.563,\"family\":\"policy_compliance\",\"input_tokens\":2360,\"reversibility\":0.6874,\"risk\":0.6013,\"summary\":\"Identify evidence required before approving prompt caching for production.\"},\"workers\":[{\"credits_per_call\":3.784,\"id\":\"worker_a\",\"p95_latency_ms\":50695,\"pass_rate\":0.7681,\"provider_group\":\"provider_y\",\"quality\":0.8833,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.308,\"id\":\"worker_b\",\"p95_latency_ms\":58489,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.968,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.903,\"id\":\"worker_c\",\"p95_latency_ms\":25713,\"pass_rate\":0.4972,\"provider_group\":\"provider_x\",\"quality\":0.7705,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.995,\"estimated_latency_ms\":134897,\"estimated_pass_rate\":0.7367,\"estimated_quality\":0.9076,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.997,\"max_latency_ms\":160270,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7666, "category": "review", "complexity": 0.6416, "consequence": 0.5253, "evidence_need": 0.563, "family": "policy_compliance", "input_tokens": 2360, "request_id": "synthetic-policy_compliance-000324", "reversibility": 0.6874, "risk": 0.6013, "summary": "Identify evidence required before approving prompt caching for production."}, "workers": [{"credits_per_call": 3.784, "id": "worker_a", "p95_latency_ms": 50695, "pass_rate": 0.7681, "provider_group": "provider_y", "quality": 0.8833, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.308, "id": "worker_b", "p95_latency_ms": 58489, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.968, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.903, "id": "worker_c", "p95_latency_ms": 25713, "pass_rate": 0.4972, "provider_group": "provider_x", "quality": 0.7705, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.042, "max_latency_ms": 233738, "minimum_quality": 0.7575}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000996", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.166, "estimated_latency_ms": 91475, "estimated_pass_rate": 0.6322, "estimated_quality": 0.8006, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.042, "max_latency_ms": 233738, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6322, "quality": 0.8006, "quality_target_met": true, "utility": 67.277842}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.042,\"max_latency_ms\":233738,\"minimum_quality\":0.7575},\"task\":{\"ambiguity\":0.5271,\"category\":\"review\",\"complexity\":0.4604,\"consequence\":0.938,\"evidence_need\":0.9494,\"family\":\"policy_compliance\",\"input_tokens\":2480,\"reversibility\":0.5229,\"risk\":0.8184,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":3.001,\"id\":\"worker_a\",\"p95_latency_ms\":66003,\"pass_rate\":0.7508,\"provider_group\":\"provider_z\",\"quality\":0.8363,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.716,\"id\":\"worker_b\",\"p95_latency_ms\":73353,\"pass_rate\":0.9717,\"provider_group\":\"provider_y\",\"quality\":0.9894,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.165,\"id\":\"worker_c\",\"p95_latency_ms\":25472,\"pass_rate\":0.5462,\"provider_group\":\"provider_x\",\"quality\":0.6817,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.166,\"estimated_latency_ms\":91475,\"estimated_pass_rate\":0.6322,\"estimated_quality\":0.8006,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.042,\"max_latency_ms\":233738,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5271, "category": "review", "complexity": 0.4604, "consequence": 0.938, "evidence_need": 0.9494, "family": "policy_compliance", "input_tokens": 2480, "request_id": "synthetic-policy_compliance-000996", "reversibility": 0.5229, "risk": 0.8184, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 3.001, "id": "worker_a", "p95_latency_ms": 66003, "pass_rate": 0.7508, "provider_group": "provider_z", "quality": 0.8363, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.716, "id": "worker_b", "p95_latency_ms": 73353, "pass_rate": 0.9717, "provider_group": "provider_y", "quality": 0.9894, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.165, "id": "worker_c", "p95_latency_ms": 25472, "pass_rate": 0.5462, "provider_group": "provider_x", "quality": 0.6817, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.436, "max_latency_ms": 28582, "minimum_quality": 0.6865}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000009", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.057, "estimated_latency_ms": 20033, "estimated_pass_rate": 0.3063, "estimated_quality": 0.579, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.436, "max_latency_ms": 28582, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3063, "quality": 0.579, "quality_target_met": false, "utility": 30.121363}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.436,\"max_latency_ms\":28582,\"minimum_quality\":0.6865},\"task\":{\"ambiguity\":0.4036,\"category\":\"review\",\"complexity\":0.8365,\"consequence\":0.8248,\"evidence_need\":0.8236,\"family\":\"policy_compliance\",\"input_tokens\":280,\"reversibility\":0.5143,\"risk\":0.4602,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.057,\"id\":\"worker_a\",\"p95_latency_ms\":20033,\"pass_rate\":0.4689,\"provider_group\":\"provider_x\",\"quality\":0.7105,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.918,\"id\":\"worker_b\",\"p95_latency_ms\":77046,\"pass_rate\":0.9013,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.169,\"id\":\"worker_c\",\"p95_latency_ms\":64003,\"pass_rate\":0.8419,\"provider_group\":\"provider_y\",\"quality\":0.8864,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.057,\"estimated_latency_ms\":20033,\"estimated_pass_rate\":0.3063,\"estimated_quality\":0.579,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.436,\"max_latency_ms\":28582,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4036, "category": "review", "complexity": 0.8365, "consequence": 0.8248, "evidence_need": 0.8236, "family": "policy_compliance", "input_tokens": 280, "request_id": "synthetic-policy_compliance-000009", "reversibility": 0.5143, "risk": 0.4602, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 1.057, "id": "worker_a", "p95_latency_ms": 20033, "pass_rate": 0.4689, "provider_group": "provider_x", "quality": 0.7105, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.918, "id": "worker_b", "p95_latency_ms": 77046, "pass_rate": 0.9013, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.169, "id": "worker_c", "p95_latency_ms": 64003, "pass_rate": 0.8419, "provider_group": "provider_y", "quality": 0.8864, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.328, "max_latency_ms": 110953, "minimum_quality": 0.8638}, "family": "migration_planning", "id": "synthetic-migration_planning-001018", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.104, "estimated_latency_ms": 79060, "estimated_pass_rate": 0.7024, "estimated_quality": 0.8954, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.328, "max_latency_ms": 110953, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7024, "quality": 0.8954, "quality_target_met": true, "utility": 68.284946}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.328,\"max_latency_ms\":110953,\"minimum_quality\":0.8638},\"task\":{\"ambiguity\":0.5958,\"category\":\"coding\",\"complexity\":0.7002,\"consequence\":0.8208,\"evidence_need\":0.8304,\"family\":\"migration_planning\",\"input_tokens\":7320,\"reversibility\":0.647,\"risk\":0.7271,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.069,\"id\":\"worker_a\",\"p95_latency_ms\":64306,\"pass_rate\":0.872,\"provider_group\":\"provider_y\",\"quality\":0.9125,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.913,\"id\":\"worker_b\",\"p95_latency_ms\":57764,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9456,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.191,\"id\":\"worker_c\",\"p95_latency_ms\":21296,\"pass_rate\":0.4891,\"provider_group\":\"provider_x\",\"quality\":0.751,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.104,\"estimated_latency_ms\":79060,\"estimated_pass_rate\":0.7024,\"estimated_quality\":0.8954,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.328,\"max_latency_ms\":110953,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5958, "category": "coding", "complexity": 0.7002, "consequence": 0.8208, "evidence_need": 0.8304, "family": "migration_planning", "input_tokens": 7320, "request_id": "synthetic-migration_planning-001018", "reversibility": 0.647, "risk": 0.7271, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 4.069, "id": "worker_a", "p95_latency_ms": 64306, "pass_rate": 0.872, "provider_group": "provider_y", "quality": 0.9125, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.913, "id": "worker_b", "p95_latency_ms": 57764, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9456, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.191, "id": "worker_c", "p95_latency_ms": 21296, "pass_rate": 0.4891, "provider_group": "provider_x", "quality": 0.751, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.847, "max_latency_ms": 158923, "minimum_quality": 0.8385}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000885", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.791, "estimated_latency_ms": 144933, "estimated_pass_rate": 0.762, "estimated_quality": 0.884, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.847, "max_latency_ms": 158923, "policy": "strict", "reason_codes": ["independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.762, "quality": 0.884, "quality_target_met": true, "utility": 71.921879}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.847,\"max_latency_ms\":158923,\"minimum_quality\":0.8385},\"task\":{\"ambiguity\":0.4481,\"category\":\"review\",\"complexity\":0.5842,\"consequence\":0.6999,\"evidence_need\":0.6249,\"family\":\"policy_compliance\",\"input_tokens\":7600,\"reversibility\":0.6618,\"risk\":0.7111,\"summary\":\"Identify evidence required before approving streaming inference for production.\"},\"workers\":[{\"credits_per_call\":3.089,\"id\":\"worker_a\",\"p95_latency_ms\":73029,\"pass_rate\":0.7374,\"provider_group\":\"provider_y\",\"quality\":0.8634,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.802,\"id\":\"worker_b\",\"p95_latency_ms\":19535,\"pass_rate\":0.5416,\"provider_group\":\"provider_z\",\"quality\":0.7794,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.9,\"id\":\"worker_c\",\"p95_latency_ms\":52369,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9103,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.791,\"estimated_latency_ms\":144933,\"estimated_pass_rate\":0.762,\"estimated_quality\":0.884,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.847,\"max_latency_ms\":158923,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4481, "category": "review", "complexity": 0.5842, "consequence": 0.6999, "evidence_need": 0.6249, "family": "policy_compliance", "input_tokens": 7600, "request_id": "synthetic-policy_compliance-000885", "reversibility": 0.6618, "risk": 0.7111, "summary": "Identify evidence required before approving streaming inference for production."}, "workers": [{"credits_per_call": 3.089, "id": "worker_a", "p95_latency_ms": 73029, "pass_rate": 0.7374, "provider_group": "provider_y", "quality": 0.8634, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.802, "id": "worker_b", "p95_latency_ms": 19535, "pass_rate": 0.5416, "provider_group": "provider_z", "quality": 0.7794, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.9, "id": "worker_c", "p95_latency_ms": 52369, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9103, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.926, "max_latency_ms": 227461, "minimum_quality": 0.7948}, "family": "migration_planning", "id": "synthetic-migration_planning-000292", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.89, "estimated_latency_ms": 157021, "estimated_pass_rate": 0.7326, "estimated_quality": 0.9008, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.926, "max_latency_ms": 227461, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7326, "quality": 0.9008, "quality_target_met": true, "utility": 74.074921}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.926,\"max_latency_ms\":227461,\"minimum_quality\":0.7948},\"task\":{\"ambiguity\":0.4491,\"category\":\"coding\",\"complexity\":0.9339,\"consequence\":0.4184,\"evidence_need\":0.7066,\"family\":\"migration_planning\",\"input_tokens\":7320,\"reversibility\":0.6457,\"risk\":0.5997,\"summary\":\"Review a staged migration for access control and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.305,\"id\":\"worker_a\",\"p95_latency_ms\":49450,\"pass_rate\":0.855,\"provider_group\":\"provider_y\",\"quality\":0.8001,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.086,\"id\":\"worker_b\",\"p95_latency_ms\":20240,\"pass_rate\":0.562,\"provider_group\":\"provider_x\",\"quality\":0.6751,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.499,\"id\":\"worker_c\",\"p95_latency_ms\":87331,\"pass_rate\":0.9251,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.89,\"estimated_latency_ms\":157021,\"estimated_pass_rate\":0.7326,\"estimated_quality\":0.9008,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.926,\"max_latency_ms\":227461,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4491, "category": "coding", "complexity": 0.9339, "consequence": 0.4184, "evidence_need": 0.7066, "family": "migration_planning", "input_tokens": 7320, "request_id": "synthetic-migration_planning-000292", "reversibility": 0.6457, "risk": 0.5997, "summary": "Review a staged migration for access control and identify irreversible steps."}, "workers": [{"credits_per_call": 3.305, "id": "worker_a", "p95_latency_ms": 49450, "pass_rate": 0.855, "provider_group": "provider_y", "quality": 0.8001, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.086, "id": "worker_b", "p95_latency_ms": 20240, "pass_rate": 0.562, "provider_group": "provider_x", "quality": 0.6751, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.499, "id": "worker_c", "p95_latency_ms": 87331, "pass_rate": 0.9251, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.773, "max_latency_ms": 210780, "minimum_quality": 0.8563}, "family": "migration_planning", "id": "synthetic-migration_planning-001198", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.385, "estimated_latency_ms": 165295, "estimated_pass_rate": 0.7334, "estimated_quality": 0.9032, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.773, "max_latency_ms": 210780, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7334, "quality": 0.9032, "quality_target_met": true, "utility": 74.7978}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.773,\"max_latency_ms\":210780,\"minimum_quality\":0.8563},\"task\":{\"ambiguity\":0.4316,\"category\":\"coding\",\"complexity\":0.8634,\"consequence\":0.6855,\"evidence_need\":0.6058,\"family\":\"migration_planning\",\"input_tokens\":1320,\"reversibility\":0.4361,\"risk\":0.7163,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.534,\"id\":\"worker_a\",\"p95_latency_ms\":76108,\"pass_rate\":0.8169,\"provider_group\":\"provider_z\",\"quality\":0.8093,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.082,\"id\":\"worker_b\",\"p95_latency_ms\":28093,\"pass_rate\":0.5886,\"provider_group\":\"provider_y\",\"quality\":0.7805,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.769,\"id\":\"worker_c\",\"p95_latency_ms\":61094,\"pass_rate\":0.8879,\"provider_group\":\"provider_x\",\"quality\":0.9766,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.385,\"estimated_latency_ms\":165295,\"estimated_pass_rate\":0.7334,\"estimated_quality\":0.9032,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.773,\"max_latency_ms\":210780,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4316, "category": "coding", "complexity": 0.8634, "consequence": 0.6855, "evidence_need": 0.6058, "family": "migration_planning", "input_tokens": 1320, "request_id": "synthetic-migration_planning-001198", "reversibility": 0.4361, "risk": 0.7163, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 3.534, "id": "worker_a", "p95_latency_ms": 76108, "pass_rate": 0.8169, "provider_group": "provider_z", "quality": 0.8093, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.082, "id": "worker_b", "p95_latency_ms": 28093, "pass_rate": 0.5886, "provider_group": "provider_y", "quality": 0.7805, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.769, "id": "worker_c", "p95_latency_ms": 61094, "pass_rate": 0.8879, "provider_group": "provider_x", "quality": 0.9766, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.07, "max_latency_ms": 189393, "minimum_quality": 0.8672}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000254", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.686, "estimated_latency_ms": 160221, "estimated_pass_rate": 0.8853, "estimated_quality": 0.9147, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.07, "max_latency_ms": 189393, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8853, "quality": 0.9147, "quality_target_met": true, "utility": 66.36135}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.07,\"max_latency_ms\":189393,\"minimum_quality\":0.8672},\"task\":{\"ambiguity\":0.7467,\"category\":\"analysis\",\"complexity\":0.8883,\"consequence\":0.3032,\"evidence_need\":0.4453,\"family\":\"performance_tuning\",\"input_tokens\":2480,\"reversibility\":0.5891,\"risk\":0.2261,\"summary\":\"Diagnose latency variance in webhook delivery and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.036,\"id\":\"worker_a\",\"p95_latency_ms\":81816,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.992,\"id\":\"worker_b\",\"p95_latency_ms\":29693,\"pass_rate\":0.3946,\"provider_group\":\"provider_x\",\"quality\":0.6573,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.65,\"id\":\"worker_c\",\"p95_latency_ms\":78405,\"pass_rate\":0.879,\"provider_group\":\"provider_z\",\"quality\":0.9542,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.686,\"estimated_latency_ms\":160221,\"estimated_pass_rate\":0.8853,\"estimated_quality\":0.9147,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.07,\"max_latency_ms\":189393,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7467, "category": "analysis", "complexity": 0.8883, "consequence": 0.3032, "evidence_need": 0.4453, "family": "performance_tuning", "input_tokens": 2480, "request_id": "synthetic-performance_tuning-000254", "reversibility": 0.5891, "risk": 0.2261, "summary": "Diagnose latency variance in webhook delivery and rank the next experiments."}, "workers": [{"credits_per_call": 11.036, "id": "worker_a", "p95_latency_ms": 81816, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.992, "id": "worker_b", "p95_latency_ms": 29693, "pass_rate": 0.3946, "provider_group": "provider_x", "quality": 0.6573, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.65, "id": "worker_c", "p95_latency_ms": 78405, "pass_rate": 0.879, "provider_group": "provider_z", "quality": 0.9542, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.005, "max_latency_ms": 200549, "minimum_quality": 0.7619}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001020", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.808, "estimated_latency_ms": 150065, "estimated_pass_rate": 0.8276, "estimated_quality": 0.9223, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.005, "max_latency_ms": 200549, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8276, "quality": 0.9223, "quality_target_met": true, "utility": 71.899142}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.005,\"max_latency_ms\":200549,\"minimum_quality\":0.7619},\"task\":{\"ambiguity\":0.7725,\"category\":\"review\",\"complexity\":0.6117,\"consequence\":0.6932,\"evidence_need\":0.9181,\"family\":\"policy_compliance\",\"input_tokens\":7160,\"reversibility\":0.4443,\"risk\":0.8026,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":14.033,\"id\":\"worker_a\",\"p95_latency_ms\":74693,\"pass_rate\":0.8862,\"provider_group\":\"provider_y\",\"quality\":0.9643,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.889,\"id\":\"worker_b\",\"p95_latency_ms\":32558,\"pass_rate\":0.6057,\"provider_group\":\"provider_z\",\"quality\":0.803,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.775,\"id\":\"worker_c\",\"p95_latency_ms\":75372,\"pass_rate\":0.8439,\"provider_group\":\"provider_x\",\"quality\":0.8275,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.808,\"estimated_latency_ms\":150065,\"estimated_pass_rate\":0.8276,\"estimated_quality\":0.9223,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.005,\"max_latency_ms\":200549,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7725, "category": "review", "complexity": 0.6117, "consequence": 0.6932, "evidence_need": 0.9181, "family": "policy_compliance", "input_tokens": 7160, "request_id": "synthetic-policy_compliance-001020", "reversibility": 0.4443, "risk": 0.8026, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 14.033, "id": "worker_a", "p95_latency_ms": 74693, "pass_rate": 0.8862, "provider_group": "provider_y", "quality": 0.9643, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.889, "id": "worker_b", "p95_latency_ms": 32558, "pass_rate": 0.6057, "provider_group": "provider_z", "quality": 0.803, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.775, "id": "worker_c", "p95_latency_ms": 75372, "pass_rate": 0.8439, "provider_group": "provider_x", "quality": 0.8275, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.725, "max_latency_ms": 166044, "minimum_quality": 0.8615}, "family": "migration_planning", "id": "synthetic-migration_planning-000223", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.448, "estimated_latency_ms": 106271, "estimated_pass_rate": 0.9035, "estimated_quality": 0.9308, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.725, "max_latency_ms": 166044, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9035, "quality": 0.9308, "quality_target_met": true, "utility": 70.247993}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.725,\"max_latency_ms\":166044,\"minimum_quality\":0.8615},\"task\":{\"ambiguity\":0.4721,\"category\":\"coding\",\"complexity\":0.5832,\"consequence\":0.5851,\"evidence_need\":0.7095,\"family\":\"migration_planning\",\"input_tokens\":7880,\"reversibility\":0.6411,\"risk\":0.421,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.602,\"id\":\"worker_a\",\"p95_latency_ms\":49826,\"pass_rate\":0.8143,\"provider_group\":\"provider_x\",\"quality\":0.8024,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.948,\"id\":\"worker_b\",\"p95_latency_ms\":19882,\"pass_rate\":0.4652,\"provider_group\":\"provider_y\",\"quality\":0.7327,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.846,\"id\":\"worker_c\",\"p95_latency_ms\":56445,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9714,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.448,\"estimated_latency_ms\":106271,\"estimated_pass_rate\":0.9035,\"estimated_quality\":0.9308,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.725,\"max_latency_ms\":166044,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4721, "category": "coding", "complexity": 0.5832, "consequence": 0.5851, "evidence_need": 0.7095, "family": "migration_planning", "input_tokens": 7880, "request_id": "synthetic-migration_planning-000223", "reversibility": 0.6411, "risk": 0.421, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 3.602, "id": "worker_a", "p95_latency_ms": 49826, "pass_rate": 0.8143, "provider_group": "provider_x", "quality": 0.8024, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.948, "id": "worker_b", "p95_latency_ms": 19882, "pass_rate": 0.4652, "provider_group": "provider_y", "quality": 0.7327, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.846, "id": "worker_c", "p95_latency_ms": 56445, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9714, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.476, "max_latency_ms": 40197, "minimum_quality": 0.7419}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000479", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.866, "estimated_latency_ms": 27056, "estimated_pass_rate": 0.3394, "estimated_quality": 0.5915, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.476, "max_latency_ms": 40197, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3394, "quality": 0.5915, "quality_target_met": false, "utility": 33.389559}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.476,\"max_latency_ms\":40197,\"minimum_quality\":0.7419},\"task\":{\"ambiguity\":0.4307,\"category\":\"analysis\",\"complexity\":0.6513,\"consequence\":0.6239,\"evidence_need\":0.5092,\"family\":\"performance_tuning\",\"input_tokens\":2560,\"reversibility\":0.7149,\"risk\":0.2899,\"summary\":\"Diagnose latency variance in prompt caching and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":12.25,\"id\":\"worker_a\",\"p95_latency_ms\":83741,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.81,\"id\":\"worker_b\",\"p95_latency_ms\":64863,\"pass_rate\":0.8318,\"provider_group\":\"provider_z\",\"quality\":0.9535,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.866,\"id\":\"worker_c\",\"p95_latency_ms\":27056,\"pass_rate\":0.4693,\"provider_group\":\"provider_x\",\"quality\":0.6791,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.866,\"estimated_latency_ms\":27056,\"estimated_pass_rate\":0.3394,\"estimated_quality\":0.5915,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.476,\"max_latency_ms\":40197,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4307, "category": "analysis", "complexity": 0.6513, "consequence": 0.6239, "evidence_need": 0.5092, "family": "performance_tuning", "input_tokens": 2560, "request_id": "synthetic-performance_tuning-000479", "reversibility": 0.7149, "risk": 0.2899, "summary": "Diagnose latency variance in prompt caching and rank the next experiments."}, "workers": [{"credits_per_call": 12.25, "id": "worker_a", "p95_latency_ms": 83741, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.81, "id": "worker_b", "p95_latency_ms": 64863, "pass_rate": 0.8318, "provider_group": "provider_z", "quality": 0.9535, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.866, "id": "worker_c", "p95_latency_ms": 27056, "pass_rate": 0.4693, "provider_group": "provider_x", "quality": 0.6791, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.455, "max_latency_ms": 157770, "minimum_quality": 0.8365}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000585", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.337, "estimated_latency_ms": 143240, "estimated_pass_rate": 0.7018, "estimated_quality": 0.9098, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.455, "max_latency_ms": 157770, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7018, "quality": 0.9098, "quality_target_met": true, "utility": 76.271799}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.455,\"max_latency_ms\":157770,\"minimum_quality\":0.8365},\"task\":{\"ambiguity\":0.6273,\"category\":\"review\",\"complexity\":0.6318,\"consequence\":0.8583,\"evidence_need\":0.8873,\"family\":\"policy_compliance\",\"input_tokens\":2080,\"reversibility\":0.4341,\"risk\":0.6566,\"summary\":\"Review whether tenant-scoped caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.196,\"id\":\"worker_a\",\"p95_latency_ms\":69083,\"pass_rate\":0.9889,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.027,\"id\":\"worker_b\",\"p95_latency_ms\":44877,\"pass_rate\":0.7339,\"provider_group\":\"provider_x\",\"quality\":0.7948,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.114,\"id\":\"worker_c\",\"p95_latency_ms\":29280,\"pass_rate\":0.4551,\"provider_group\":\"provider_y\",\"quality\":0.7066,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.337,\"estimated_latency_ms\":143240,\"estimated_pass_rate\":0.7018,\"estimated_quality\":0.9098,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.455,\"max_latency_ms\":157770,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6273, "category": "review", "complexity": 0.6318, "consequence": 0.8583, "evidence_need": 0.8873, "family": "policy_compliance", "input_tokens": 2080, "request_id": "synthetic-policy_compliance-000585", "reversibility": 0.4341, "risk": 0.6566, "summary": "Review whether tenant-scoped caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.196, "id": "worker_a", "p95_latency_ms": 69083, "pass_rate": 0.9889, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.027, "id": "worker_b", "p95_latency_ms": 44877, "pass_rate": 0.7339, "provider_group": "provider_x", "quality": 0.7948, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.114, "id": "worker_c", "p95_latency_ms": 29280, "pass_rate": 0.4551, "provider_group": "provider_y", "quality": 0.7066, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.01, "max_latency_ms": 170784, "minimum_quality": 0.8261}, "family": "migration_planning", "id": "synthetic-migration_planning-000301", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.165, "estimated_latency_ms": 84794, "estimated_pass_rate": 0.6787, "estimated_quality": 0.816, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.01, "max_latency_ms": 170784, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6787, "quality": 0.816, "quality_target_met": false, "utility": 56.295532}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.01,\"max_latency_ms\":170784,\"minimum_quality\":0.8261},\"task\":{\"ambiguity\":0.3493,\"category\":\"coding\",\"complexity\":0.6287,\"consequence\":0.8361,\"evidence_need\":0.5298,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.5355,\"risk\":0.3978,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.146,\"id\":\"worker_a\",\"p95_latency_ms\":61788,\"pass_rate\":0.7468,\"provider_group\":\"provider_z\",\"quality\":0.8682,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.502,\"id\":\"worker_b\",\"p95_latency_ms\":78107,\"pass_rate\":0.8871,\"provider_group\":\"provider_y\",\"quality\":0.9638,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.019,\"id\":\"worker_c\",\"p95_latency_ms\":23006,\"pass_rate\":0.4769,\"provider_group\":\"provider_x\",\"quality\":0.7342,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.165,\"estimated_latency_ms\":84794,\"estimated_pass_rate\":0.6787,\"estimated_quality\":0.816,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.01,\"max_latency_ms\":170784,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3493, "category": "coding", "complexity": 0.6287, "consequence": 0.8361, "evidence_need": 0.5298, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-000301", "reversibility": 0.5355, "risk": 0.3978, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 4.146, "id": "worker_a", "p95_latency_ms": 61788, "pass_rate": 0.7468, "provider_group": "provider_z", "quality": 0.8682, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.502, "id": "worker_b", "p95_latency_ms": 78107, "pass_rate": 0.8871, "provider_group": "provider_y", "quality": 0.9638, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.019, "id": "worker_c", "p95_latency_ms": 23006, "pass_rate": 0.4769, "provider_group": "provider_x", "quality": 0.7342, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.966, "max_latency_ms": 212417, "minimum_quality": 0.8428}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001008", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.148, "estimated_latency_ms": 100175, "estimated_pass_rate": 0.5929, "estimated_quality": 0.749, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.966, "max_latency_ms": 212417, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5929, "quality": 0.749, "quality_target_met": false, "utility": 55.913366}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.966,\"max_latency_ms\":212417,\"minimum_quality\":0.8428},\"task\":{\"ambiguity\":0.5346,\"category\":\"review\",\"complexity\":0.8134,\"consequence\":0.8477,\"evidence_need\":0.647,\"family\":\"policy_compliance\",\"input_tokens\":5520,\"reversibility\":0.605,\"risk\":0.82,\"summary\":\"Review whether usage metering satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.03,\"id\":\"worker_a\",\"p95_latency_ms\":29508,\"pass_rate\":0.4941,\"provider_group\":\"provider_y\",\"quality\":0.6827,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.118,\"id\":\"worker_b\",\"p95_latency_ms\":70667,\"pass_rate\":0.7798,\"provider_group\":\"provider_x\",\"quality\":0.8499,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.867,\"id\":\"worker_c\",\"p95_latency_ms\":74771,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.148,\"estimated_latency_ms\":100175,\"estimated_pass_rate\":0.5929,\"estimated_quality\":0.749,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.966,\"max_latency_ms\":212417,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5346, "category": "review", "complexity": 0.8134, "consequence": 0.8477, "evidence_need": 0.647, "family": "policy_compliance", "input_tokens": 5520, "request_id": "synthetic-policy_compliance-001008", "reversibility": 0.605, "risk": 0.82, "summary": "Review whether usage metering satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.03, "id": "worker_a", "p95_latency_ms": 29508, "pass_rate": 0.4941, "provider_group": "provider_y", "quality": 0.6827, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.118, "id": "worker_b", "p95_latency_ms": 70667, "pass_rate": 0.7798, "provider_group": "provider_x", "quality": 0.8499, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.867, "id": "worker_c", "p95_latency_ms": 74771, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.951, "max_latency_ms": 222163, "minimum_quality": 0.8675}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000572", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.053, "estimated_latency_ms": 167605, "estimated_pass_rate": 0.7268, "estimated_quality": 0.9309, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.951, "max_latency_ms": 222163, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7268, "quality": 0.9309, "quality_target_met": true, "utility": 70.609805}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.951,\"max_latency_ms\":222163,\"minimum_quality\":0.8675},\"task\":{\"ambiguity\":0.6572,\"category\":\"analysis\",\"complexity\":0.6533,\"consequence\":0.3239,\"evidence_need\":0.7374,\"family\":\"performance_tuning\",\"input_tokens\":4520,\"reversibility\":0.739,\"risk\":0.3484,\"summary\":\"Diagnose latency variance in data retention and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.708,\"id\":\"worker_a\",\"p95_latency_ms\":75680,\"pass_rate\":0.9146,\"provider_group\":\"provider_y\",\"quality\":0.9197,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.439,\"id\":\"worker_b\",\"p95_latency_ms\":64170,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.906,\"id\":\"worker_c\",\"p95_latency_ms\":27755,\"pass_rate\":0.4198,\"provider_group\":\"provider_z\",\"quality\":0.6975,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.053,\"estimated_latency_ms\":167605,\"estimated_pass_rate\":0.7268,\"estimated_quality\":0.9309,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.951,\"max_latency_ms\":222163,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6572, "category": "analysis", "complexity": 0.6533, "consequence": 0.3239, "evidence_need": 0.7374, "family": "performance_tuning", "input_tokens": 4520, "request_id": "synthetic-performance_tuning-000572", "reversibility": 0.739, "risk": 0.3484, "summary": "Diagnose latency variance in data retention and rank the next experiments."}, "workers": [{"credits_per_call": 3.708, "id": "worker_a", "p95_latency_ms": 75680, "pass_rate": 0.9146, "provider_group": "provider_y", "quality": 0.9197, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.439, "id": "worker_b", "p95_latency_ms": 64170, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.906, "id": "worker_c", "p95_latency_ms": 27755, "pass_rate": 0.4198, "provider_group": "provider_z", "quality": 0.6975, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 23.836, "max_latency_ms": 173006, "minimum_quality": 0.8028}, "family": "migration_planning", "id": "synthetic-migration_planning-000823", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.585, "estimated_latency_ms": 146216, "estimated_pass_rate": 0.892, "estimated_quality": 0.9129, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 23.836, "max_latency_ms": 173006, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.892, "quality": 0.9129, "quality_target_met": true, "utility": 69.485005}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":23.836,\"max_latency_ms\":173006,\"minimum_quality\":0.8028},\"task\":{\"ambiguity\":0.7379,\"category\":\"coding\",\"complexity\":0.5981,\"consequence\":0.6684,\"evidence_need\":0.7787,\"family\":\"migration_planning\",\"input_tokens\":7360,\"reversibility\":0.3342,\"risk\":0.4741,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":15.364,\"id\":\"worker_a\",\"p95_latency_ms\":67022,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9468,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.221,\"id\":\"worker_b\",\"p95_latency_ms\":79194,\"pass_rate\":0.8433,\"provider_group\":\"provider_y\",\"quality\":0.8256,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.823,\"id\":\"worker_c\",\"p95_latency_ms\":20089,\"pass_rate\":0.5517,\"provider_group\":\"provider_z\",\"quality\":0.7646,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.585,\"estimated_latency_ms\":146216,\"estimated_pass_rate\":0.892,\"estimated_quality\":0.9129,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":23.836,\"max_latency_ms\":173006,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7379, "category": "coding", "complexity": 0.5981, "consequence": 0.6684, "evidence_need": 0.7787, "family": "migration_planning", "input_tokens": 7360, "request_id": "synthetic-migration_planning-000823", "reversibility": 0.3342, "risk": 0.4741, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 15.364, "id": "worker_a", "p95_latency_ms": 67022, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9468, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.221, "id": "worker_b", "p95_latency_ms": 79194, "pass_rate": 0.8433, "provider_group": "provider_y", "quality": 0.8256, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.823, "id": "worker_c", "p95_latency_ms": 20089, "pass_rate": 0.5517, "provider_group": "provider_z", "quality": 0.7646, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.506, "max_latency_ms": 196645, "minimum_quality": 0.8206}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000348", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.756, "estimated_latency_ms": 136484, "estimated_pass_rate": 0.7552, "estimated_quality": 0.9012, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.506, "max_latency_ms": 196645, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7552, "quality": 0.9012, "quality_target_met": true, "utility": 79.278451}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.506,\"max_latency_ms\":196645,\"minimum_quality\":0.8206},\"task\":{\"ambiguity\":0.6504,\"category\":\"review\",\"complexity\":0.7643,\"consequence\":0.6692,\"evidence_need\":0.7834,\"family\":\"policy_compliance\",\"input_tokens\":840,\"reversibility\":0.6345,\"risk\":0.7834,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.39,\"id\":\"worker_a\",\"p95_latency_ms\":56584,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9781,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.267,\"id\":\"worker_b\",\"p95_latency_ms\":49988,\"pass_rate\":0.7332,\"provider_group\":\"provider_x\",\"quality\":0.9118,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.099,\"id\":\"worker_c\",\"p95_latency_ms\":29912,\"pass_rate\":0.5263,\"provider_group\":\"provider_y\",\"quality\":0.6836,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.756,\"estimated_latency_ms\":136484,\"estimated_pass_rate\":0.7552,\"estimated_quality\":0.9012,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.506,\"max_latency_ms\":196645,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6504, "category": "review", "complexity": 0.7643, "consequence": 0.6692, "evidence_need": 0.7834, "family": "policy_compliance", "input_tokens": 840, "request_id": "synthetic-policy_compliance-000348", "reversibility": 0.6345, "risk": 0.7834, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.39, "id": "worker_a", "p95_latency_ms": 56584, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9781, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.267, "id": "worker_b", "p95_latency_ms": 49988, "pass_rate": 0.7332, "provider_group": "provider_x", "quality": 0.9118, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.099, "id": "worker_c", "p95_latency_ms": 29912, "pass_rate": 0.5263, "provider_group": "provider_y", "quality": 0.6836, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.792, "max_latency_ms": 31843, "minimum_quality": 0.5006}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000513", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.191, "estimated_latency_ms": 19336, "estimated_pass_rate": 0.3133, "estimated_quality": 0.5409, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.792, "max_latency_ms": 31843, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3133, "quality": 0.5409, "quality_target_met": true, "utility": 33.356825}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.792,\"max_latency_ms\":31843,\"minimum_quality\":0.5006},\"task\":{\"ambiguity\":0.616,\"category\":\"review\",\"complexity\":0.7939,\"consequence\":0.7164,\"evidence_need\":0.5677,\"family\":\"policy_compliance\",\"input_tokens\":8000,\"reversibility\":0.4578,\"risk\":0.6027,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.191,\"id\":\"worker_a\",\"p95_latency_ms\":19336,\"pass_rate\":0.4987,\"provider_group\":\"provider_y\",\"quality\":0.6892,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.143,\"id\":\"worker_b\",\"p95_latency_ms\":63550,\"pass_rate\":0.7672,\"provider_group\":\"provider_x\",\"quality\":0.8269,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.101,\"id\":\"worker_c\",\"p95_latency_ms\":52102,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9285,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.191,\"estimated_latency_ms\":19336,\"estimated_pass_rate\":0.3133,\"estimated_quality\":0.5409,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.792,\"max_latency_ms\":31843,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.616, "category": "review", "complexity": 0.7939, "consequence": 0.7164, "evidence_need": 0.5677, "family": "policy_compliance", "input_tokens": 8000, "request_id": "synthetic-policy_compliance-000513", "reversibility": 0.4578, "risk": 0.6027, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 1.191, "id": "worker_a", "p95_latency_ms": 19336, "pass_rate": 0.4987, "provider_group": "provider_y", "quality": 0.6892, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.143, "id": "worker_b", "p95_latency_ms": 63550, "pass_rate": 0.7672, "provider_group": "provider_x", "quality": 0.8269, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.101, "id": "worker_c", "p95_latency_ms": 52102, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9285, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.925, "max_latency_ms": 51902, "minimum_quality": 0.5345}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000038", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.942, "estimated_latency_ms": 30170, "estimated_pass_rate": 0.3825, "estimated_quality": 0.5639, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.925, "max_latency_ms": 51902, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3825, "quality": 0.5639, "quality_target_met": true, "utility": 37.71805}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.925,\"max_latency_ms\":51902,\"minimum_quality\":0.5345},\"task\":{\"ambiguity\":0.3406,\"category\":\"analysis\",\"complexity\":0.7267,\"consequence\":0.6165,\"evidence_need\":0.3281,\"family\":\"performance_tuning\",\"input_tokens\":6240,\"reversibility\":0.6864,\"risk\":0.2301,\"summary\":\"Review a proposed optimization for usage metering for hidden regressions.\"},\"workers\":[{\"credits_per_call\":0.942,\"id\":\"worker_a\",\"p95_latency_ms\":30170,\"pass_rate\":0.5093,\"provider_group\":\"provider_z\",\"quality\":0.6547,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.349,\"id\":\"worker_b\",\"p95_latency_ms\":76700,\"pass_rate\":0.8792,\"provider_group\":\"provider_y\",\"quality\":0.8816,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.643,\"id\":\"worker_c\",\"p95_latency_ms\":64491,\"pass_rate\":0.9723,\"provider_group\":\"provider_x\",\"quality\":0.9933,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.942,\"estimated_latency_ms\":30170,\"estimated_pass_rate\":0.3825,\"estimated_quality\":0.5639,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.925,\"max_latency_ms\":51902,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3406, "category": "analysis", "complexity": 0.7267, "consequence": 0.6165, "evidence_need": 0.3281, "family": "performance_tuning", "input_tokens": 6240, "request_id": "synthetic-performance_tuning-000038", "reversibility": 0.6864, "risk": 0.2301, "summary": "Review a proposed optimization for usage metering for hidden regressions."}, "workers": [{"credits_per_call": 0.942, "id": "worker_a", "p95_latency_ms": 30170, "pass_rate": 0.5093, "provider_group": "provider_z", "quality": 0.6547, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.349, "id": "worker_b", "p95_latency_ms": 76700, "pass_rate": 0.8792, "provider_group": "provider_y", "quality": 0.8816, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.643, "id": "worker_c", "p95_latency_ms": 64491, "pass_rate": 0.9723, "provider_group": "provider_x", "quality": 0.9933, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.788, "max_latency_ms": 195870, "minimum_quality": 0.8643}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000504", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.116, "estimated_latency_ms": 145251, "estimated_pass_rate": 0.6963, "estimated_quality": 0.8766, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.788, "max_latency_ms": 195870, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6963, "quality": 0.8766, "quality_target_met": true, "utility": 72.690552}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.788,\"max_latency_ms\":195870,\"minimum_quality\":0.8643},\"task\":{\"ambiguity\":0.6772,\"category\":\"review\",\"complexity\":0.7286,\"consequence\":0.9044,\"evidence_need\":0.5728,\"family\":\"policy_compliance\",\"input_tokens\":2360,\"reversibility\":0.4884,\"risk\":0.6433,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.866,\"id\":\"worker_a\",\"p95_latency_ms\":26454,\"pass_rate\":0.5178,\"provider_group\":\"provider_z\",\"quality\":0.7058,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.587,\"id\":\"worker_b\",\"p95_latency_ms\":52837,\"pass_rate\":0.9044,\"provider_group\":\"provider_x\",\"quality\":0.9568,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.663,\"id\":\"worker_c\",\"p95_latency_ms\":65960,\"pass_rate\":0.7497,\"provider_group\":\"provider_y\",\"quality\":0.8205,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.116,\"estimated_latency_ms\":145251,\"estimated_pass_rate\":0.6963,\"estimated_quality\":0.8766,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.788,\"max_latency_ms\":195870,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6772, "category": "review", "complexity": 0.7286, "consequence": 0.9044, "evidence_need": 0.5728, "family": "policy_compliance", "input_tokens": 2360, "request_id": "synthetic-policy_compliance-000504", "reversibility": 0.4884, "risk": 0.6433, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.866, "id": "worker_a", "p95_latency_ms": 26454, "pass_rate": 0.5178, "provider_group": "provider_z", "quality": 0.7058, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.587, "id": "worker_b", "p95_latency_ms": 52837, "pass_rate": 0.9044, "provider_group": "provider_x", "quality": 0.9568, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.663, "id": "worker_c", "p95_latency_ms": 65960, "pass_rate": 0.7497, "provider_group": "provider_y", "quality": 0.8205, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.722, "max_latency_ms": 40756, "minimum_quality": 0.4854}, "family": "migration_planning", "id": "synthetic-migration_planning-000217", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.109, "estimated_latency_ms": 25018, "estimated_pass_rate": 0.3077, "estimated_quality": 0.5758, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.722, "max_latency_ms": 40756, "policy": "lean", "reason_codes": ["complex_task", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3077, "quality": 0.5758, "quality_target_met": true, "utility": 37.928676}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.722,\"max_latency_ms\":40756,\"minimum_quality\":0.4854},\"task\":{\"ambiguity\":0.4353,\"category\":\"coding\",\"complexity\":0.7875,\"consequence\":0.7939,\"evidence_need\":0.5973,\"family\":\"migration_planning\",\"input_tokens\":3120,\"reversibility\":0.3538,\"risk\":0.6364,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":1.109,\"id\":\"worker_a\",\"p95_latency_ms\":25018,\"pass_rate\":0.4815,\"provider_group\":\"provider_x\",\"quality\":0.7133,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.323,\"id\":\"worker_b\",\"p95_latency_ms\":56648,\"pass_rate\":0.8715,\"provider_group\":\"provider_y\",\"quality\":0.8623,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.219,\"id\":\"worker_c\",\"p95_latency_ms\":63808,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9458,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.109,\"estimated_latency_ms\":25018,\"estimated_pass_rate\":0.3077,\"estimated_quality\":0.5758,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.722,\"max_latency_ms\":40756,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4353, "category": "coding", "complexity": 0.7875, "consequence": 0.7939, "evidence_need": 0.5973, "family": "migration_planning", "input_tokens": 3120, "request_id": "synthetic-migration_planning-000217", "reversibility": 0.3538, "risk": 0.6364, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 1.109, "id": "worker_a", "p95_latency_ms": 25018, "pass_rate": 0.4815, "provider_group": "provider_x", "quality": 0.7133, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.323, "id": "worker_b", "p95_latency_ms": 56648, "pass_rate": 0.8715, "provider_group": "provider_y", "quality": 0.8623, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.219, "id": "worker_c", "p95_latency_ms": 63808, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9458, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.155, "max_latency_ms": 53057, "minimum_quality": 0.6547}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000540", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.961, "estimated_latency_ms": 31852, "estimated_pass_rate": 0.2403, "estimated_quality": 0.5464, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.155, "max_latency_ms": 53057, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2403, "quality": 0.5464, "quality_target_met": false, "utility": 16.427101}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.155,\"max_latency_ms\":53057,\"minimum_quality\":0.6547},\"task\":{\"ambiguity\":0.772,\"category\":\"review\",\"complexity\":0.8269,\"consequence\":0.7589,\"evidence_need\":0.6956,\"family\":\"policy_compliance\",\"input_tokens\":280,\"reversibility\":0.5018,\"risk\":0.7413,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.283,\"id\":\"worker_a\",\"p95_latency_ms\":51357,\"pass_rate\":0.9701,\"provider_group\":\"provider_x\",\"quality\":0.9697,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.972,\"id\":\"worker_b\",\"p95_latency_ms\":52618,\"pass_rate\":0.8516,\"provider_group\":\"provider_z\",\"quality\":0.8454,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.961,\"id\":\"worker_c\",\"p95_latency_ms\":31852,\"pass_rate\":0.4525,\"provider_group\":\"provider_y\",\"quality\":0.7212,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.961,\"estimated_latency_ms\":31852,\"estimated_pass_rate\":0.2403,\"estimated_quality\":0.5464,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.155,\"max_latency_ms\":53057,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.772, "category": "review", "complexity": 0.8269, "consequence": 0.7589, "evidence_need": 0.6956, "family": "policy_compliance", "input_tokens": 280, "request_id": "synthetic-policy_compliance-000540", "reversibility": 0.5018, "risk": 0.7413, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 10.283, "id": "worker_a", "p95_latency_ms": 51357, "pass_rate": 0.9701, "provider_group": "provider_x", "quality": 0.9697, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.972, "id": "worker_b", "p95_latency_ms": 52618, "pass_rate": 0.8516, "provider_group": "provider_z", "quality": 0.8454, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.961, "id": "worker_c", "p95_latency_ms": 31852, "pass_rate": 0.4525, "provider_group": "provider_y", "quality": 0.7212, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.647, "max_latency_ms": 48937, "minimum_quality": 0.7689}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000689", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.018, "estimated_latency_ms": 31904, "estimated_pass_rate": 0.4133, "estimated_quality": 0.6082, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.647, "max_latency_ms": 48937, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4133, "quality": 0.6082, "quality_target_met": false, "utility": 33.204514}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.647,\"max_latency_ms\":48937,\"minimum_quality\":0.7689},\"task\":{\"ambiguity\":0.4362,\"category\":\"analysis\",\"complexity\":0.4546,\"consequence\":0.567,\"evidence_need\":0.4402,\"family\":\"performance_tuning\",\"input_tokens\":3240,\"reversibility\":0.781,\"risk\":0.3312,\"summary\":\"Choose between throughput and tail-latency changes for OAuth callbacks.\"},\"workers\":[{\"credits_per_call\":1.018,\"id\":\"worker_a\",\"p95_latency_ms\":31904,\"pass_rate\":0.5247,\"provider_group\":\"provider_z\",\"quality\":0.6636,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":13.528,\"id\":\"worker_b\",\"p95_latency_ms\":90420,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.072,\"id\":\"worker_c\",\"p95_latency_ms\":51067,\"pass_rate\":0.7851,\"provider_group\":\"provider_y\",\"quality\":0.9318,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.018,\"estimated_latency_ms\":31904,\"estimated_pass_rate\":0.4133,\"estimated_quality\":0.6082,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.647,\"max_latency_ms\":48937,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4362, "category": "analysis", "complexity": 0.4546, "consequence": 0.567, "evidence_need": 0.4402, "family": "performance_tuning", "input_tokens": 3240, "request_id": "synthetic-performance_tuning-000689", "reversibility": 0.781, "risk": 0.3312, "summary": "Choose between throughput and tail-latency changes for OAuth callbacks."}, "workers": [{"credits_per_call": 1.018, "id": "worker_a", "p95_latency_ms": 31904, "pass_rate": 0.5247, "provider_group": "provider_z", "quality": 0.6636, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 13.528, "id": "worker_b", "p95_latency_ms": 90420, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.072, "id": "worker_c", "p95_latency_ms": 51067, "pass_rate": 0.7851, "provider_group": "provider_y", "quality": 0.9318, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.684, "max_latency_ms": 230469, "minimum_quality": 0.9282}, "family": "migration_planning", "id": "synthetic-migration_planning-000379", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.326, "estimated_latency_ms": 166949, "estimated_pass_rate": 0.7391, "estimated_quality": 0.9026, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.684, "max_latency_ms": 230469, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7391, "quality": 0.9026, "quality_target_met": false, "utility": 71.049898}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.684,\"max_latency_ms\":230469,\"minimum_quality\":0.9282},\"task\":{\"ambiguity\":0.687,\"category\":\"coding\",\"complexity\":0.6319,\"consequence\":0.6749,\"evidence_need\":0.6004,\"family\":\"migration_planning\",\"input_tokens\":4640,\"reversibility\":0.4607,\"risk\":0.5184,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":4.016,\"id\":\"worker_a\",\"p95_latency_ms\":77269,\"pass_rate\":0.7637,\"provider_group\":\"provider_y\",\"quality\":0.8714,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.833,\"id\":\"worker_b\",\"p95_latency_ms\":20984,\"pass_rate\":0.5425,\"provider_group\":\"provider_x\",\"quality\":0.7909,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.477,\"id\":\"worker_c\",\"p95_latency_ms\":68696,\"pass_rate\":0.9686,\"provider_group\":\"provider_z\",\"quality\":0.9478,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.326,\"estimated_latency_ms\":166949,\"estimated_pass_rate\":0.7391,\"estimated_quality\":0.9026,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.684,\"max_latency_ms\":230469,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.687, "category": "coding", "complexity": 0.6319, "consequence": 0.6749, "evidence_need": 0.6004, "family": "migration_planning", "input_tokens": 4640, "request_id": "synthetic-migration_planning-000379", "reversibility": 0.4607, "risk": 0.5184, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 4.016, "id": "worker_a", "p95_latency_ms": 77269, "pass_rate": 0.7637, "provider_group": "provider_y", "quality": 0.8714, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.833, "id": "worker_b", "p95_latency_ms": 20984, "pass_rate": 0.5425, "provider_group": "provider_x", "quality": 0.7909, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.477, "id": "worker_c", "p95_latency_ms": 68696, "pass_rate": 0.9686, "provider_group": "provider_z", "quality": 0.9478, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.995, "max_latency_ms": 57344, "minimum_quality": 0.6203}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000041", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.889, "estimated_latency_ms": 29510, "estimated_pass_rate": 0.2631, "estimated_quality": 0.5126, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.995, "max_latency_ms": 57344, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2631, "quality": 0.5126, "quality_target_met": false, "utility": 28.915543}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.995,\"max_latency_ms\":57344,\"minimum_quality\":0.6203},\"task\":{\"ambiguity\":0.4606,\"category\":\"analysis\",\"complexity\":0.8366,\"consequence\":0.5774,\"evidence_need\":0.324,\"family\":\"performance_tuning\",\"input_tokens\":6360,\"reversibility\":0.5959,\"risk\":0.5869,\"summary\":\"Choose between throughput and tail-latency changes for usage metering.\"},\"workers\":[{\"credits_per_call\":3.424,\"id\":\"worker_a\",\"p95_latency_ms\":54346,\"pass_rate\":0.8047,\"provider_group\":\"provider_y\",\"quality\":0.9371,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.03,\"id\":\"worker_b\",\"p95_latency_ms\":69335,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.889,\"id\":\"worker_c\",\"p95_latency_ms\":29510,\"pass_rate\":0.4403,\"provider_group\":\"provider_z\",\"quality\":0.6571,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.889,\"estimated_latency_ms\":29510,\"estimated_pass_rate\":0.2631,\"estimated_quality\":0.5126,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.995,\"max_latency_ms\":57344,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4606, "category": "analysis", "complexity": 0.8366, "consequence": 0.5774, "evidence_need": 0.324, "family": "performance_tuning", "input_tokens": 6360, "request_id": "synthetic-performance_tuning-000041", "reversibility": 0.5959, "risk": 0.5869, "summary": "Choose between throughput and tail-latency changes for usage metering."}, "workers": [{"credits_per_call": 3.424, "id": "worker_a", "p95_latency_ms": 54346, "pass_rate": 0.8047, "provider_group": "provider_y", "quality": 0.9371, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.03, "id": "worker_b", "p95_latency_ms": 69335, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.889, "id": "worker_c", "p95_latency_ms": 29510, "pass_rate": 0.4403, "provider_group": "provider_z", "quality": 0.6571, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.547, "max_latency_ms": 156574, "minimum_quality": 0.8767}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000500", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.463, "estimated_latency_ms": 149058, "estimated_pass_rate": 0.875, "estimated_quality": 0.9357, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.547, "max_latency_ms": 156574, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.875, "quality": 0.9357, "quality_target_met": true, "utility": 71.976218}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.547,\"max_latency_ms\":156574,\"minimum_quality\":0.8767},\"task\":{\"ambiguity\":0.6168,\"category\":\"analysis\",\"complexity\":0.8274,\"consequence\":0.2826,\"evidence_need\":0.6339,\"family\":\"performance_tuning\",\"input_tokens\":7960,\"reversibility\":0.6991,\"risk\":0.2847,\"summary\":\"Diagnose latency variance in prompt caching and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":10.639,\"id\":\"worker_a\",\"p95_latency_ms\":77963,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.776,\"id\":\"worker_b\",\"p95_latency_ms\":51646,\"pass_rate\":0.9388,\"provider_group\":\"provider_y\",\"quality\":0.938,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.048,\"id\":\"worker_c\",\"p95_latency_ms\":19449,\"pass_rate\":0.4591,\"provider_group\":\"provider_z\",\"quality\":0.7174,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.463,\"estimated_latency_ms\":149058,\"estimated_pass_rate\":0.875,\"estimated_quality\":0.9357,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.547,\"max_latency_ms\":156574,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6168, "category": "analysis", "complexity": 0.8274, "consequence": 0.2826, "evidence_need": 0.6339, "family": "performance_tuning", "input_tokens": 7960, "request_id": "synthetic-performance_tuning-000500", "reversibility": 0.6991, "risk": 0.2847, "summary": "Diagnose latency variance in prompt caching and rank the next experiments."}, "workers": [{"credits_per_call": 10.639, "id": "worker_a", "p95_latency_ms": 77963, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.776, "id": "worker_b", "p95_latency_ms": 51646, "pass_rate": 0.9388, "provider_group": "provider_y", "quality": 0.938, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.048, "id": "worker_c", "p95_latency_ms": 19449, "pass_rate": 0.4591, "provider_group": "provider_z", "quality": 0.7174, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.563, "max_latency_ms": 155526, "minimum_quality": 0.7675}, "family": "migration_planning", "id": "synthetic-migration_planning-000799", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.9, "estimated_latency_ms": 140932, "estimated_pass_rate": 0.8612, "estimated_quality": 0.9204, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.563, "max_latency_ms": 155526, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8612, "quality": 0.9204, "quality_target_met": true, "utility": 68.656779}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.563,\"max_latency_ms\":155526,\"minimum_quality\":0.7675},\"task\":{\"ambiguity\":0.6336,\"category\":\"coding\",\"complexity\":0.6649,\"consequence\":0.6679,\"evidence_need\":0.7886,\"family\":\"migration_planning\",\"input_tokens\":2240,\"reversibility\":0.4875,\"risk\":0.3826,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":1.179,\"id\":\"worker_a\",\"p95_latency_ms\":28598,\"pass_rate\":0.4522,\"provider_group\":\"provider_z\",\"quality\":0.7874,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.183,\"id\":\"worker_b\",\"p95_latency_ms\":84577,\"pass_rate\":0.9486,\"provider_group\":\"provider_y\",\"quality\":0.9664,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.717,\"id\":\"worker_c\",\"p95_latency_ms\":56355,\"pass_rate\":0.7732,\"provider_group\":\"provider_x\",\"quality\":0.8767,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.9,\"estimated_latency_ms\":140932,\"estimated_pass_rate\":0.8612,\"estimated_quality\":0.9204,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.563,\"max_latency_ms\":155526,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6336, "category": "coding", "complexity": 0.6649, "consequence": 0.6679, "evidence_need": 0.7886, "family": "migration_planning", "input_tokens": 2240, "request_id": "synthetic-migration_planning-000799", "reversibility": 0.4875, "risk": 0.3826, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 1.179, "id": "worker_a", "p95_latency_ms": 28598, "pass_rate": 0.4522, "provider_group": "provider_z", "quality": 0.7874, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.183, "id": "worker_b", "p95_latency_ms": 84577, "pass_rate": 0.9486, "provider_group": "provider_y", "quality": 0.9664, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.717, "id": "worker_c", "p95_latency_ms": 56355, "pass_rate": 0.7732, "provider_group": "provider_x", "quality": 0.8767, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.532, "max_latency_ms": 254088, "minimum_quality": 0.8936}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000369", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.04, "estimated_latency_ms": 183060, "estimated_pass_rate": 0.7529, "estimated_quality": 0.9315, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.532, "max_latency_ms": 254088, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7529, "quality": 0.9315, "quality_target_met": true, "utility": 74.188745}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.532,\"max_latency_ms\":254088,\"minimum_quality\":0.8936},\"task\":{\"ambiguity\":0.4921,\"category\":\"review\",\"complexity\":0.5336,\"consequence\":0.7145,\"evidence_need\":0.7147,\"family\":\"policy_compliance\",\"input_tokens\":7120,\"reversibility\":0.3803,\"risk\":0.6005,\"summary\":\"Review whether schema evolution satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.256,\"id\":\"worker_a\",\"p95_latency_ms\":68767,\"pass_rate\":0.9089,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.641,\"id\":\"worker_b\",\"p95_latency_ms\":81638,\"pass_rate\":0.7939,\"provider_group\":\"provider_x\",\"quality\":0.906,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.143,\"id\":\"worker_c\",\"p95_latency_ms\":32655,\"pass_rate\":0.5704,\"provider_group\":\"provider_y\",\"quality\":0.7758,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.04,\"estimated_latency_ms\":183060,\"estimated_pass_rate\":0.7529,\"estimated_quality\":0.9315,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.532,\"max_latency_ms\":254088,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4921, "category": "review", "complexity": 0.5336, "consequence": 0.7145, "evidence_need": 0.7147, "family": "policy_compliance", "input_tokens": 7120, "request_id": "synthetic-policy_compliance-000369", "reversibility": 0.3803, "risk": 0.6005, "summary": "Review whether schema evolution satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.256, "id": "worker_a", "p95_latency_ms": 68767, "pass_rate": 0.9089, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.641, "id": "worker_b", "p95_latency_ms": 81638, "pass_rate": 0.7939, "provider_group": "provider_x", "quality": 0.906, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.143, "id": "worker_c", "p95_latency_ms": 32655, "pass_rate": 0.5704, "provider_group": "provider_y", "quality": 0.7758, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.566, "max_latency_ms": 37152, "minimum_quality": 0.626}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000810", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.199, "estimated_latency_ms": 19799, "estimated_pass_rate": 0.4309, "estimated_quality": 0.6119, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.566, "max_latency_ms": 37152, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4309, "quality": 0.6119, "quality_target_met": false, "utility": 41.587316}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.566,\"max_latency_ms\":37152,\"minimum_quality\":0.626},\"task\":{\"ambiguity\":0.7067,\"category\":\"review\",\"complexity\":0.6073,\"consequence\":0.6891,\"evidence_need\":0.829,\"family\":\"policy_compliance\",\"input_tokens\":4320,\"reversibility\":0.6272,\"risk\":0.5742,\"summary\":\"Identify evidence required before approving API-key rotation for production.\"},\"workers\":[{\"credits_per_call\":1.199,\"id\":\"worker_a\",\"p95_latency_ms\":19799,\"pass_rate\":0.5997,\"provider_group\":\"provider_z\",\"quality\":0.7305,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.733,\"id\":\"worker_b\",\"p95_latency_ms\":78097,\"pass_rate\":0.8222,\"provider_group\":\"provider_x\",\"quality\":0.814,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.758,\"id\":\"worker_c\",\"p95_latency_ms\":55331,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.99,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.199,\"estimated_latency_ms\":19799,\"estimated_pass_rate\":0.4309,\"estimated_quality\":0.6119,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.566,\"max_latency_ms\":37152,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7067, "category": "review", "complexity": 0.6073, "consequence": 0.6891, "evidence_need": 0.829, "family": "policy_compliance", "input_tokens": 4320, "request_id": "synthetic-policy_compliance-000810", "reversibility": 0.6272, "risk": 0.5742, "summary": "Identify evidence required before approving API-key rotation for production."}, "workers": [{"credits_per_call": 1.199, "id": "worker_a", "p95_latency_ms": 19799, "pass_rate": 0.5997, "provider_group": "provider_z", "quality": 0.7305, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.733, "id": "worker_b", "p95_latency_ms": 78097, "pass_rate": 0.8222, "provider_group": "provider_x", "quality": 0.814, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.758, "id": "worker_c", "p95_latency_ms": 55331, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.99, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.043, "max_latency_ms": 232359, "minimum_quality": 0.9125}, "family": "migration_planning", "id": "synthetic-migration_planning-000532", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.007, "estimated_latency_ms": 168407, "estimated_pass_rate": 0.722, "estimated_quality": 0.9251, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.043, "max_latency_ms": 232359, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.722, "quality": 0.9251, "quality_target_met": true, "utility": 72.485557}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.043,\"max_latency_ms\":232359,\"minimum_quality\":0.9125},\"task\":{\"ambiguity\":0.5463,\"category\":\"coding\",\"complexity\":0.6667,\"consequence\":0.6668,\"evidence_need\":0.789,\"family\":\"migration_planning\",\"input_tokens\":5440,\"reversibility\":0.3208,\"risk\":0.3998,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.866,\"id\":\"worker_a\",\"p95_latency_ms\":74154,\"pass_rate\":0.8989,\"provider_group\":\"provider_z\",\"quality\":0.9682,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.233,\"id\":\"worker_b\",\"p95_latency_ms\":69313,\"pass_rate\":0.8829,\"provider_group\":\"provider_y\",\"quality\":0.8566,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.908,\"id\":\"worker_c\",\"p95_latency_ms\":24940,\"pass_rate\":0.5315,\"provider_group\":\"provider_x\",\"quality\":0.7978,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.007,\"estimated_latency_ms\":168407,\"estimated_pass_rate\":0.722,\"estimated_quality\":0.9251,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.043,\"max_latency_ms\":232359,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5463, "category": "coding", "complexity": 0.6667, "consequence": 0.6668, "evidence_need": 0.789, "family": "migration_planning", "input_tokens": 5440, "request_id": "synthetic-migration_planning-000532", "reversibility": 0.3208, "risk": 0.3998, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 12.866, "id": "worker_a", "p95_latency_ms": 74154, "pass_rate": 0.8989, "provider_group": "provider_z", "quality": 0.9682, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.233, "id": "worker_b", "p95_latency_ms": 69313, "pass_rate": 0.8829, "provider_group": "provider_y", "quality": 0.8566, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.908, "id": "worker_c", "p95_latency_ms": 24940, "pass_rate": 0.5315, "provider_group": "provider_x", "quality": 0.7978, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.704, "max_latency_ms": 212596, "minimum_quality": 0.883}, "family": "migration_planning", "id": "synthetic-migration_planning-000625", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.191, "estimated_latency_ms": 148686, "estimated_pass_rate": 0.7725, "estimated_quality": 0.8801, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.704, "max_latency_ms": 212596, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7725, "quality": 0.8801, "quality_target_met": false, "utility": 72.803736}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.704,\"max_latency_ms\":212596,\"minimum_quality\":0.883},\"task\":{\"ambiguity\":0.3844,\"category\":\"coding\",\"complexity\":0.812,\"consequence\":0.5204,\"evidence_need\":0.6472,\"family\":\"migration_planning\",\"input_tokens\":4520,\"reversibility\":0.4825,\"risk\":0.6284,\"summary\":\"Review a staged migration for credit reservation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.214,\"id\":\"worker_a\",\"p95_latency_ms\":32714,\"pass_rate\":0.6028,\"provider_group\":\"provider_x\",\"quality\":0.7232,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.722,\"id\":\"worker_b\",\"p95_latency_ms\":51906,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9361,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.255,\"id\":\"worker_c\",\"p95_latency_ms\":64066,\"pass_rate\":0.7871,\"provider_group\":\"provider_y\",\"quality\":0.8028,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.191,\"estimated_latency_ms\":148686,\"estimated_pass_rate\":0.7725,\"estimated_quality\":0.8801,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.704,\"max_latency_ms\":212596,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3844, "category": "coding", "complexity": 0.812, "consequence": 0.5204, "evidence_need": 0.6472, "family": "migration_planning", "input_tokens": 4520, "request_id": "synthetic-migration_planning-000625", "reversibility": 0.4825, "risk": 0.6284, "summary": "Review a staged migration for credit reservation and identify irreversible steps."}, "workers": [{"credits_per_call": 1.214, "id": "worker_a", "p95_latency_ms": 32714, "pass_rate": 0.6028, "provider_group": "provider_x", "quality": 0.7232, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.722, "id": "worker_b", "p95_latency_ms": 51906, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9361, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.255, "id": "worker_c", "p95_latency_ms": 64066, "pass_rate": 0.7871, "provider_group": "provider_y", "quality": 0.8028, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.902, "max_latency_ms": 221346, "minimum_quality": 0.8797}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000413", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.038, "estimated_latency_ms": 76630, "estimated_pass_rate": 0.6967, "estimated_quality": 0.8285, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.902, "max_latency_ms": 221346, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6967, "quality": 0.8285, "quality_target_met": false, "utility": 56.310945}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.902,\"max_latency_ms\":221346,\"minimum_quality\":0.8797},\"task\":{\"ambiguity\":0.6445,\"category\":\"analysis\",\"complexity\":0.8977,\"consequence\":0.6287,\"evidence_need\":0.7393,\"family\":\"performance_tuning\",\"input_tokens\":4120,\"reversibility\":0.881,\"risk\":0.5208,\"summary\":\"Review a proposed optimization for webhook delivery for hidden regressions.\"},\"workers\":[{\"credits_per_call\":13.568,\"id\":\"worker_a\",\"p95_latency_ms\":88029,\"pass_rate\":0.9889,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.154,\"id\":\"worker_b\",\"p95_latency_ms\":29887,\"pass_rate\":0.5547,\"provider_group\":\"provider_z\",\"quality\":0.7283,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.884,\"id\":\"worker_c\",\"p95_latency_ms\":46743,\"pass_rate\":0.8117,\"provider_group\":\"provider_x\",\"quality\":0.9412,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.038,\"estimated_latency_ms\":76630,\"estimated_pass_rate\":0.6967,\"estimated_quality\":0.8285,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.902,\"max_latency_ms\":221346,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6445, "category": "analysis", "complexity": 0.8977, "consequence": 0.6287, "evidence_need": 0.7393, "family": "performance_tuning", "input_tokens": 4120, "request_id": "synthetic-performance_tuning-000413", "reversibility": 0.881, "risk": 0.5208, "summary": "Review a proposed optimization for webhook delivery for hidden regressions."}, "workers": [{"credits_per_call": 13.568, "id": "worker_a", "p95_latency_ms": 88029, "pass_rate": 0.9889, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.154, "id": "worker_b", "p95_latency_ms": 29887, "pass_rate": 0.5547, "provider_group": "provider_z", "quality": 0.7283, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.884, "id": "worker_c", "p95_latency_ms": 46743, "pass_rate": 0.8117, "provider_group": "provider_x", "quality": 0.9412, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.73, "max_latency_ms": 31088, "minimum_quality": 0.6734}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000180", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.196, "estimated_latency_ms": 21725, "estimated_pass_rate": 0.4236, "estimated_quality": 0.5961, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.73, "max_latency_ms": 31088, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4236, "quality": 0.5961, "quality_target_met": false, "utility": 35.470392}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.73,\"max_latency_ms\":31088,\"minimum_quality\":0.6734},\"task\":{\"ambiguity\":0.605,\"category\":\"review\",\"complexity\":0.6317,\"consequence\":0.5596,\"evidence_need\":0.6453,\"family\":\"policy_compliance\",\"input_tokens\":920,\"reversibility\":0.3338,\"risk\":0.8145,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.129,\"id\":\"worker_a\",\"p95_latency_ms\":58683,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.196,\"id\":\"worker_b\",\"p95_latency_ms\":21725,\"pass_rate\":0.6069,\"provider_group\":\"provider_z\",\"quality\":0.7296,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.493,\"id\":\"worker_c\",\"p95_latency_ms\":66359,\"pass_rate\":0.7485,\"provider_group\":\"provider_x\",\"quality\":0.8971,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.196,\"estimated_latency_ms\":21725,\"estimated_pass_rate\":0.4236,\"estimated_quality\":0.5961,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.73,\"max_latency_ms\":31088,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.605, "category": "review", "complexity": 0.6317, "consequence": 0.5596, "evidence_need": 0.6453, "family": "policy_compliance", "input_tokens": 920, "request_id": "synthetic-policy_compliance-000180", "reversibility": 0.3338, "risk": 0.8145, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.129, "id": "worker_a", "p95_latency_ms": 58683, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.196, "id": "worker_b", "p95_latency_ms": 21725, "pass_rate": 0.6069, "provider_group": "provider_z", "quality": 0.7296, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.493, "id": "worker_c", "p95_latency_ms": 66359, "pass_rate": 0.7485, "provider_group": "provider_x", "quality": 0.8971, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.353, "max_latency_ms": 31553, "minimum_quality": 0.5248}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000729", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.826, "estimated_latency_ms": 24325, "estimated_pass_rate": 0.369, "estimated_quality": 0.6045, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.353, "max_latency_ms": 31553, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.369, "quality": 0.6045, "quality_target_met": true, "utility": 40.039548}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.353,\"max_latency_ms\":31553,\"minimum_quality\":0.5248},\"task\":{\"ambiguity\":0.6104,\"category\":\"review\",\"complexity\":0.6865,\"consequence\":0.7431,\"evidence_need\":0.6674,\"family\":\"policy_compliance\",\"input_tokens\":1280,\"reversibility\":0.6553,\"risk\":0.47,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.398,\"id\":\"worker_a\",\"p95_latency_ms\":65268,\"pass_rate\":0.9353,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.223,\"id\":\"worker_b\",\"p95_latency_ms\":50412,\"pass_rate\":0.7893,\"provider_group\":\"provider_z\",\"quality\":0.8608,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.826,\"id\":\"worker_c\",\"p95_latency_ms\":24325,\"pass_rate\":0.5311,\"provider_group\":\"provider_x\",\"quality\":0.7235,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.826,\"estimated_latency_ms\":24325,\"estimated_pass_rate\":0.369,\"estimated_quality\":0.6045,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.353,\"max_latency_ms\":31553,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6104, "category": "review", "complexity": 0.6865, "consequence": 0.7431, "evidence_need": 0.6674, "family": "policy_compliance", "input_tokens": 1280, "request_id": "synthetic-policy_compliance-000729", "reversibility": 0.6553, "risk": 0.47, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 15.398, "id": "worker_a", "p95_latency_ms": 65268, "pass_rate": 0.9353, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.223, "id": "worker_b", "p95_latency_ms": 50412, "pass_rate": 0.7893, "provider_group": "provider_z", "quality": 0.8608, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.826, "id": "worker_c", "p95_latency_ms": 24325, "pass_rate": 0.5311, "provider_group": "provider_x", "quality": 0.7235, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.61, "max_latency_ms": 228081, "minimum_quality": 0.8166}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001040", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.145, "estimated_latency_ms": 158551, "estimated_pass_rate": 0.81, "estimated_quality": 0.9492, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.61, "max_latency_ms": 228081, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.81, "quality": 0.9492, "quality_target_met": true, "utility": 71.164999}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.61,\"max_latency_ms\":228081,\"minimum_quality\":0.8166},\"task\":{\"ambiguity\":0.3527,\"category\":\"analysis\",\"complexity\":0.7925,\"consequence\":0.6962,\"evidence_need\":0.3108,\"family\":\"performance_tuning\",\"input_tokens\":7200,\"reversibility\":0.5591,\"risk\":0.4047,\"summary\":\"Diagnose latency variance in webhook delivery and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.041,\"id\":\"worker_a\",\"p95_latency_ms\":79035,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.227,\"id\":\"worker_b\",\"p95_latency_ms\":54741,\"pass_rate\":0.7845,\"provider_group\":\"provider_y\",\"quality\":0.9817,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.877,\"id\":\"worker_c\",\"p95_latency_ms\":24775,\"pass_rate\":0.5026,\"provider_group\":\"provider_x\",\"quality\":0.725,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.145,\"estimated_latency_ms\":158551,\"estimated_pass_rate\":0.81,\"estimated_quality\":0.9492,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.61,\"max_latency_ms\":228081,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3527, "category": "analysis", "complexity": 0.7925, "consequence": 0.6962, "evidence_need": 0.3108, "family": "performance_tuning", "input_tokens": 7200, "request_id": "synthetic-performance_tuning-001040", "reversibility": 0.5591, "risk": 0.4047, "summary": "Diagnose latency variance in webhook delivery and rank the next experiments."}, "workers": [{"credits_per_call": 11.041, "id": "worker_a", "p95_latency_ms": 79035, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.227, "id": "worker_b", "p95_latency_ms": 54741, "pass_rate": 0.7845, "provider_group": "provider_y", "quality": 0.9817, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.877, "id": "worker_c", "p95_latency_ms": 24775, "pass_rate": 0.5026, "provider_group": "provider_x", "quality": 0.725, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.882, "max_latency_ms": 116709, "minimum_quality": 0.734}, "family": "migration_planning", "id": "synthetic-migration_planning-000598", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.961, "estimated_latency_ms": 101671, "estimated_pass_rate": 0.5856, "estimated_quality": 0.8345, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.882, "max_latency_ms": 116709, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5856, "quality": 0.8345, "quality_target_met": true, "utility": 59.407119}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.882,\"max_latency_ms\":116709,\"minimum_quality\":0.734},\"task\":{\"ambiguity\":0.3748,\"category\":\"coding\",\"complexity\":0.8344,\"consequence\":0.6602,\"evidence_need\":0.7136,\"family\":\"migration_planning\",\"input_tokens\":5680,\"reversibility\":0.3829,\"risk\":0.6279,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":3.062,\"id\":\"worker_a\",\"p95_latency_ms\":78019,\"pass_rate\":0.8494,\"provider_group\":\"provider_z\",\"quality\":0.8612,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.899,\"id\":\"worker_b\",\"p95_latency_ms\":23652,\"pass_rate\":0.4605,\"provider_group\":\"provider_y\",\"quality\":0.799,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.676,\"id\":\"worker_c\",\"p95_latency_ms\":67606,\"pass_rate\":0.9053,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.961,\"estimated_latency_ms\":101671,\"estimated_pass_rate\":0.5856,\"estimated_quality\":0.8345,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.882,\"max_latency_ms\":116709,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3748, "category": "coding", "complexity": 0.8344, "consequence": 0.6602, "evidence_need": 0.7136, "family": "migration_planning", "input_tokens": 5680, "request_id": "synthetic-migration_planning-000598", "reversibility": 0.3829, "risk": 0.6279, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 3.062, "id": "worker_a", "p95_latency_ms": 78019, "pass_rate": 0.8494, "provider_group": "provider_z", "quality": 0.8612, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.899, "id": "worker_b", "p95_latency_ms": 23652, "pass_rate": 0.4605, "provider_group": "provider_y", "quality": 0.799, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.676, "id": "worker_c", "p95_latency_ms": 67606, "pass_rate": 0.9053, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.793, "max_latency_ms": 162471, "minimum_quality": 0.8237}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000029", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.727, "estimated_latency_ms": 156519, "estimated_pass_rate": 0.7096, "estimated_quality": 0.9135, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.793, "max_latency_ms": 162471, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7096, "quality": 0.9135, "quality_target_met": true, "utility": 72.087539}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.793,\"max_latency_ms\":162471,\"minimum_quality\":0.8237},\"task\":{\"ambiguity\":0.37,\"category\":\"analysis\",\"complexity\":0.8831,\"consequence\":0.6891,\"evidence_need\":0.6536,\"family\":\"performance_tuning\",\"input_tokens\":960,\"reversibility\":0.7617,\"risk\":0.4441,\"summary\":\"Diagnose latency variance in database failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":9.969,\"id\":\"worker_a\",\"p95_latency_ms\":61044,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.822,\"id\":\"worker_b\",\"p95_latency_ms\":27920,\"pass_rate\":0.4078,\"provider_group\":\"provider_x\",\"quality\":0.6812,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.936,\"id\":\"worker_c\",\"p95_latency_ms\":67555,\"pass_rate\":0.8092,\"provider_group\":\"provider_y\",\"quality\":0.9309,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.727,\"estimated_latency_ms\":156519,\"estimated_pass_rate\":0.7096,\"estimated_quality\":0.9135,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.793,\"max_latency_ms\":162471,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.37, "category": "analysis", "complexity": 0.8831, "consequence": 0.6891, "evidence_need": 0.6536, "family": "performance_tuning", "input_tokens": 960, "request_id": "synthetic-performance_tuning-000029", "reversibility": 0.7617, "risk": 0.4441, "summary": "Diagnose latency variance in database failover and rank the next experiments."}, "workers": [{"credits_per_call": 9.969, "id": "worker_a", "p95_latency_ms": 61044, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.822, "id": "worker_b", "p95_latency_ms": 27920, "pass_rate": 0.4078, "provider_group": "provider_x", "quality": 0.6812, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.936, "id": "worker_c", "p95_latency_ms": 67555, "pass_rate": 0.8092, "provider_group": "provider_y", "quality": 0.9309, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.1, "max_latency_ms": 214493, "minimum_quality": 0.8601}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000221", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.024, "estimated_latency_ms": 183486, "estimated_pass_rate": 0.7006, "estimated_quality": 0.9113, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.1, "max_latency_ms": 214493, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7006, "quality": 0.9113, "quality_target_met": true, "utility": 70.319915}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.1,\"max_latency_ms\":214493,\"minimum_quality\":0.8601},\"task\":{\"ambiguity\":0.4438,\"category\":\"analysis\",\"complexity\":0.8838,\"consequence\":0.5381,\"evidence_need\":0.5419,\"family\":\"performance_tuning\",\"input_tokens\":6680,\"reversibility\":0.8306,\"risk\":0.4094,\"summary\":\"Review a proposed optimization for streaming inference for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.706,\"id\":\"worker_a\",\"p95_latency_ms\":61484,\"pass_rate\":0.7899,\"provider_group\":\"provider_y\",\"quality\":0.9768,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.856,\"id\":\"worker_b\",\"p95_latency_ms\":31725,\"pass_rate\":0.3867,\"provider_group\":\"provider_x\",\"quality\":0.6855,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.462,\"id\":\"worker_c\",\"p95_latency_ms\":90277,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.024,\"estimated_latency_ms\":183486,\"estimated_pass_rate\":0.7006,\"estimated_quality\":0.9113,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.1,\"max_latency_ms\":214493,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4438, "category": "analysis", "complexity": 0.8838, "consequence": 0.5381, "evidence_need": 0.5419, "family": "performance_tuning", "input_tokens": 6680, "request_id": "synthetic-performance_tuning-000221", "reversibility": 0.8306, "risk": 0.4094, "summary": "Review a proposed optimization for streaming inference for hidden regressions."}, "workers": [{"credits_per_call": 3.706, "id": "worker_a", "p95_latency_ms": 61484, "pass_rate": 0.7899, "provider_group": "provider_y", "quality": 0.9768, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.856, "id": "worker_b", "p95_latency_ms": 31725, "pass_rate": 0.3867, "provider_group": "provider_x", "quality": 0.6855, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.462, "id": "worker_c", "p95_latency_ms": 90277, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.146, "max_latency_ms": 42757, "minimum_quality": 0.5956}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000614", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.788, "estimated_latency_ms": 24002, "estimated_pass_rate": 0.3033, "estimated_quality": 0.59, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.146, "max_latency_ms": 42757, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3033, "quality": 0.59, "quality_target_met": false, "utility": 38.224224}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.146,\"max_latency_ms\":42757,\"minimum_quality\":0.5956},\"task\":{\"ambiguity\":0.7321,\"category\":\"analysis\",\"complexity\":0.7388,\"consequence\":0.2559,\"evidence_need\":0.3256,\"family\":\"performance_tuning\",\"input_tokens\":5480,\"reversibility\":0.6426,\"risk\":0.4808,\"summary\":\"Review a proposed optimization for streaming inference for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.469,\"id\":\"worker_a\",\"p95_latency_ms\":59314,\"pass_rate\":0.8405,\"provider_group\":\"provider_x\",\"quality\":0.9685,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.788,\"id\":\"worker_b\",\"p95_latency_ms\":24002,\"pass_rate\":0.4814,\"provider_group\":\"provider_y\",\"quality\":0.7274,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.749,\"id\":\"worker_c\",\"p95_latency_ms\":82015,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.788,\"estimated_latency_ms\":24002,\"estimated_pass_rate\":0.3033,\"estimated_quality\":0.59,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.146,\"max_latency_ms\":42757,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7321, "category": "analysis", "complexity": 0.7388, "consequence": 0.2559, "evidence_need": 0.3256, "family": "performance_tuning", "input_tokens": 5480, "request_id": "synthetic-performance_tuning-000614", "reversibility": 0.6426, "risk": 0.4808, "summary": "Review a proposed optimization for streaming inference for hidden regressions."}, "workers": [{"credits_per_call": 3.469, "id": "worker_a", "p95_latency_ms": 59314, "pass_rate": 0.8405, "provider_group": "provider_x", "quality": 0.9685, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.788, "id": "worker_b", "p95_latency_ms": 24002, "pass_rate": 0.4814, "provider_group": "provider_y", "quality": 0.7274, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.749, "id": "worker_c", "p95_latency_ms": 82015, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.94, "max_latency_ms": 41465, "minimum_quality": 0.703}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000599", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.039, "estimated_latency_ms": 30225, "estimated_pass_rate": 0.3141, "estimated_quality": 0.5888, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.94, "max_latency_ms": 41465, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3141, "quality": 0.5888, "quality_target_met": false, "utility": 33.179514}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.94,\"max_latency_ms\":41465,\"minimum_quality\":0.703},\"task\":{\"ambiguity\":0.4501,\"category\":\"analysis\",\"complexity\":0.8919,\"consequence\":0.3712,\"evidence_need\":0.4918,\"family\":\"performance_tuning\",\"input_tokens\":5920,\"reversibility\":0.8362,\"risk\":0.3142,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.196,\"id\":\"worker_a\",\"p95_latency_ms\":79670,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9839,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.772,\"id\":\"worker_b\",\"p95_latency_ms\":46114,\"pass_rate\":0.9129,\"provider_group\":\"provider_y\",\"quality\":0.9372,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.039,\"id\":\"worker_c\",\"p95_latency_ms\":30225,\"pass_rate\":0.4747,\"provider_group\":\"provider_x\",\"quality\":0.7229,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.039,\"estimated_latency_ms\":30225,\"estimated_pass_rate\":0.3141,\"estimated_quality\":0.5888,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.94,\"max_latency_ms\":41465,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4501, "category": "analysis", "complexity": 0.8919, "consequence": 0.3712, "evidence_need": 0.4918, "family": "performance_tuning", "input_tokens": 5920, "request_id": "synthetic-performance_tuning-000599", "reversibility": 0.8362, "risk": 0.3142, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 11.196, "id": "worker_a", "p95_latency_ms": 79670, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9839, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.772, "id": "worker_b", "p95_latency_ms": 46114, "pass_rate": 0.9129, "provider_group": "provider_y", "quality": 0.9372, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.039, "id": "worker_c", "p95_latency_ms": 30225, "pass_rate": 0.4747, "provider_group": "provider_x", "quality": 0.7229, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.052, "max_latency_ms": 167246, "minimum_quality": 0.8625}, "family": "migration_planning", "id": "synthetic-migration_planning-000460", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.516, "estimated_latency_ms": 123320, "estimated_pass_rate": 0.7538, "estimated_quality": 0.9042, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.052, "max_latency_ms": 167246, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7538, "quality": 0.9042, "quality_target_met": true, "utility": 73.475451}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.052,\"max_latency_ms\":167246,\"minimum_quality\":0.8625},\"task\":{\"ambiguity\":0.5632,\"category\":\"coding\",\"complexity\":0.702,\"consequence\":0.6347,\"evidence_need\":0.4289,\"family\":\"migration_planning\",\"input_tokens\":6640,\"reversibility\":0.6183,\"risk\":0.6413,\"summary\":\"Design compatibility testing for a artifact storage schema migration.\"},\"workers\":[{\"credits_per_call\":11.153,\"id\":\"worker_a\",\"p95_latency_ms\":52423,\"pass_rate\":0.9697,\"provider_group\":\"provider_x\",\"quality\":0.9829,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.348,\"id\":\"worker_b\",\"p95_latency_ms\":49697,\"pass_rate\":0.8079,\"provider_group\":\"provider_z\",\"quality\":0.8026,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.015,\"id\":\"worker_c\",\"p95_latency_ms\":21200,\"pass_rate\":0.5702,\"provider_group\":\"provider_y\",\"quality\":0.7599,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.516,\"estimated_latency_ms\":123320,\"estimated_pass_rate\":0.7538,\"estimated_quality\":0.9042,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.052,\"max_latency_ms\":167246,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5632, "category": "coding", "complexity": 0.702, "consequence": 0.6347, "evidence_need": 0.4289, "family": "migration_planning", "input_tokens": 6640, "request_id": "synthetic-migration_planning-000460", "reversibility": 0.6183, "risk": 0.6413, "summary": "Design compatibility testing for a artifact storage schema migration."}, "workers": [{"credits_per_call": 11.153, "id": "worker_a", "p95_latency_ms": 52423, "pass_rate": 0.9697, "provider_group": "provider_x", "quality": 0.9829, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.348, "id": "worker_b", "p95_latency_ms": 49697, "pass_rate": 0.8079, "provider_group": "provider_z", "quality": 0.8026, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.015, "id": "worker_c", "p95_latency_ms": 21200, "pass_rate": 0.5702, "provider_group": "provider_y", "quality": 0.7599, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.266, "max_latency_ms": 236375, "minimum_quality": 0.7525}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001257", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.431, "estimated_latency_ms": 131154, "estimated_pass_rate": 0.8789, "estimated_quality": 0.8876, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.266, "max_latency_ms": 236375, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8789, "quality": 0.8876, "quality_target_met": true, "utility": 69.046541}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.266,\"max_latency_ms\":236375,\"minimum_quality\":0.7525},\"task\":{\"ambiguity\":0.5208,\"category\":\"review\",\"complexity\":0.834,\"consequence\":0.685,\"evidence_need\":0.7065,\"family\":\"policy_compliance\",\"input_tokens\":6640,\"reversibility\":0.4159,\"risk\":0.625,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.773,\"id\":\"worker_a\",\"p95_latency_ms\":80062,\"pass_rate\":0.7231,\"provider_group\":\"provider_x\",\"quality\":0.8909,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.964,\"id\":\"worker_b\",\"p95_latency_ms\":26696,\"pass_rate\":0.5836,\"provider_group\":\"provider_z\",\"quality\":0.7674,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.658,\"id\":\"worker_c\",\"p95_latency_ms\":51092,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9493,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.431,\"estimated_latency_ms\":131154,\"estimated_pass_rate\":0.8789,\"estimated_quality\":0.8876,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.266,\"max_latency_ms\":236375,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5208, "category": "review", "complexity": 0.834, "consequence": 0.685, "evidence_need": 0.7065, "family": "policy_compliance", "input_tokens": 6640, "request_id": "synthetic-policy_compliance-001257", "reversibility": 0.4159, "risk": 0.625, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.773, "id": "worker_a", "p95_latency_ms": 80062, "pass_rate": 0.7231, "provider_group": "provider_x", "quality": 0.8909, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.964, "id": "worker_b", "p95_latency_ms": 26696, "pass_rate": 0.5836, "provider_group": "provider_z", "quality": 0.7674, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.658, "id": "worker_c", "p95_latency_ms": 51092, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9493, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.486, "max_latency_ms": 181763, "minimum_quality": 0.7954}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000051", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.476, "estimated_latency_ms": 173525, "estimated_pass_rate": 0.7543, "estimated_quality": 0.8852, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.486, "max_latency_ms": 181763, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7543, "quality": 0.8852, "quality_target_met": true, "utility": 76.153507}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.486,\"max_latency_ms\":181763,\"minimum_quality\":0.7954},\"task\":{\"ambiguity\":0.4527,\"category\":\"review\",\"complexity\":0.8266,\"consequence\":0.9477,\"evidence_need\":0.775,\"family\":\"policy_compliance\",\"input_tokens\":440,\"reversibility\":0.5026,\"risk\":0.634,\"summary\":\"Identify evidence required before approving background jobs for production.\"},\"workers\":[{\"credits_per_call\":12.748,\"id\":\"worker_a\",\"p95_latency_ms\":81447,\"pass_rate\":0.9147,\"provider_group\":\"provider_y\",\"quality\":0.9107,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.746,\"id\":\"worker_b\",\"p95_latency_ms\":72037,\"pass_rate\":0.8222,\"provider_group\":\"provider_z\",\"quality\":0.8735,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.982,\"id\":\"worker_c\",\"p95_latency_ms\":20041,\"pass_rate\":0.5938,\"provider_group\":\"provider_x\",\"quality\":0.7711,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.476,\"estimated_latency_ms\":173525,\"estimated_pass_rate\":0.7543,\"estimated_quality\":0.8852,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.486,\"max_latency_ms\":181763,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4527, "category": "review", "complexity": 0.8266, "consequence": 0.9477, "evidence_need": 0.775, "family": "policy_compliance", "input_tokens": 440, "request_id": "synthetic-policy_compliance-000051", "reversibility": 0.5026, "risk": 0.634, "summary": "Identify evidence required before approving background jobs for production."}, "workers": [{"credits_per_call": 12.748, "id": "worker_a", "p95_latency_ms": 81447, "pass_rate": 0.9147, "provider_group": "provider_y", "quality": 0.9107, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.746, "id": "worker_b", "p95_latency_ms": 72037, "pass_rate": 0.8222, "provider_group": "provider_z", "quality": 0.8735, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.982, "id": "worker_c", "p95_latency_ms": 20041, "pass_rate": 0.5938, "provider_group": "provider_x", "quality": 0.7711, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 0.995, "max_latency_ms": 37653, "minimum_quality": 0.6967}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000444", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.865, "estimated_latency_ms": 24887, "estimated_pass_rate": 0.3293, "estimated_quality": 0.5387, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 0.995, "max_latency_ms": 37653, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3293, "quality": 0.5387, "quality_target_met": false, "utility": 11.288172}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":0.995,\"max_latency_ms\":37653,\"minimum_quality\":0.6967},\"task\":{\"ambiguity\":0.4316,\"category\":\"review\",\"complexity\":0.8357,\"consequence\":0.7248,\"evidence_need\":0.6386,\"family\":\"policy_compliance\",\"input_tokens\":1480,\"reversibility\":0.6109,\"risk\":0.8232,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":3.786,\"id\":\"worker_a\",\"p95_latency_ms\":69523,\"pass_rate\":0.8427,\"provider_group\":\"provider_y\",\"quality\":0.8611,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.865,\"id\":\"worker_b\",\"p95_latency_ms\":24887,\"pass_rate\":0.5234,\"provider_group\":\"provider_z\",\"quality\":0.6981,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.977,\"id\":\"worker_c\",\"p95_latency_ms\":60398,\"pass_rate\":0.9135,\"provider_group\":\"provider_x\",\"quality\":0.9796,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.865,\"estimated_latency_ms\":24887,\"estimated_pass_rate\":0.3293,\"estimated_quality\":0.5387,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":0.995,\"max_latency_ms\":37653,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4316, "category": "review", "complexity": 0.8357, "consequence": 0.7248, "evidence_need": 0.6386, "family": "policy_compliance", "input_tokens": 1480, "request_id": "synthetic-policy_compliance-000444", "reversibility": 0.6109, "risk": 0.8232, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 3.786, "id": "worker_a", "p95_latency_ms": 69523, "pass_rate": 0.8427, "provider_group": "provider_y", "quality": 0.8611, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.865, "id": "worker_b", "p95_latency_ms": 24887, "pass_rate": 0.5234, "provider_group": "provider_z", "quality": 0.6981, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.977, "id": "worker_c", "p95_latency_ms": 60398, "pass_rate": 0.9135, "provider_group": "provider_x", "quality": 0.9796, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.76, "max_latency_ms": 136235, "minimum_quality": 0.7945}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000801", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.536, "estimated_latency_ms": 125746, "estimated_pass_rate": 0.9085, "estimated_quality": 0.9518, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.76, "max_latency_ms": 136235, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9085, "quality": 0.9518, "quality_target_met": true, "utility": 73.518334}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.76,\"max_latency_ms\":136235,\"minimum_quality\":0.7945},\"task\":{\"ambiguity\":0.5575,\"category\":\"review\",\"complexity\":0.4399,\"consequence\":0.6284,\"evidence_need\":0.7942,\"family\":\"policy_compliance\",\"input_tokens\":5440,\"reversibility\":0.6592,\"risk\":0.7567,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.987,\"id\":\"worker_a\",\"p95_latency_ms\":27927,\"pass_rate\":0.5909,\"provider_group\":\"provider_y\",\"quality\":0.7808,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.566,\"id\":\"worker_b\",\"p95_latency_ms\":55836,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9933,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.97,\"id\":\"worker_c\",\"p95_latency_ms\":69910,\"pass_rate\":0.7681,\"provider_group\":\"provider_x\",\"quality\":0.9056,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.536,\"estimated_latency_ms\":125746,\"estimated_pass_rate\":0.9085,\"estimated_quality\":0.9518,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.76,\"max_latency_ms\":136235,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5575, "category": "review", "complexity": 0.4399, "consequence": 0.6284, "evidence_need": 0.7942, "family": "policy_compliance", "input_tokens": 5440, "request_id": "synthetic-policy_compliance-000801", "reversibility": 0.6592, "risk": 0.7567, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.987, "id": "worker_a", "p95_latency_ms": 27927, "pass_rate": 0.5909, "provider_group": "provider_y", "quality": 0.7808, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.566, "id": "worker_b", "p95_latency_ms": 55836, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9933, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.97, "id": "worker_c", "p95_latency_ms": 69910, "pass_rate": 0.7681, "provider_group": "provider_x", "quality": 0.9056, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.339, "max_latency_ms": 45487, "minimum_quality": 0.5685}, "family": "migration_planning", "id": "synthetic-migration_planning-000070", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.053, "estimated_latency_ms": 28871, "estimated_pass_rate": 0.4179, "estimated_quality": 0.6407, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.339, "max_latency_ms": 45487, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4179, "quality": 0.6407, "quality_target_met": true, "utility": 43.327564}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.339,\"max_latency_ms\":45487,\"minimum_quality\":0.5685},\"task\":{\"ambiguity\":0.4341,\"category\":\"coding\",\"complexity\":0.6783,\"consequence\":0.662,\"evidence_need\":0.847,\"family\":\"migration_planning\",\"input_tokens\":400,\"reversibility\":0.557,\"risk\":0.6526,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":12.036,\"id\":\"worker_a\",\"p95_latency_ms\":54796,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9554,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.053,\"id\":\"worker_b\",\"p95_latency_ms\":28871,\"pass_rate\":0.5805,\"provider_group\":\"provider_z\",\"quality\":0.7595,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.931,\"id\":\"worker_c\",\"p95_latency_ms\":68541,\"pass_rate\":0.8145,\"provider_group\":\"provider_x\",\"quality\":0.912,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.053,\"estimated_latency_ms\":28871,\"estimated_pass_rate\":0.4179,\"estimated_quality\":0.6407,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.339,\"max_latency_ms\":45487,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4341, "category": "coding", "complexity": 0.6783, "consequence": 0.662, "evidence_need": 0.847, "family": "migration_planning", "input_tokens": 400, "request_id": "synthetic-migration_planning-000070", "reversibility": 0.557, "risk": 0.6526, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 12.036, "id": "worker_a", "p95_latency_ms": 54796, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9554, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.053, "id": "worker_b", "p95_latency_ms": 28871, "pass_rate": 0.5805, "provider_group": "provider_z", "quality": 0.7595, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.931, "id": "worker_c", "p95_latency_ms": 68541, "pass_rate": 0.8145, "provider_group": "provider_x", "quality": 0.912, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.61, "max_latency_ms": 219328, "minimum_quality": 0.8654}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000315", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.608, "estimated_latency_ms": 138709, "estimated_pass_rate": 0.8235, "estimated_quality": 0.9187, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.61, "max_latency_ms": 219328, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8235, "quality": 0.9187, "quality_target_met": true, "utility": 69.285563}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.61,\"max_latency_ms\":219328,\"minimum_quality\":0.8654},\"task\":{\"ambiguity\":0.3637,\"category\":\"review\",\"complexity\":0.8276,\"consequence\":0.5171,\"evidence_need\":0.8944,\"family\":\"policy_compliance\",\"input_tokens\":3560,\"reversibility\":0.4852,\"risk\":0.8004,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":11.429,\"id\":\"worker_a\",\"p95_latency_ms\":86107,\"pass_rate\":0.8965,\"provider_group\":\"provider_x\",\"quality\":0.9927,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.179,\"id\":\"worker_b\",\"p95_latency_ms\":52602,\"pass_rate\":0.7217,\"provider_group\":\"provider_y\",\"quality\":0.8669,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.953,\"id\":\"worker_c\",\"p95_latency_ms\":25696,\"pass_rate\":0.5435,\"provider_group\":\"provider_z\",\"quality\":0.7474,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.608,\"estimated_latency_ms\":138709,\"estimated_pass_rate\":0.8235,\"estimated_quality\":0.9187,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.61,\"max_latency_ms\":219328,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3637, "category": "review", "complexity": 0.8276, "consequence": 0.5171, "evidence_need": 0.8944, "family": "policy_compliance", "input_tokens": 3560, "request_id": "synthetic-policy_compliance-000315", "reversibility": 0.4852, "risk": 0.8004, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 11.429, "id": "worker_a", "p95_latency_ms": 86107, "pass_rate": 0.8965, "provider_group": "provider_x", "quality": 0.9927, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.179, "id": "worker_b", "p95_latency_ms": 52602, "pass_rate": 0.7217, "provider_group": "provider_y", "quality": 0.8669, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.953, "id": "worker_c", "p95_latency_ms": 25696, "pass_rate": 0.5435, "provider_group": "provider_z", "quality": 0.7474, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.284, "max_latency_ms": 35404, "minimum_quality": 0.7055}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000861", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.108, "estimated_latency_ms": 23712, "estimated_pass_rate": 0.42, "estimated_quality": 0.5871, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.284, "max_latency_ms": 35404, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.42, "quality": 0.5871, "quality_target_met": false, "utility": 18.514052}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.284,\"max_latency_ms\":35404,\"minimum_quality\":0.7055},\"task\":{\"ambiguity\":0.7102,\"category\":\"review\",\"complexity\":0.6359,\"consequence\":0.7907,\"evidence_need\":0.8291,\"family\":\"policy_compliance\",\"input_tokens\":3760,\"reversibility\":0.3639,\"risk\":0.8415,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.375,\"id\":\"worker_a\",\"p95_latency_ms\":72084,\"pass_rate\":0.7959,\"provider_group\":\"provider_z\",\"quality\":0.8385,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.41,\"id\":\"worker_b\",\"p95_latency_ms\":59990,\"pass_rate\":0.9337,\"provider_group\":\"provider_x\",\"quality\":0.9321,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.108,\"id\":\"worker_c\",\"p95_latency_ms\":23712,\"pass_rate\":0.614,\"provider_group\":\"provider_y\",\"quality\":0.7304,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.108,\"estimated_latency_ms\":23712,\"estimated_pass_rate\":0.42,\"estimated_quality\":0.5871,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.284,\"max_latency_ms\":35404,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7102, "category": "review", "complexity": 0.6359, "consequence": 0.7907, "evidence_need": 0.8291, "family": "policy_compliance", "input_tokens": 3760, "request_id": "synthetic-policy_compliance-000861", "reversibility": 0.3639, "risk": 0.8415, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.375, "id": "worker_a", "p95_latency_ms": 72084, "pass_rate": 0.7959, "provider_group": "provider_z", "quality": 0.8385, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.41, "id": "worker_b", "p95_latency_ms": 59990, "pass_rate": 0.9337, "provider_group": "provider_x", "quality": 0.9321, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.108, "id": "worker_c", "p95_latency_ms": 23712, "pass_rate": 0.614, "provider_group": "provider_y", "quality": 0.7304, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.532, "max_latency_ms": 208657, "minimum_quality": 0.8796}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001089", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.848, "estimated_latency_ms": 157947, "estimated_pass_rate": 0.8409, "estimated_quality": 0.8872, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.532, "max_latency_ms": 208657, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8409, "quality": 0.8872, "quality_target_met": true, "utility": 69.242427}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.532,\"max_latency_ms\":208657,\"minimum_quality\":0.8796},\"task\":{\"ambiguity\":0.6318,\"category\":\"review\",\"complexity\":0.5521,\"consequence\":0.8691,\"evidence_need\":0.5579,\"family\":\"policy_compliance\",\"input_tokens\":4080,\"reversibility\":0.6945,\"risk\":0.623,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":4.161,\"id\":\"worker_a\",\"p95_latency_ms\":72985,\"pass_rate\":0.7352,\"provider_group\":\"provider_y\",\"quality\":0.8521,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.687,\"id\":\"worker_b\",\"p95_latency_ms\":84962,\"pass_rate\":0.916,\"provider_group\":\"provider_z\",\"quality\":0.9236,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.995,\"id\":\"worker_c\",\"p95_latency_ms\":27585,\"pass_rate\":0.4679,\"provider_group\":\"provider_x\",\"quality\":0.7118,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.848,\"estimated_latency_ms\":157947,\"estimated_pass_rate\":0.8409,\"estimated_quality\":0.8872,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.532,\"max_latency_ms\":208657,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6318, "category": "review", "complexity": 0.5521, "consequence": 0.8691, "evidence_need": 0.5579, "family": "policy_compliance", "input_tokens": 4080, "request_id": "synthetic-policy_compliance-001089", "reversibility": 0.6945, "risk": 0.623, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 4.161, "id": "worker_a", "p95_latency_ms": 72985, "pass_rate": 0.7352, "provider_group": "provider_y", "quality": 0.8521, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 11.687, "id": "worker_b", "p95_latency_ms": 84962, "pass_rate": 0.916, "provider_group": "provider_z", "quality": 0.9236, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.995, "id": "worker_c", "p95_latency_ms": 27585, "pass_rate": 0.4679, "provider_group": "provider_x", "quality": 0.7118, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.721, "max_latency_ms": 46880, "minimum_quality": 0.6447}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000473", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.92, "estimated_latency_ms": 26797, "estimated_pass_rate": 0.2349, "estimated_quality": 0.5101, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.721, "max_latency_ms": 46880, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2349, "quality": 0.5101, "quality_target_met": false, "utility": 27.072745}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.721,\"max_latency_ms\":46880,\"minimum_quality\":0.6447},\"task\":{\"ambiguity\":0.6432,\"category\":\"analysis\",\"complexity\":0.8092,\"consequence\":0.3004,\"evidence_need\":0.734,\"family\":\"performance_tuning\",\"input_tokens\":6640,\"reversibility\":0.8798,\"risk\":0.2907,\"summary\":\"Review a proposed optimization for rate limiting for hidden regressions.\"},\"workers\":[{\"credits_per_call\":2.75,\"id\":\"worker_a\",\"p95_latency_ms\":49694,\"pass_rate\":0.9245,\"provider_group\":\"provider_z\",\"quality\":0.9257,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.885,\"id\":\"worker_b\",\"p95_latency_ms\":79593,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.92,\"id\":\"worker_c\",\"p95_latency_ms\":26797,\"pass_rate\":0.3988,\"provider_group\":\"provider_x\",\"quality\":0.6406,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.92,\"estimated_latency_ms\":26797,\"estimated_pass_rate\":0.2349,\"estimated_quality\":0.5101,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.721,\"max_latency_ms\":46880,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6432, "category": "analysis", "complexity": 0.8092, "consequence": 0.3004, "evidence_need": 0.734, "family": "performance_tuning", "input_tokens": 6640, "request_id": "synthetic-performance_tuning-000473", "reversibility": 0.8798, "risk": 0.2907, "summary": "Review a proposed optimization for rate limiting for hidden regressions."}, "workers": [{"credits_per_call": 2.75, "id": "worker_a", "p95_latency_ms": 49694, "pass_rate": 0.9245, "provider_group": "provider_z", "quality": 0.9257, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.885, "id": "worker_b", "p95_latency_ms": 79593, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.92, "id": "worker_c", "p95_latency_ms": 26797, "pass_rate": 0.3988, "provider_group": "provider_x", "quality": 0.6406, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.995, "max_latency_ms": 164747, "minimum_quality": 0.8346}, "family": "migration_planning", "id": "synthetic-migration_planning-000955", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.808, "estimated_latency_ms": 150227, "estimated_pass_rate": 0.7512, "estimated_quality": 0.8925, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.995, "max_latency_ms": 164747, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7512, "quality": 0.8925, "quality_target_met": true, "utility": 72.751488}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.995,\"max_latency_ms\":164747,\"minimum_quality\":0.8346},\"task\":{\"ambiguity\":0.7288,\"category\":\"coding\",\"complexity\":0.7595,\"consequence\":0.4872,\"evidence_need\":0.808,\"family\":\"migration_planning\",\"input_tokens\":6200,\"reversibility\":0.4101,\"risk\":0.4048,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":14.252,\"id\":\"worker_a\",\"p95_latency_ms\":52301,\"pass_rate\":0.979,\"provider_group\":\"provider_y\",\"quality\":0.9133,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.197,\"id\":\"worker_b\",\"p95_latency_ms\":17797,\"pass_rate\":0.5364,\"provider_group\":\"provider_x\",\"quality\":0.7393,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.359,\"id\":\"worker_c\",\"p95_latency_ms\":80129,\"pass_rate\":0.8676,\"provider_group\":\"provider_z\",\"quality\":0.9065,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.808,\"estimated_latency_ms\":150227,\"estimated_pass_rate\":0.7512,\"estimated_quality\":0.8925,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.995,\"max_latency_ms\":164747,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7288, "category": "coding", "complexity": 0.7595, "consequence": 0.4872, "evidence_need": 0.808, "family": "migration_planning", "input_tokens": 6200, "request_id": "synthetic-migration_planning-000955", "reversibility": 0.4101, "risk": 0.4048, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 14.252, "id": "worker_a", "p95_latency_ms": 52301, "pass_rate": 0.979, "provider_group": "provider_y", "quality": 0.9133, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.197, "id": "worker_b", "p95_latency_ms": 17797, "pass_rate": 0.5364, "provider_group": "provider_x", "quality": 0.7393, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.359, "id": "worker_c", "p95_latency_ms": 80129, "pass_rate": 0.8676, "provider_group": "provider_z", "quality": 0.9065, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.859, "max_latency_ms": 24803, "minimum_quality": 0.5699}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000765", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.126, "estimated_latency_ms": 22175, "estimated_pass_rate": 0.3423, "estimated_quality": 0.6747, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.859, "max_latency_ms": 24803, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3423, "quality": 0.6747, "quality_target_met": true, "utility": 43.638599}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.859,\"max_latency_ms\":24803,\"minimum_quality\":0.5699},\"task\":{\"ambiguity\":0.4927,\"category\":\"review\",\"complexity\":0.5703,\"consequence\":0.9163,\"evidence_need\":0.7102,\"family\":\"policy_compliance\",\"input_tokens\":3960,\"reversibility\":0.5659,\"risk\":0.4711,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.126,\"id\":\"worker_a\",\"p95_latency_ms\":22175,\"pass_rate\":0.4824,\"provider_group\":\"provider_x\",\"quality\":0.7648,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.376,\"id\":\"worker_b\",\"p95_latency_ms\":47321,\"pass_rate\":0.8606,\"provider_group\":\"provider_z\",\"quality\":0.8536,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.902,\"id\":\"worker_c\",\"p95_latency_ms\":53898,\"pass_rate\":0.9731,\"provider_group\":\"provider_y\",\"quality\":0.9585,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.126,\"estimated_latency_ms\":22175,\"estimated_pass_rate\":0.3423,\"estimated_quality\":0.6747,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.859,\"max_latency_ms\":24803,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4927, "category": "review", "complexity": 0.5703, "consequence": 0.9163, "evidence_need": 0.7102, "family": "policy_compliance", "input_tokens": 3960, "request_id": "synthetic-policy_compliance-000765", "reversibility": 0.5659, "risk": 0.4711, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 1.126, "id": "worker_a", "p95_latency_ms": 22175, "pass_rate": 0.4824, "provider_group": "provider_x", "quality": 0.7648, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.376, "id": "worker_b", "p95_latency_ms": 47321, "pass_rate": 0.8606, "provider_group": "provider_z", "quality": 0.8536, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.902, "id": "worker_c", "p95_latency_ms": 53898, "pass_rate": 0.9731, "provider_group": "provider_y", "quality": 0.9585, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.422, "max_latency_ms": 34317, "minimum_quality": 0.521}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000302", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.918, "estimated_latency_ms": 31423, "estimated_pass_rate": 0.3394, "estimated_quality": 0.595, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.422, "max_latency_ms": 34317, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3394, "quality": 0.595, "quality_target_met": true, "utility": 38.697842}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.422,\"max_latency_ms\":34317,\"minimum_quality\":0.521},\"task\":{\"ambiguity\":0.7245,\"category\":\"analysis\",\"complexity\":0.7398,\"consequence\":0.6362,\"evidence_need\":0.6048,\"family\":\"performance_tuning\",\"input_tokens\":6320,\"reversibility\":0.6025,\"risk\":0.3718,\"summary\":\"Diagnose latency variance in deployment rollback and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.918,\"id\":\"worker_a\",\"p95_latency_ms\":31423,\"pass_rate\":0.5082,\"provider_group\":\"provider_z\",\"quality\":0.7242,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.439,\"id\":\"worker_b\",\"p95_latency_ms\":91600,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.215,\"id\":\"worker_c\",\"p95_latency_ms\":74836,\"pass_rate\":0.8663,\"provider_group\":\"provider_y\",\"quality\":0.9412,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.918,\"estimated_latency_ms\":31423,\"estimated_pass_rate\":0.3394,\"estimated_quality\":0.595,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.422,\"max_latency_ms\":34317,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7245, "category": "analysis", "complexity": 0.7398, "consequence": 0.6362, "evidence_need": 0.6048, "family": "performance_tuning", "input_tokens": 6320, "request_id": "synthetic-performance_tuning-000302", "reversibility": 0.6025, "risk": 0.3718, "summary": "Diagnose latency variance in deployment rollback and rank the next experiments."}, "workers": [{"credits_per_call": 0.918, "id": "worker_a", "p95_latency_ms": 31423, "pass_rate": 0.5082, "provider_group": "provider_z", "quality": 0.7242, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.439, "id": "worker_b", "p95_latency_ms": 91600, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.215, "id": "worker_c", "p95_latency_ms": 74836, "pass_rate": 0.8663, "provider_group": "provider_y", "quality": 0.9412, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.087, "max_latency_ms": 198552, "minimum_quality": 0.8074}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001103", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.656, "estimated_latency_ms": 167739, "estimated_pass_rate": 0.7125, "estimated_quality": 0.9144, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.087, "max_latency_ms": 198552, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7125, "quality": 0.9144, "quality_target_met": true, "utility": 71.725698}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.087,\"max_latency_ms\":198552,\"minimum_quality\":0.8074},\"task\":{\"ambiguity\":0.5393,\"category\":\"analysis\",\"complexity\":0.7992,\"consequence\":0.5809,\"evidence_need\":0.5365,\"family\":\"performance_tuning\",\"input_tokens\":6040,\"reversibility\":0.6384,\"risk\":0.4951,\"summary\":\"Choose between throughput and tail-latency changes for schema evolution.\"},\"workers\":[{\"credits_per_call\":12.549,\"id\":\"worker_a\",\"p95_latency_ms\":81242,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.155,\"id\":\"worker_b\",\"p95_latency_ms\":20838,\"pass_rate\":0.4297,\"provider_group\":\"provider_z\",\"quality\":0.7291,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.952,\"id\":\"worker_c\",\"p95_latency_ms\":65659,\"pass_rate\":0.8312,\"provider_group\":\"provider_x\",\"quality\":0.864,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.656,\"estimated_latency_ms\":167739,\"estimated_pass_rate\":0.7125,\"estimated_quality\":0.9144,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.087,\"max_latency_ms\":198552,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5393, "category": "analysis", "complexity": 0.7992, "consequence": 0.5809, "evidence_need": 0.5365, "family": "performance_tuning", "input_tokens": 6040, "request_id": "synthetic-performance_tuning-001103", "reversibility": 0.6384, "risk": 0.4951, "summary": "Choose between throughput and tail-latency changes for schema evolution."}, "workers": [{"credits_per_call": 12.549, "id": "worker_a", "p95_latency_ms": 81242, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.155, "id": "worker_b", "p95_latency_ms": 20838, "pass_rate": 0.4297, "provider_group": "provider_z", "quality": 0.7291, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.952, "id": "worker_c", "p95_latency_ms": 65659, "pass_rate": 0.8312, "provider_group": "provider_x", "quality": 0.864, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.394, "max_latency_ms": 41012, "minimum_quality": 0.7331}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000243", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.1, "estimated_latency_ms": 29365, "estimated_pass_rate": 0.4668, "estimated_quality": 0.6745, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.394, "max_latency_ms": 41012, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4668, "quality": 0.6745, "quality_target_met": false, "utility": 43.655781}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.394,\"max_latency_ms\":41012,\"minimum_quality\":0.7331},\"task\":{\"ambiguity\":0.4691,\"category\":\"review\",\"complexity\":0.6034,\"consequence\":0.9161,\"evidence_need\":0.5513,\"family\":\"policy_compliance\",\"input_tokens\":1480,\"reversibility\":0.3789,\"risk\":0.5434,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":1.1,\"id\":\"worker_a\",\"p95_latency_ms\":29365,\"pass_rate\":0.6147,\"provider_group\":\"provider_y\",\"quality\":0.7742,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.786,\"id\":\"worker_b\",\"p95_latency_ms\":49218,\"pass_rate\":0.8403,\"provider_group\":\"provider_z\",\"quality\":0.849,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.749,\"id\":\"worker_c\",\"p95_latency_ms\":88372,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9339,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.1,\"estimated_latency_ms\":29365,\"estimated_pass_rate\":0.4668,\"estimated_quality\":0.6745,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.394,\"max_latency_ms\":41012,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4691, "category": "review", "complexity": 0.6034, "consequence": 0.9161, "evidence_need": 0.5513, "family": "policy_compliance", "input_tokens": 1480, "request_id": "synthetic-policy_compliance-000243", "reversibility": 0.3789, "risk": 0.5434, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 1.1, "id": "worker_a", "p95_latency_ms": 29365, "pass_rate": 0.6147, "provider_group": "provider_y", "quality": 0.7742, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.786, "id": "worker_b", "p95_latency_ms": 49218, "pass_rate": 0.8403, "provider_group": "provider_z", "quality": 0.849, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.749, "id": "worker_c", "p95_latency_ms": 88372, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9339, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.601, "max_latency_ms": 184338, "minimum_quality": 0.8813}, "family": "migration_planning", "id": "synthetic-migration_planning-001252", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.759, "estimated_latency_ms": 147853, "estimated_pass_rate": 0.758, "estimated_quality": 0.9141, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.601, "max_latency_ms": 184338, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.758, "quality": 0.9141, "quality_target_met": true, "utility": 75.285792}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.601,\"max_latency_ms\":184338,\"minimum_quality\":0.8813},\"task\":{\"ambiguity\":0.5241,\"category\":\"coding\",\"complexity\":0.6844,\"consequence\":0.8471,\"evidence_need\":0.7713,\"family\":\"migration_planning\",\"input_tokens\":2600,\"reversibility\":0.3677,\"risk\":0.6188,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.072,\"id\":\"worker_a\",\"p95_latency_ms\":62580,\"pass_rate\":0.8152,\"provider_group\":\"provider_x\",\"quality\":0.8291,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.85,\"id\":\"worker_b\",\"p95_latency_ms\":32312,\"pass_rate\":0.5646,\"provider_group\":\"provider_z\",\"quality\":0.7723,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.837,\"id\":\"worker_c\",\"p95_latency_ms\":52961,\"pass_rate\":0.9686,\"provider_group\":\"provider_y\",\"quality\":0.9693,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.759,\"estimated_latency_ms\":147853,\"estimated_pass_rate\":0.758,\"estimated_quality\":0.9141,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.601,\"max_latency_ms\":184338,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5241, "category": "coding", "complexity": 0.6844, "consequence": 0.8471, "evidence_need": 0.7713, "family": "migration_planning", "input_tokens": 2600, "request_id": "synthetic-migration_planning-001252", "reversibility": 0.3677, "risk": 0.6188, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 3.072, "id": "worker_a", "p95_latency_ms": 62580, "pass_rate": 0.8152, "provider_group": "provider_x", "quality": 0.8291, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.85, "id": "worker_b", "p95_latency_ms": 32312, "pass_rate": 0.5646, "provider_group": "provider_z", "quality": 0.7723, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.837, "id": "worker_c", "p95_latency_ms": 52961, "pass_rate": 0.9686, "provider_group": "provider_y", "quality": 0.9693, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.251, "max_latency_ms": 35976, "minimum_quality": 0.5919}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000807", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.024, "estimated_latency_ms": 18628, "estimated_pass_rate": 0.4427, "estimated_quality": 0.6268, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.251, "max_latency_ms": 35976, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4427, "quality": 0.6268, "quality_target_met": true, "utility": 40.98383}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.251,\"max_latency_ms\":35976,\"minimum_quality\":0.5919},\"task\":{\"ambiguity\":0.5175,\"category\":\"review\",\"complexity\":0.5254,\"consequence\":0.7435,\"evidence_need\":0.8117,\"family\":\"policy_compliance\",\"input_tokens\":3840,\"reversibility\":0.5859,\"risk\":0.5627,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.024,\"id\":\"worker_a\",\"p95_latency_ms\":18628,\"pass_rate\":0.587,\"provider_group\":\"provider_x\",\"quality\":0.7171,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.952,\"id\":\"worker_b\",\"p95_latency_ms\":79384,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.861,\"id\":\"worker_c\",\"p95_latency_ms\":49183,\"pass_rate\":0.723,\"provider_group\":\"provider_y\",\"quality\":0.8692,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.024,\"estimated_latency_ms\":18628,\"estimated_pass_rate\":0.4427,\"estimated_quality\":0.6268,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.251,\"max_latency_ms\":35976,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5175, "category": "review", "complexity": 0.5254, "consequence": 0.7435, "evidence_need": 0.8117, "family": "policy_compliance", "input_tokens": 3840, "request_id": "synthetic-policy_compliance-000807", "reversibility": 0.5859, "risk": 0.5627, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.024, "id": "worker_a", "p95_latency_ms": 18628, "pass_rate": 0.587, "provider_group": "provider_x", "quality": 0.7171, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.952, "id": "worker_b", "p95_latency_ms": 79384, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.861, "id": "worker_c", "p95_latency_ms": 49183, "pass_rate": 0.723, "provider_group": "provider_y", "quality": 0.8692, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.064, "max_latency_ms": 239032, "minimum_quality": 0.8653}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000042", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.058, "estimated_latency_ms": 170550, "estimated_pass_rate": 0.784, "estimated_quality": 0.9314, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.064, "max_latency_ms": 239032, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.784, "quality": 0.9314, "quality_target_met": true, "utility": 77.692564}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.064,\"max_latency_ms\":239032,\"minimum_quality\":0.8653},\"task\":{\"ambiguity\":0.78,\"category\":\"review\",\"complexity\":0.4322,\"consequence\":0.6133,\"evidence_need\":0.9158,\"family\":\"policy_compliance\",\"input_tokens\":7240,\"reversibility\":0.66,\"risk\":0.7111,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":3.768,\"id\":\"worker_a\",\"p95_latency_ms\":63782,\"pass_rate\":0.8425,\"provider_group\":\"provider_y\",\"quality\":0.8938,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.094,\"id\":\"worker_b\",\"p95_latency_ms\":77259,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.196,\"id\":\"worker_c\",\"p95_latency_ms\":29509,\"pass_rate\":0.5348,\"provider_group\":\"provider_x\",\"quality\":0.6916,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.058,\"estimated_latency_ms\":170550,\"estimated_pass_rate\":0.784,\"estimated_quality\":0.9314,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.064,\"max_latency_ms\":239032,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.78, "category": "review", "complexity": 0.4322, "consequence": 0.6133, "evidence_need": 0.9158, "family": "policy_compliance", "input_tokens": 7240, "request_id": "synthetic-policy_compliance-000042", "reversibility": 0.66, "risk": 0.7111, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 3.768, "id": "worker_a", "p95_latency_ms": 63782, "pass_rate": 0.8425, "provider_group": "provider_y", "quality": 0.8938, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.094, "id": "worker_b", "p95_latency_ms": 77259, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.196, "id": "worker_c", "p95_latency_ms": 29509, "pass_rate": 0.5348, "provider_group": "provider_x", "quality": 0.6916, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.874, "max_latency_ms": 263814, "minimum_quality": 0.7438}, "family": "migration_planning", "id": "synthetic-migration_planning-000313", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.202, "estimated_latency_ms": 103576, "estimated_pass_rate": 0.6345, "estimated_quality": 0.793, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.874, "max_latency_ms": 263814, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6345, "quality": 0.793, "quality_target_met": true, "utility": 60.704712}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.874,\"max_latency_ms\":263814,\"minimum_quality\":0.7438},\"task\":{\"ambiguity\":0.3825,\"category\":\"coding\",\"complexity\":0.8243,\"consequence\":0.7165,\"evidence_need\":0.5409,\"family\":\"migration_planning\",\"input_tokens\":920,\"reversibility\":0.4916,\"risk\":0.6326,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.968,\"id\":\"worker_a\",\"p95_latency_ms\":25375,\"pass_rate\":0.5669,\"provider_group\":\"provider_y\",\"quality\":0.7685,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.234,\"id\":\"worker_b\",\"p95_latency_ms\":78201,\"pass_rate\":0.8557,\"provider_group\":\"provider_z\",\"quality\":0.8098,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.545,\"id\":\"worker_c\",\"p95_latency_ms\":68456,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.202,\"estimated_latency_ms\":103576,\"estimated_pass_rate\":0.6345,\"estimated_quality\":0.793,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.874,\"max_latency_ms\":263814,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3825, "category": "coding", "complexity": 0.8243, "consequence": 0.7165, "evidence_need": 0.5409, "family": "migration_planning", "input_tokens": 920, "request_id": "synthetic-migration_planning-000313", "reversibility": 0.4916, "risk": 0.6326, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 0.968, "id": "worker_a", "p95_latency_ms": 25375, "pass_rate": 0.5669, "provider_group": "provider_y", "quality": 0.7685, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.234, "id": "worker_b", "p95_latency_ms": 78201, "pass_rate": 0.8557, "provider_group": "provider_z", "quality": 0.8098, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.545, "id": "worker_c", "p95_latency_ms": 68456, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.378, "max_latency_ms": 194051, "minimum_quality": 0.8611}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001167", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.651, "estimated_latency_ms": 177968, "estimated_pass_rate": 0.7768, "estimated_quality": 0.9052, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 26.378, "max_latency_ms": 194051, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7768, "quality": 0.9052, "quality_target_met": true, "utility": 77.953013}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.378,\"max_latency_ms\":194051,\"minimum_quality\":0.8611},\"task\":{\"ambiguity\":0.6637,\"category\":\"review\",\"complexity\":0.6643,\"consequence\":0.6719,\"evidence_need\":0.8266,\"family\":\"policy_compliance\",\"input_tokens\":4880,\"reversibility\":0.6613,\"risk\":0.7546,\"summary\":\"Review whether private credentials satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.192,\"id\":\"worker_a\",\"p95_latency_ms\":27006,\"pass_rate\":0.5476,\"provider_group\":\"provider_z\",\"quality\":0.7951,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.523,\"id\":\"worker_b\",\"p95_latency_ms\":83224,\"pass_rate\":0.9937,\"provider_group\":\"provider_x\",\"quality\":0.9319,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.936,\"id\":\"worker_c\",\"p95_latency_ms\":67738,\"pass_rate\":0.8237,\"provider_group\":\"provider_y\",\"quality\":0.8731,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.651,\"estimated_latency_ms\":177968,\"estimated_pass_rate\":0.7768,\"estimated_quality\":0.9052,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.378,\"max_latency_ms\":194051,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6637, "category": "review", "complexity": 0.6643, "consequence": 0.6719, "evidence_need": 0.8266, "family": "policy_compliance", "input_tokens": 4880, "request_id": "synthetic-policy_compliance-001167", "reversibility": 0.6613, "risk": 0.7546, "summary": "Review whether private credentials satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.192, "id": "worker_a", "p95_latency_ms": 27006, "pass_rate": 0.5476, "provider_group": "provider_z", "quality": 0.7951, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.523, "id": "worker_b", "p95_latency_ms": 83224, "pass_rate": 0.9937, "provider_group": "provider_x", "quality": 0.9319, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.936, "id": "worker_c", "p95_latency_ms": 67738, "pass_rate": 0.8237, "provider_group": "provider_y", "quality": 0.8731, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 14.956, "max_latency_ms": 222119, "minimum_quality": 0.8053}, "family": "migration_planning", "id": "synthetic-migration_planning-000508", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.965, "estimated_latency_ms": 157355, "estimated_pass_rate": 0.7748, "estimated_quality": 0.9329, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 14.956, "max_latency_ms": 222119, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7748, "quality": 0.9329, "quality_target_met": true, "utility": 74.814616}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":14.956,\"max_latency_ms\":222119,\"minimum_quality\":0.8053},\"task\":{\"ambiguity\":0.4254,\"category\":\"coding\",\"complexity\":0.6565,\"consequence\":0.4929,\"evidence_need\":0.7634,\"family\":\"migration_planning\",\"input_tokens\":4560,\"reversibility\":0.3233,\"risk\":0.6978,\"summary\":\"Design compatibility testing for a observability schema migration.\"},\"workers\":[{\"credits_per_call\":2.9,\"id\":\"worker_a\",\"p95_latency_ms\":78618,\"pass_rate\":0.8567,\"provider_group\":\"provider_y\",\"quality\":0.8969,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.908,\"id\":\"worker_b\",\"p95_latency_ms\":22434,\"pass_rate\":0.6025,\"provider_group\":\"provider_x\",\"quality\":0.7708,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.157,\"id\":\"worker_c\",\"p95_latency_ms\":56303,\"pass_rate\":0.903,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.965,\"estimated_latency_ms\":157355,\"estimated_pass_rate\":0.7748,\"estimated_quality\":0.9329,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":14.956,\"max_latency_ms\":222119,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4254, "category": "coding", "complexity": 0.6565, "consequence": 0.4929, "evidence_need": 0.7634, "family": "migration_planning", "input_tokens": 4560, "request_id": "synthetic-migration_planning-000508", "reversibility": 0.3233, "risk": 0.6978, "summary": "Design compatibility testing for a observability schema migration."}, "workers": [{"credits_per_call": 2.9, "id": "worker_a", "p95_latency_ms": 78618, "pass_rate": 0.8567, "provider_group": "provider_y", "quality": 0.8969, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.908, "id": "worker_b", "p95_latency_ms": 22434, "pass_rate": 0.6025, "provider_group": "provider_x", "quality": 0.7708, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.157, "id": "worker_c", "p95_latency_ms": 56303, "pass_rate": 0.903, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 14.681, "max_latency_ms": 89507, "minimum_quality": 0.8287}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001221", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.372, "estimated_latency_ms": 75049, "estimated_pass_rate": 0.8595, "estimated_quality": 0.8884, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.681, "max_latency_ms": 89507, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 5, "pass_rate": 0.8595, "quality": 0.8884, "quality_target_met": true, "utility": 64.409092}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":14.681,\"max_latency_ms\":89507,\"minimum_quality\":0.8287},\"task\":{\"ambiguity\":0.4201,\"category\":\"review\",\"complexity\":0.7425,\"consequence\":0.5657,\"evidence_need\":0.6543,\"family\":\"policy_compliance\",\"input_tokens\":3760,\"reversibility\":0.4146,\"risk\":0.5326,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.344,\"id\":\"worker_a\",\"p95_latency_ms\":79510,\"pass_rate\":0.8796,\"provider_group\":\"provider_z\",\"quality\":0.9209,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.328,\"id\":\"worker_b\",\"p95_latency_ms\":50037,\"pass_rate\":0.9767,\"provider_group\":\"provider_y\",\"quality\":0.9713,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.044,\"id\":\"worker_c\",\"p95_latency_ms\":25012,\"pass_rate\":0.6168,\"provider_group\":\"provider_x\",\"quality\":0.7582,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.372,\"estimated_latency_ms\":75049,\"estimated_pass_rate\":0.8595,\"estimated_quality\":0.8884,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.681,\"max_latency_ms\":89507,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4201, "category": "review", "complexity": 0.7425, "consequence": 0.5657, "evidence_need": 0.6543, "family": "policy_compliance", "input_tokens": 3760, "request_id": "synthetic-policy_compliance-001221", "reversibility": 0.4146, "risk": 0.5326, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.344, "id": "worker_a", "p95_latency_ms": 79510, "pass_rate": 0.8796, "provider_group": "provider_z", "quality": 0.9209, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.328, "id": "worker_b", "p95_latency_ms": 50037, "pass_rate": 0.9767, "provider_group": "provider_y", "quality": 0.9713, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.044, "id": "worker_c", "p95_latency_ms": 25012, "pass_rate": 0.6168, "provider_group": "provider_x", "quality": 0.7582, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.979, "max_latency_ms": 265983, "minimum_quality": 0.7966}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000768", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.333, "estimated_latency_ms": 93636, "estimated_pass_rate": 0.7819, "estimated_quality": 0.8674, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.979, "max_latency_ms": 265983, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7819, "quality": 0.8674, "quality_target_met": true, "utility": 67.351966}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.979,\"max_latency_ms\":265983,\"minimum_quality\":0.7966},\"task\":{\"ambiguity\":0.6975,\"category\":\"review\",\"complexity\":0.5315,\"consequence\":0.5961,\"evidence_need\":0.5968,\"family\":\"policy_compliance\",\"input_tokens\":3120,\"reversibility\":0.4832,\"risk\":0.571,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":11.135,\"id\":\"worker_a\",\"p95_latency_ms\":89900,\"pass_rate\":0.9053,\"provider_group\":\"provider_y\",\"quality\":0.9786,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.85,\"id\":\"worker_b\",\"p95_latency_ms\":19564,\"pass_rate\":0.6077,\"provider_group\":\"provider_x\",\"quality\":0.8021,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.483,\"id\":\"worker_c\",\"p95_latency_ms\":74072,\"pass_rate\":0.8655,\"provider_group\":\"provider_z\",\"quality\":0.9219,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.333,\"estimated_latency_ms\":93636,\"estimated_pass_rate\":0.7819,\"estimated_quality\":0.8674,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.979,\"max_latency_ms\":265983,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6975, "category": "review", "complexity": 0.5315, "consequence": 0.5961, "evidence_need": 0.5968, "family": "policy_compliance", "input_tokens": 3120, "request_id": "synthetic-policy_compliance-000768", "reversibility": 0.4832, "risk": 0.571, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 11.135, "id": "worker_a", "p95_latency_ms": 89900, "pass_rate": 0.9053, "provider_group": "provider_y", "quality": 0.9786, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.85, "id": "worker_b", "p95_latency_ms": 19564, "pass_rate": 0.6077, "provider_group": "provider_x", "quality": 0.8021, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.483, "id": "worker_c", "p95_latency_ms": 74072, "pass_rate": 0.8655, "provider_group": "provider_z", "quality": 0.9219, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.784, "max_latency_ms": 198363, "minimum_quality": 0.846}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000426", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.961, "estimated_latency_ms": 89635, "estimated_pass_rate": 0.7026, "estimated_quality": 0.8386, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.784, "max_latency_ms": 198363, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7026, "quality": 0.8386, "quality_target_met": false, "utility": 61.35994}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.784,\"max_latency_ms\":198363,\"minimum_quality\":0.846},\"task\":{\"ambiguity\":0.6609,\"category\":\"review\",\"complexity\":0.4005,\"consequence\":0.6896,\"evidence_need\":0.5936,\"family\":\"policy_compliance\",\"input_tokens\":2920,\"reversibility\":0.4636,\"risk\":0.6561,\"summary\":\"Review whether observability satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.833,\"id\":\"worker_a\",\"p95_latency_ms\":26280,\"pass_rate\":0.5854,\"provider_group\":\"provider_z\",\"quality\":0.6899,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.128,\"id\":\"worker_b\",\"p95_latency_ms\":63355,\"pass_rate\":0.7664,\"provider_group\":\"provider_x\",\"quality\":0.8812,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.002,\"id\":\"worker_c\",\"p95_latency_ms\":85263,\"pass_rate\":0.9751,\"provider_group\":\"provider_y\",\"quality\":0.9661,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.961,\"estimated_latency_ms\":89635,\"estimated_pass_rate\":0.7026,\"estimated_quality\":0.8386,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.784,\"max_latency_ms\":198363,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6609, "category": "review", "complexity": 0.4005, "consequence": 0.6896, "evidence_need": 0.5936, "family": "policy_compliance", "input_tokens": 2920, "request_id": "synthetic-policy_compliance-000426", "reversibility": 0.4636, "risk": 0.6561, "summary": "Review whether observability satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.833, "id": "worker_a", "p95_latency_ms": 26280, "pass_rate": 0.5854, "provider_group": "provider_z", "quality": 0.6899, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.128, "id": "worker_b", "p95_latency_ms": 63355, "pass_rate": 0.7664, "provider_group": "provider_x", "quality": 0.8812, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.002, "id": "worker_c", "p95_latency_ms": 85263, "pass_rate": 0.9751, "provider_group": "provider_y", "quality": 0.9661, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.603, "max_latency_ms": 189161, "minimum_quality": 0.8194}, "family": "migration_planning", "id": "synthetic-migration_planning-000946", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.208, "estimated_latency_ms": 168120, "estimated_pass_rate": 0.7374, "estimated_quality": 0.8731, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.603, "max_latency_ms": 189161, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7374, "quality": 0.8731, "quality_target_met": true, "utility": 69.882063}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.603,\"max_latency_ms\":189161,\"minimum_quality\":0.8194},\"task\":{\"ambiguity\":0.5326,\"category\":\"coding\",\"complexity\":0.9293,\"consequence\":0.6516,\"evidence_need\":0.4134,\"family\":\"migration_planning\",\"input_tokens\":5000,\"reversibility\":0.4807,\"risk\":0.4578,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.922,\"id\":\"worker_a\",\"p95_latency_ms\":22338,\"pass_rate\":0.5763,\"provider_group\":\"provider_z\",\"quality\":0.7459,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.174,\"id\":\"worker_b\",\"p95_latency_ms\":89516,\"pass_rate\":0.9759,\"provider_group\":\"provider_x\",\"quality\":0.9342,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.112,\"id\":\"worker_c\",\"p95_latency_ms\":56266,\"pass_rate\":0.7518,\"provider_group\":\"provider_y\",\"quality\":0.8449,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.208,\"estimated_latency_ms\":168120,\"estimated_pass_rate\":0.7374,\"estimated_quality\":0.8731,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.603,\"max_latency_ms\":189161,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5326, "category": "coding", "complexity": 0.9293, "consequence": 0.6516, "evidence_need": 0.4134, "family": "migration_planning", "input_tokens": 5000, "request_id": "synthetic-migration_planning-000946", "reversibility": 0.4807, "risk": 0.4578, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 0.922, "id": "worker_a", "p95_latency_ms": 22338, "pass_rate": 0.5763, "provider_group": "provider_z", "quality": 0.7459, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.174, "id": "worker_b", "p95_latency_ms": 89516, "pass_rate": 0.9759, "provider_group": "provider_x", "quality": 0.9342, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.112, "id": "worker_c", "p95_latency_ms": 56266, "pass_rate": 0.7518, "provider_group": "provider_y", "quality": 0.8449, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.809, "max_latency_ms": 196637, "minimum_quality": 0.9272}, "family": "migration_planning", "id": "synthetic-migration_planning-000094", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.057, "estimated_latency_ms": 135742, "estimated_pass_rate": 0.702, "estimated_quality": 0.9031, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 15.809, "max_latency_ms": 196637, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.702, "quality": 0.9031, "quality_target_met": false, "utility": 72.482851}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.809,\"max_latency_ms\":196637,\"minimum_quality\":0.9272},\"task\":{\"ambiguity\":0.7262,\"category\":\"coding\",\"complexity\":0.652,\"consequence\":0.6705,\"evidence_need\":0.836,\"family\":\"migration_planning\",\"input_tokens\":5120,\"reversibility\":0.3616,\"risk\":0.5648,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":0.782,\"id\":\"worker_a\",\"p95_latency_ms\":25262,\"pass_rate\":0.4955,\"provider_group\":\"provider_y\",\"quality\":0.778,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.595,\"id\":\"worker_b\",\"p95_latency_ms\":59792,\"pass_rate\":0.7599,\"provider_group\":\"provider_x\",\"quality\":0.869,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.68,\"id\":\"worker_c\",\"p95_latency_ms\":50688,\"pass_rate\":0.9211,\"provider_group\":\"provider_z\",\"quality\":0.9483,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.057,\"estimated_latency_ms\":135742,\"estimated_pass_rate\":0.702,\"estimated_quality\":0.9031,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":15.809,\"max_latency_ms\":196637,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7262, "category": "coding", "complexity": 0.652, "consequence": 0.6705, "evidence_need": 0.836, "family": "migration_planning", "input_tokens": 5120, "request_id": "synthetic-migration_planning-000094", "reversibility": 0.3616, "risk": 0.5648, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 0.782, "id": "worker_a", "p95_latency_ms": 25262, "pass_rate": 0.4955, "provider_group": "provider_y", "quality": 0.778, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.595, "id": "worker_b", "p95_latency_ms": 59792, "pass_rate": 0.7599, "provider_group": "provider_x", "quality": 0.869, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.68, "id": "worker_c", "p95_latency_ms": 50688, "pass_rate": 0.9211, "provider_group": "provider_z", "quality": 0.9483, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.625, "max_latency_ms": 48348, "minimum_quality": 0.6805}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000063", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.127, "estimated_latency_ms": 31594, "estimated_pass_rate": 0.433, "estimated_quality": 0.5886, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.625, "max_latency_ms": 48348, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.433, "quality": 0.5886, "quality_target_met": false, "utility": 34.797013}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.625,\"max_latency_ms\":48348,\"minimum_quality\":0.6805},\"task\":{\"ambiguity\":0.6204,\"category\":\"review\",\"complexity\":0.6608,\"consequence\":0.5539,\"evidence_need\":0.605,\"family\":\"policy_compliance\",\"input_tokens\":7240,\"reversibility\":0.6856,\"risk\":0.7782,\"summary\":\"Audit a workflow for tenant-scoped caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":9.982,\"id\":\"worker_a\",\"p95_latency_ms\":62161,\"pass_rate\":0.9795,\"provider_group\":\"provider_y\",\"quality\":0.9257,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.127,\"id\":\"worker_b\",\"p95_latency_ms\":31594,\"pass_rate\":0.6179,\"provider_group\":\"provider_z\",\"quality\":0.7258,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.99,\"id\":\"worker_c\",\"p95_latency_ms\":77807,\"pass_rate\":0.8389,\"provider_group\":\"provider_x\",\"quality\":0.8722,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.127,\"estimated_latency_ms\":31594,\"estimated_pass_rate\":0.433,\"estimated_quality\":0.5886,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.625,\"max_latency_ms\":48348,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6204, "category": "review", "complexity": 0.6608, "consequence": 0.5539, "evidence_need": 0.605, "family": "policy_compliance", "input_tokens": 7240, "request_id": "synthetic-policy_compliance-000063", "reversibility": 0.6856, "risk": 0.7782, "summary": "Audit a workflow for tenant-scoped caching against explicit access requirements."}, "workers": [{"credits_per_call": 9.982, "id": "worker_a", "p95_latency_ms": 62161, "pass_rate": 0.9795, "provider_group": "provider_y", "quality": 0.9257, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.127, "id": "worker_b", "p95_latency_ms": 31594, "pass_rate": 0.6179, "provider_group": "provider_z", "quality": 0.7258, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.99, "id": "worker_c", "p95_latency_ms": 77807, "pass_rate": 0.8389, "provider_group": "provider_x", "quality": 0.8722, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.588, "max_latency_ms": 32015, "minimum_quality": 0.7691}, "family": "migration_planning", "id": "synthetic-migration_planning-000151", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.932, "estimated_latency_ms": 26389, "estimated_pass_rate": 0.3523, "estimated_quality": 0.5703, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.588, "max_latency_ms": 32015, "policy": "lean", "reason_codes": ["best_feasible_below_target", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3523, "quality": 0.5703, "quality_target_met": false, "utility": 21.38182}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.588,\"max_latency_ms\":32015,\"minimum_quality\":0.7691},\"task\":{\"ambiguity\":0.5844,\"category\":\"coding\",\"complexity\":0.6758,\"consequence\":0.7311,\"evidence_need\":0.59,\"family\":\"migration_planning\",\"input_tokens\":2480,\"reversibility\":0.2755,\"risk\":0.5054,\"summary\":\"Review a staged migration for access control and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.873,\"id\":\"worker_a\",\"p95_latency_ms\":68056,\"pass_rate\":0.8755,\"provider_group\":\"provider_y\",\"quality\":0.811,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.932,\"id\":\"worker_b\",\"p95_latency_ms\":26389,\"pass_rate\":0.5141,\"provider_group\":\"provider_x\",\"quality\":0.6881,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.365,\"id\":\"worker_c\",\"p95_latency_ms\":66262,\"pass_rate\":0.9074,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.932,\"estimated_latency_ms\":26389,\"estimated_pass_rate\":0.3523,\"estimated_quality\":0.5703,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.588,\"max_latency_ms\":32015,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5844, "category": "coding", "complexity": 0.6758, "consequence": 0.7311, "evidence_need": 0.59, "family": "migration_planning", "input_tokens": 2480, "request_id": "synthetic-migration_planning-000151", "reversibility": 0.2755, "risk": 0.5054, "summary": "Review a staged migration for access control and identify irreversible steps."}, "workers": [{"credits_per_call": 2.873, "id": "worker_a", "p95_latency_ms": 68056, "pass_rate": 0.8755, "provider_group": "provider_y", "quality": 0.811, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.932, "id": "worker_b", "p95_latency_ms": 26389, "pass_rate": 0.5141, "provider_group": "provider_x", "quality": 0.6881, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.365, "id": "worker_c", "p95_latency_ms": 66262, "pass_rate": 0.9074, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.99, "max_latency_ms": 37011, "minimum_quality": 0.5593}, "family": "migration_planning", "id": "synthetic-migration_planning-000592", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.845, "estimated_latency_ms": 22293, "estimated_pass_rate": 0.3269, "estimated_quality": 0.5952, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.99, "max_latency_ms": 37011, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3269, "quality": 0.5952, "quality_target_met": true, "utility": 38.738233}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.99,\"max_latency_ms\":37011,\"minimum_quality\":0.5593},\"task\":{\"ambiguity\":0.4131,\"category\":\"coding\",\"complexity\":0.7933,\"consequence\":0.8257,\"evidence_need\":0.6978,\"family\":\"migration_planning\",\"input_tokens\":2400,\"reversibility\":0.5948,\"risk\":0.4663,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.756,\"id\":\"worker_a\",\"p95_latency_ms\":80010,\"pass_rate\":0.9744,\"provider_group\":\"provider_y\",\"quality\":0.9478,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.845,\"id\":\"worker_b\",\"p95_latency_ms\":22293,\"pass_rate\":0.4859,\"provider_group\":\"provider_z\",\"quality\":0.72,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.194,\"id\":\"worker_c\",\"p95_latency_ms\":64991,\"pass_rate\":0.8883,\"provider_group\":\"provider_x\",\"quality\":0.8151,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.845,\"estimated_latency_ms\":22293,\"estimated_pass_rate\":0.3269,\"estimated_quality\":0.5952,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.99,\"max_latency_ms\":37011,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4131, "category": "coding", "complexity": 0.7933, "consequence": 0.8257, "evidence_need": 0.6978, "family": "migration_planning", "input_tokens": 2400, "request_id": "synthetic-migration_planning-000592", "reversibility": 0.5948, "risk": 0.4663, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 11.756, "id": "worker_a", "p95_latency_ms": 80010, "pass_rate": 0.9744, "provider_group": "provider_y", "quality": 0.9478, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.845, "id": "worker_b", "p95_latency_ms": 22293, "pass_rate": 0.4859, "provider_group": "provider_z", "quality": 0.72, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.194, "id": "worker_c", "p95_latency_ms": 64991, "pass_rate": 0.8883, "provider_group": "provider_x", "quality": 0.8151, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.291, "max_latency_ms": 260734, "minimum_quality": 0.8212}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000392", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.151, "estimated_latency_ms": 187614, "estimated_pass_rate": 0.7395, "estimated_quality": 0.9233, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.291, "max_latency_ms": 260734, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7395, "quality": 0.9233, "quality_target_met": true, "utility": 72.241368}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.291,\"max_latency_ms\":260734,\"minimum_quality\":0.8212},\"task\":{\"ambiguity\":0.7344,\"category\":\"analysis\",\"complexity\":0.7363,\"consequence\":0.289,\"evidence_need\":0.6116,\"family\":\"performance_tuning\",\"input_tokens\":2920,\"reversibility\":0.8608,\"risk\":0.353,\"summary\":\"Choose between throughput and tail-latency changes for access control.\"},\"workers\":[{\"credits_per_call\":1.173,\"id\":\"worker_a\",\"p95_latency_ms\":31229,\"pass_rate\":0.4911,\"provider_group\":\"provider_x\",\"quality\":0.732,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.17,\"id\":\"worker_b\",\"p95_latency_ms\":75065,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9908,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.808,\"id\":\"worker_c\",\"p95_latency_ms\":81320,\"pass_rate\":0.8669,\"provider_group\":\"provider_y\",\"quality\":0.9222,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.151,\"estimated_latency_ms\":187614,\"estimated_pass_rate\":0.7395,\"estimated_quality\":0.9233,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.291,\"max_latency_ms\":260734,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7344, "category": "analysis", "complexity": 0.7363, "consequence": 0.289, "evidence_need": 0.6116, "family": "performance_tuning", "input_tokens": 2920, "request_id": "synthetic-performance_tuning-000392", "reversibility": 0.8608, "risk": 0.353, "summary": "Choose between throughput and tail-latency changes for access control."}, "workers": [{"credits_per_call": 1.173, "id": "worker_a", "p95_latency_ms": 31229, "pass_rate": 0.4911, "provider_group": "provider_x", "quality": 0.732, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.17, "id": "worker_b", "p95_latency_ms": 75065, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9908, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.808, "id": "worker_c", "p95_latency_ms": 81320, "pass_rate": 0.8669, "provider_group": "provider_y", "quality": 0.9222, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 6.087, "max_latency_ms": 254917, "minimum_quality": 0.8461}, "family": "migration_planning", "id": "synthetic-migration_planning-000529", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.989, "estimated_latency_ms": 103999, "estimated_pass_rate": 0.5878, "estimated_quality": 0.78, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.087, "max_latency_ms": 254917, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5878, "quality": 0.78, "quality_target_met": false, "utility": 53.563876}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":6.087,\"max_latency_ms\":254917,\"minimum_quality\":0.8461},\"task\":{\"ambiguity\":0.4276,\"category\":\"coding\",\"complexity\":0.9152,\"consequence\":0.7877,\"evidence_need\":0.7423,\"family\":\"migration_planning\",\"input_tokens\":4840,\"reversibility\":0.292,\"risk\":0.4229,\"summary\":\"Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.74,\"id\":\"worker_a\",\"p95_latency_ms\":65744,\"pass_rate\":0.9152,\"provider_group\":\"provider_z\",\"quality\":0.929,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.013,\"id\":\"worker_b\",\"p95_latency_ms\":81366,\"pass_rate\":0.792,\"provider_group\":\"provider_y\",\"quality\":0.8552,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.976,\"id\":\"worker_c\",\"p95_latency_ms\":22633,\"pass_rate\":0.5457,\"provider_group\":\"provider_x\",\"quality\":0.7398,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.989,\"estimated_latency_ms\":103999,\"estimated_pass_rate\":0.5878,\"estimated_quality\":0.78,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.087,\"max_latency_ms\":254917,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4276, "category": "coding", "complexity": 0.9152, "consequence": 0.7877, "evidence_need": 0.7423, "family": "migration_planning", "input_tokens": 4840, "request_id": "synthetic-migration_planning-000529", "reversibility": 0.292, "risk": 0.4229, "summary": "Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints."}, "workers": [{"credits_per_call": 12.74, "id": "worker_a", "p95_latency_ms": 65744, "pass_rate": 0.9152, "provider_group": "provider_z", "quality": 0.929, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.013, "id": "worker_b", "p95_latency_ms": 81366, "pass_rate": 0.792, "provider_group": "provider_y", "quality": 0.8552, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.976, "id": "worker_c", "p95_latency_ms": 22633, "pass_rate": 0.5457, "provider_group": "provider_x", "quality": 0.7398, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.766, "max_latency_ms": 164309, "minimum_quality": 0.782}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000107", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.838, "estimated_latency_ms": 127242, "estimated_pass_rate": 0.9196, "estimated_quality": 0.9597, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.766, "max_latency_ms": 164309, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9196, "quality": 0.9597, "quality_target_met": true, "utility": 71.326226}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.766,\"max_latency_ms\":164309,\"minimum_quality\":0.782},\"task\":{\"ambiguity\":0.4543,\"category\":\"analysis\",\"complexity\":0.4624,\"consequence\":0.5964,\"evidence_need\":0.55,\"family\":\"performance_tuning\",\"input_tokens\":1120,\"reversibility\":0.6286,\"risk\":0.4052,\"summary\":\"Diagnose latency variance in database failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.119,\"id\":\"worker_a\",\"p95_latency_ms\":29391,\"pass_rate\":0.3897,\"provider_group\":\"provider_y\",\"quality\":0.6112,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.163,\"id\":\"worker_b\",\"p95_latency_ms\":74289,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.675,\"id\":\"worker_c\",\"p95_latency_ms\":52953,\"pass_rate\":0.809,\"provider_group\":\"provider_x\",\"quality\":0.9009,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.838,\"estimated_latency_ms\":127242,\"estimated_pass_rate\":0.9196,\"estimated_quality\":0.9597,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.766,\"max_latency_ms\":164309,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4543, "category": "analysis", "complexity": 0.4624, "consequence": 0.5964, "evidence_need": 0.55, "family": "performance_tuning", "input_tokens": 1120, "request_id": "synthetic-performance_tuning-000107", "reversibility": 0.6286, "risk": 0.4052, "summary": "Diagnose latency variance in database failover and rank the next experiments."}, "workers": [{"credits_per_call": 1.119, "id": "worker_a", "p95_latency_ms": 29391, "pass_rate": 0.3897, "provider_group": "provider_y", "quality": 0.6112, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.163, "id": "worker_b", "p95_latency_ms": 74289, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.675, "id": "worker_c", "p95_latency_ms": 52953, "pass_rate": 0.809, "provider_group": "provider_x", "quality": 0.9009, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.204, "max_latency_ms": 128178, "minimum_quality": 0.8374}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000821", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.716, "estimated_latency_ms": 104121, "estimated_pass_rate": 0.7811, "estimated_quality": 0.8487, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.204, "max_latency_ms": 128178, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7811, "quality": 0.8487, "quality_target_met": true, "utility": 57.94478}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.204,\"max_latency_ms\":128178,\"minimum_quality\":0.8374},\"task\":{\"ambiguity\":0.6751,\"category\":\"analysis\",\"complexity\":0.4981,\"consequence\":0.3233,\"evidence_need\":0.5297,\"family\":\"performance_tuning\",\"input_tokens\":2240,\"reversibility\":0.5738,\"risk\":0.2221,\"summary\":\"Choose between throughput and tail-latency changes for artifact storage.\"},\"workers\":[{\"credits_per_call\":14.334,\"id\":\"worker_a\",\"p95_latency_ms\":71813,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.706,\"id\":\"worker_b\",\"p95_latency_ms\":77446,\"pass_rate\":0.8795,\"provider_group\":\"provider_x\",\"quality\":0.9076,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.01,\"id\":\"worker_c\",\"p95_latency_ms\":26675,\"pass_rate\":0.4551,\"provider_group\":\"provider_y\",\"quality\":0.7159,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.716,\"estimated_latency_ms\":104121,\"estimated_pass_rate\":0.7811,\"estimated_quality\":0.8487,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.204,\"max_latency_ms\":128178,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6751, "category": "analysis", "complexity": 0.4981, "consequence": 0.3233, "evidence_need": 0.5297, "family": "performance_tuning", "input_tokens": 2240, "request_id": "synthetic-performance_tuning-000821", "reversibility": 0.5738, "risk": 0.2221, "summary": "Choose between throughput and tail-latency changes for artifact storage."}, "workers": [{"credits_per_call": 14.334, "id": "worker_a", "p95_latency_ms": 71813, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.706, "id": "worker_b", "p95_latency_ms": 77446, "pass_rate": 0.8795, "provider_group": "provider_x", "quality": 0.9076, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.01, "id": "worker_c", "p95_latency_ms": 26675, "pass_rate": 0.4551, "provider_group": "provider_y", "quality": 0.7159, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.66, "max_latency_ms": 197530, "minimum_quality": 0.8067}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000005", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.01, "estimated_latency_ms": 174600, "estimated_pass_rate": 0.7661, "estimated_quality": 0.8958, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 25.66, "max_latency_ms": 197530, "policy": "strict", "reason_codes": ["independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7661, "quality": 0.8958, "quality_target_met": true, "utility": 71.107848}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.66,\"max_latency_ms\":197530,\"minimum_quality\":0.8067},\"task\":{\"ambiguity\":0.6979,\"category\":\"analysis\",\"complexity\":0.7095,\"consequence\":0.5486,\"evidence_need\":0.4208,\"family\":\"performance_tuning\",\"input_tokens\":7480,\"reversibility\":0.7732,\"risk\":0.4824,\"summary\":\"Choose between throughput and tail-latency changes for background jobs.\"},\"workers\":[{\"credits_per_call\":0.791,\"id\":\"worker_a\",\"p95_latency_ms\":32678,\"pass_rate\":0.5484,\"provider_group\":\"provider_z\",\"quality\":0.618,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.883,\"id\":\"worker_b\",\"p95_latency_ms\":81503,\"pass_rate\":0.8634,\"provider_group\":\"provider_x\",\"quality\":0.8862,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.336,\"id\":\"worker_c\",\"p95_latency_ms\":60419,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.01,\"estimated_latency_ms\":174600,\"estimated_pass_rate\":0.7661,\"estimated_quality\":0.8958,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.66,\"max_latency_ms\":197530,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6979, "category": "analysis", "complexity": 0.7095, "consequence": 0.5486, "evidence_need": 0.4208, "family": "performance_tuning", "input_tokens": 7480, "request_id": "synthetic-performance_tuning-000005", "reversibility": 0.7732, "risk": 0.4824, "summary": "Choose between throughput and tail-latency changes for background jobs."}, "workers": [{"credits_per_call": 0.791, "id": "worker_a", "p95_latency_ms": 32678, "pass_rate": 0.5484, "provider_group": "provider_z", "quality": 0.618, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.883, "id": "worker_b", "p95_latency_ms": 81503, "pass_rate": 0.8634, "provider_group": "provider_x", "quality": 0.8862, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.336, "id": "worker_c", "p95_latency_ms": 60419, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.06, "max_latency_ms": 192053, "minimum_quality": 0.8554}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001070", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.253, "estimated_latency_ms": 166873, "estimated_pass_rate": 0.7488, "estimated_quality": 0.915, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.06, "max_latency_ms": 192053, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7488, "quality": 0.915, "quality_target_met": true, "utility": 69.899294}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.06,\"max_latency_ms\":192053,\"minimum_quality\":0.8554},\"task\":{\"ambiguity\":0.6417,\"category\":\"analysis\",\"complexity\":0.7984,\"consequence\":0.5889,\"evidence_need\":0.4003,\"family\":\"performance_tuning\",\"input_tokens\":4400,\"reversibility\":0.8327,\"risk\":0.2647,\"summary\":\"Choose between throughput and tail-latency changes for streaming inference.\"},\"workers\":[{\"credits_per_call\":12.931,\"id\":\"worker_a\",\"p95_latency_ms\":91446,\"pass_rate\":0.9693,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.184,\"id\":\"worker_b\",\"p95_latency_ms\":55561,\"pass_rate\":0.9239,\"provider_group\":\"provider_y\",\"quality\":0.9067,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.138,\"id\":\"worker_c\",\"p95_latency_ms\":19866,\"pass_rate\":0.5407,\"provider_group\":\"provider_z\",\"quality\":0.7132,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.253,\"estimated_latency_ms\":166873,\"estimated_pass_rate\":0.7488,\"estimated_quality\":0.915,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.06,\"max_latency_ms\":192053,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6417, "category": "analysis", "complexity": 0.7984, "consequence": 0.5889, "evidence_need": 0.4003, "family": "performance_tuning", "input_tokens": 4400, "request_id": "synthetic-performance_tuning-001070", "reversibility": 0.8327, "risk": 0.2647, "summary": "Choose between throughput and tail-latency changes for streaming inference."}, "workers": [{"credits_per_call": 12.931, "id": "worker_a", "p95_latency_ms": 91446, "pass_rate": 0.9693, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.184, "id": "worker_b", "p95_latency_ms": 55561, "pass_rate": 0.9239, "provider_group": "provider_y", "quality": 0.9067, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.138, "id": "worker_c", "p95_latency_ms": 19866, "pass_rate": 0.5407, "provider_group": "provider_z", "quality": 0.7132, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.915, "max_latency_ms": 48502, "minimum_quality": 0.6177}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000755", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.047, "estimated_latency_ms": 29408, "estimated_pass_rate": 0.3638, "estimated_quality": 0.5199, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.915, "max_latency_ms": 48502, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3638, "quality": 0.5199, "quality_target_met": false, "utility": 29.937981}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.915,\"max_latency_ms\":48502,\"minimum_quality\":0.6177},\"task\":{\"ambiguity\":0.652,\"category\":\"analysis\",\"complexity\":0.5809,\"consequence\":0.2542,\"evidence_need\":0.5969,\"family\":\"performance_tuning\",\"input_tokens\":7400,\"reversibility\":0.5599,\"risk\":0.5645,\"summary\":\"Review a proposed optimization for model routing for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.505,\"id\":\"worker_a\",\"p95_latency_ms\":56053,\"pass_rate\":0.9523,\"provider_group\":\"provider_x\",\"quality\":0.9021,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.047,\"id\":\"worker_b\",\"p95_latency_ms\":29408,\"pass_rate\":0.5247,\"provider_group\":\"provider_z\",\"quality\":0.6294,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.498,\"id\":\"worker_c\",\"p95_latency_ms\":74670,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.047,\"estimated_latency_ms\":29408,\"estimated_pass_rate\":0.3638,\"estimated_quality\":0.5199,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.915,\"max_latency_ms\":48502,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.652, "category": "analysis", "complexity": 0.5809, "consequence": 0.2542, "evidence_need": 0.5969, "family": "performance_tuning", "input_tokens": 7400, "request_id": "synthetic-performance_tuning-000755", "reversibility": 0.5599, "risk": 0.5645, "summary": "Review a proposed optimization for model routing for hidden regressions."}, "workers": [{"credits_per_call": 3.505, "id": "worker_a", "p95_latency_ms": 56053, "pass_rate": 0.9523, "provider_group": "provider_x", "quality": 0.9021, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.047, "id": "worker_b", "p95_latency_ms": 29408, "pass_rate": 0.5247, "provider_group": "provider_z", "quality": 0.6294, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.498, "id": "worker_c", "p95_latency_ms": 74670, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.273, "max_latency_ms": 109323, "minimum_quality": 0.7345}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000498", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.755, "estimated_latency_ms": 70465, "estimated_pass_rate": 0.6516, "estimated_quality": 0.83, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.273, "max_latency_ms": 109323, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6516, "quality": 0.83, "quality_target_met": true, "utility": 61.95112}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.273,\"max_latency_ms\":109323,\"minimum_quality\":0.7345},\"task\":{\"ambiguity\":0.5193,\"category\":\"review\",\"complexity\":0.6009,\"consequence\":0.7284,\"evidence_need\":0.885,\"family\":\"policy_compliance\",\"input_tokens\":4560,\"reversibility\":0.4294,\"risk\":0.7501,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.801,\"id\":\"worker_a\",\"p95_latency_ms\":25968,\"pass_rate\":0.6071,\"provider_group\":\"provider_y\",\"quality\":0.7866,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.954,\"id\":\"worker_b\",\"p95_latency_ms\":44497,\"pass_rate\":0.7433,\"provider_group\":\"provider_z\",\"quality\":0.8336,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.547,\"id\":\"worker_c\",\"p95_latency_ms\":86458,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.755,\"estimated_latency_ms\":70465,\"estimated_pass_rate\":0.6516,\"estimated_quality\":0.83,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.273,\"max_latency_ms\":109323,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5193, "category": "review", "complexity": 0.6009, "consequence": 0.7284, "evidence_need": 0.885, "family": "policy_compliance", "input_tokens": 4560, "request_id": "synthetic-policy_compliance-000498", "reversibility": 0.4294, "risk": 0.7501, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 0.801, "id": "worker_a", "p95_latency_ms": 25968, "pass_rate": 0.6071, "provider_group": "provider_y", "quality": 0.7866, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.954, "id": "worker_b", "p95_latency_ms": 44497, "pass_rate": 0.7433, "provider_group": "provider_z", "quality": 0.8336, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.547, "id": "worker_c", "p95_latency_ms": 86458, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.076, "max_latency_ms": 40114, "minimum_quality": 0.6845}, "family": "migration_planning", "id": "synthetic-migration_planning-000673", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.967, "estimated_latency_ms": 24776, "estimated_pass_rate": 0.4631, "estimated_quality": 0.5891, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.076, "max_latency_ms": 40114, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4631, "quality": 0.5891, "quality_target_met": false, "utility": 38.188736}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.076,\"max_latency_ms\":40114,\"minimum_quality\":0.6845},\"task\":{\"ambiguity\":0.3186,\"category\":\"coding\",\"complexity\":0.6553,\"consequence\":0.7293,\"evidence_need\":0.775,\"family\":\"migration_planning\",\"input_tokens\":5880,\"reversibility\":0.6424,\"risk\":0.4664,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":4.048,\"id\":\"worker_a\",\"p95_latency_ms\":51073,\"pass_rate\":0.8644,\"provider_group\":\"provider_z\",\"quality\":0.8162,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.967,\"id\":\"worker_b\",\"p95_latency_ms\":24776,\"pass_rate\":0.5993,\"provider_group\":\"provider_y\",\"quality\":0.6826,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.506,\"id\":\"worker_c\",\"p95_latency_ms\":60790,\"pass_rate\":0.9644,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.967,\"estimated_latency_ms\":24776,\"estimated_pass_rate\":0.4631,\"estimated_quality\":0.5891,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.076,\"max_latency_ms\":40114,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3186, "category": "coding", "complexity": 0.6553, "consequence": 0.7293, "evidence_need": 0.775, "family": "migration_planning", "input_tokens": 5880, "request_id": "synthetic-migration_planning-000673", "reversibility": 0.6424, "risk": 0.4664, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 4.048, "id": "worker_a", "p95_latency_ms": 51073, "pass_rate": 0.8644, "provider_group": "provider_z", "quality": 0.8162, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.967, "id": "worker_b", "p95_latency_ms": 24776, "pass_rate": 0.5993, "provider_group": "provider_y", "quality": 0.6826, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.506, "id": "worker_c", "p95_latency_ms": 60790, "pass_rate": 0.9644, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.706, "max_latency_ms": 147730, "minimum_quality": 0.8477}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001110", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.041, "estimated_latency_ms": 142377, "estimated_pass_rate": 0.7348, "estimated_quality": 0.9127, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.706, "max_latency_ms": 147730, "policy": "strict", "reason_codes": ["high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7348, "quality": 0.9127, "quality_target_met": true, "utility": 74.136889}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.706,\"max_latency_ms\":147730,\"minimum_quality\":0.8477},\"task\":{\"ambiguity\":0.4701,\"category\":\"review\",\"complexity\":0.7082,\"consequence\":0.6515,\"evidence_need\":0.6201,\"family\":\"policy_compliance\",\"input_tokens\":1200,\"reversibility\":0.5521,\"risk\":0.7451,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.744,\"id\":\"worker_a\",\"p95_latency_ms\":46350,\"pass_rate\":0.8038,\"provider_group\":\"provider_y\",\"quality\":0.8528,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.984,\"id\":\"worker_b\",\"p95_latency_ms\":25406,\"pass_rate\":0.5515,\"provider_group\":\"provider_z\",\"quality\":0.7475,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.313,\"id\":\"worker_c\",\"p95_latency_ms\":70621,\"pass_rate\":0.894,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.041,\"estimated_latency_ms\":142377,\"estimated_pass_rate\":0.7348,\"estimated_quality\":0.9127,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.706,\"max_latency_ms\":147730,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4701, "category": "review", "complexity": 0.7082, "consequence": 0.6515, "evidence_need": 0.6201, "family": "policy_compliance", "input_tokens": 1200, "request_id": "synthetic-policy_compliance-001110", "reversibility": 0.5521, "risk": 0.7451, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.744, "id": "worker_a", "p95_latency_ms": 46350, "pass_rate": 0.8038, "provider_group": "provider_y", "quality": 0.8528, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.984, "id": "worker_b", "p95_latency_ms": 25406, "pass_rate": 0.5515, "provider_group": "provider_z", "quality": 0.7475, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.313, "id": "worker_c", "p95_latency_ms": 70621, "pass_rate": 0.894, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.053, "max_latency_ms": 42008, "minimum_quality": 0.7046}, "family": "migration_planning", "id": "synthetic-migration_planning-000694", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.785, "estimated_latency_ms": 28561, "estimated_pass_rate": 0.4423, "estimated_quality": 0.6296, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.053, "max_latency_ms": 42008, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4423, "quality": 0.6296, "quality_target_met": false, "utility": 37.780152}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.053,\"max_latency_ms\":42008,\"minimum_quality\":0.7046},\"task\":{\"ambiguity\":0.7321,\"category\":\"coding\",\"complexity\":0.5876,\"consequence\":0.8451,\"evidence_need\":0.4074,\"family\":\"migration_planning\",\"input_tokens\":6880,\"reversibility\":0.6322,\"risk\":0.6553,\"summary\":\"Plan a zero-downtime migration of access control with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.876,\"id\":\"worker_a\",\"p95_latency_ms\":45775,\"pass_rate\":0.7299,\"provider_group\":\"provider_z\",\"quality\":0.8949,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.79,\"id\":\"worker_b\",\"p95_latency_ms\":75082,\"pass_rate\":0.905,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.785,\"id\":\"worker_c\",\"p95_latency_ms\":28561,\"pass_rate\":0.6174,\"provider_group\":\"provider_y\",\"quality\":0.7522,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.785,\"estimated_latency_ms\":28561,\"estimated_pass_rate\":0.4423,\"estimated_quality\":0.6296,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.053,\"max_latency_ms\":42008,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7321, "category": "coding", "complexity": 0.5876, "consequence": 0.8451, "evidence_need": 0.4074, "family": "migration_planning", "input_tokens": 6880, "request_id": "synthetic-migration_planning-000694", "reversibility": 0.6322, "risk": 0.6553, "summary": "Plan a zero-downtime migration of access control with rollback checkpoints."}, "workers": [{"credits_per_call": 2.876, "id": "worker_a", "p95_latency_ms": 45775, "pass_rate": 0.7299, "provider_group": "provider_z", "quality": 0.8949, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.79, "id": "worker_b", "p95_latency_ms": 75082, "pass_rate": 0.905, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.785, "id": "worker_c", "p95_latency_ms": 28561, "pass_rate": 0.6174, "provider_group": "provider_y", "quality": 0.7522, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.293, "max_latency_ms": 47668, "minimum_quality": 0.51}, "family": "migration_planning", "id": "synthetic-migration_planning-000763", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.895, "estimated_latency_ms": 29578, "estimated_pass_rate": 0.3006, "estimated_quality": 0.6249, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.293, "max_latency_ms": 47668, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3006, "quality": 0.6249, "quality_target_met": true, "utility": 40.133121}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.293,\"max_latency_ms\":47668,\"minimum_quality\":0.51},\"task\":{\"ambiguity\":0.7473,\"category\":\"coding\",\"complexity\":0.8497,\"consequence\":0.7036,\"evidence_need\":0.4136,\"family\":\"migration_planning\",\"input_tokens\":5800,\"reversibility\":0.5752,\"risk\":0.5929,\"summary\":\"Plan a zero-downtime migration of rate limiting with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":11.583,\"id\":\"worker_a\",\"p95_latency_ms\":53237,\"pass_rate\":0.9204,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.197,\"id\":\"worker_b\",\"p95_latency_ms\":47800,\"pass_rate\":0.7317,\"provider_group\":\"provider_z\",\"quality\":0.9221,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.895,\"id\":\"worker_c\",\"p95_latency_ms\":29578,\"pass_rate\":0.5015,\"provider_group\":\"provider_x\",\"quality\":0.7914,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.895,\"estimated_latency_ms\":29578,\"estimated_pass_rate\":0.3006,\"estimated_quality\":0.6249,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.293,\"max_latency_ms\":47668,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7473, "category": "coding", "complexity": 0.8497, "consequence": 0.7036, "evidence_need": 0.4136, "family": "migration_planning", "input_tokens": 5800, "request_id": "synthetic-migration_planning-000763", "reversibility": 0.5752, "risk": 0.5929, "summary": "Plan a zero-downtime migration of rate limiting with rollback checkpoints."}, "workers": [{"credits_per_call": 11.583, "id": "worker_a", "p95_latency_ms": 53237, "pass_rate": 0.9204, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.197, "id": "worker_b", "p95_latency_ms": 47800, "pass_rate": 0.7317, "provider_group": "provider_z", "quality": 0.9221, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.895, "id": "worker_c", "p95_latency_ms": 29578, "pass_rate": 0.5015, "provider_group": "provider_x", "quality": 0.7914, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.421, "max_latency_ms": 33656, "minimum_quality": 0.7223}, "family": "migration_planning", "id": "synthetic-migration_planning-000130", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.088, "estimated_latency_ms": 19563, "estimated_pass_rate": 0.3272, "estimated_quality": 0.6134, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.421, "max_latency_ms": 33656, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3272, "quality": 0.6134, "quality_target_met": false, "utility": 35.221107}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.421,\"max_latency_ms\":33656,\"minimum_quality\":0.7223},\"task\":{\"ambiguity\":0.3098,\"category\":\"coding\",\"complexity\":0.8569,\"consequence\":0.4329,\"evidence_need\":0.4554,\"family\":\"migration_planning\",\"input_tokens\":280,\"reversibility\":0.5503,\"risk\":0.6729,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":12.882,\"id\":\"worker_a\",\"p95_latency_ms\":74348,\"pass_rate\":0.9251,\"provider_group\":\"provider_y\",\"quality\":0.956,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.088,\"id\":\"worker_b\",\"p95_latency_ms\":19563,\"pass_rate\":0.5023,\"provider_group\":\"provider_z\",\"quality\":0.7576,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.376,\"id\":\"worker_c\",\"p95_latency_ms\":62747,\"pass_rate\":0.7539,\"provider_group\":\"provider_x\",\"quality\":0.8738,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.088,\"estimated_latency_ms\":19563,\"estimated_pass_rate\":0.3272,\"estimated_quality\":0.6134,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.421,\"max_latency_ms\":33656,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3098, "category": "coding", "complexity": 0.8569, "consequence": 0.4329, "evidence_need": 0.4554, "family": "migration_planning", "input_tokens": 280, "request_id": "synthetic-migration_planning-000130", "reversibility": 0.5503, "risk": 0.6729, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 12.882, "id": "worker_a", "p95_latency_ms": 74348, "pass_rate": 0.9251, "provider_group": "provider_y", "quality": 0.956, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.088, "id": "worker_b", "p95_latency_ms": 19563, "pass_rate": 0.5023, "provider_group": "provider_z", "quality": 0.7576, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.376, "id": "worker_c", "p95_latency_ms": 62747, "pass_rate": 0.7539, "provider_group": "provider_x", "quality": 0.8738, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.492, "max_latency_ms": 130129, "minimum_quality": 0.8819}, "family": "migration_planning", "id": "synthetic-migration_planning-001234", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.316, "estimated_latency_ms": 92319, "estimated_pass_rate": 0.8698, "estimated_quality": 0.8706, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.492, "max_latency_ms": 130129, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8698, "quality": 0.8706, "quality_target_met": false, "utility": 62.892586}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.492,\"max_latency_ms\":130129,\"minimum_quality\":0.8819},\"task\":{\"ambiguity\":0.3302,\"category\":\"coding\",\"complexity\":0.7691,\"consequence\":0.7222,\"evidence_need\":0.4682,\"family\":\"migration_planning\",\"input_tokens\":8000,\"reversibility\":0.5556,\"risk\":0.5646,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":10.371,\"id\":\"worker_a\",\"p95_latency_ms\":71485,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9611,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.945,\"id\":\"worker_b\",\"p95_latency_ms\":20834,\"pass_rate\":0.513,\"provider_group\":\"provider_y\",\"quality\":0.8018,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.102,\"id\":\"worker_c\",\"p95_latency_ms\":48308,\"pass_rate\":0.7315,\"provider_group\":\"provider_z\",\"quality\":0.8792,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.316,\"estimated_latency_ms\":92319,\"estimated_pass_rate\":0.8698,\"estimated_quality\":0.8706,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.492,\"max_latency_ms\":130129,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3302, "category": "coding", "complexity": 0.7691, "consequence": 0.7222, "evidence_need": 0.4682, "family": "migration_planning", "input_tokens": 8000, "request_id": "synthetic-migration_planning-001234", "reversibility": 0.5556, "risk": 0.5646, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 10.371, "id": "worker_a", "p95_latency_ms": 71485, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9611, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.945, "id": "worker_b", "p95_latency_ms": 20834, "pass_rate": 0.513, "provider_group": "provider_y", "quality": 0.8018, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.102, "id": "worker_c", "p95_latency_ms": 48308, "pass_rate": 0.7315, "provider_group": "provider_z", "quality": 0.8792, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.12, "max_latency_ms": 166431, "minimum_quality": 0.8646}, "family": "migration_planning", "id": "synthetic-migration_planning-000526", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.038, "estimated_latency_ms": 146172, "estimated_pass_rate": 0.7657, "estimated_quality": 0.9265, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.12, "max_latency_ms": 166431, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7657, "quality": 0.9265, "quality_target_met": true, "utility": 76.041542}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.12,\"max_latency_ms\":166431,\"minimum_quality\":0.8646},\"task\":{\"ambiguity\":0.6929,\"category\":\"coding\",\"complexity\":0.5987,\"consequence\":0.8295,\"evidence_need\":0.8244,\"family\":\"migration_planning\",\"input_tokens\":4200,\"reversibility\":0.5935,\"risk\":0.5872,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":4.184,\"id\":\"worker_a\",\"p95_latency_ms\":52107,\"pass_rate\":0.81,\"provider_group\":\"provider_x\",\"quality\":0.9023,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.202,\"id\":\"worker_b\",\"p95_latency_ms\":23901,\"pass_rate\":0.5344,\"provider_group\":\"provider_z\",\"quality\":0.7365,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.652,\"id\":\"worker_c\",\"p95_latency_ms\":70164,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9884,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.038,\"estimated_latency_ms\":146172,\"estimated_pass_rate\":0.7657,\"estimated_quality\":0.9265,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.12,\"max_latency_ms\":166431,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6929, "category": "coding", "complexity": 0.5987, "consequence": 0.8295, "evidence_need": 0.8244, "family": "migration_planning", "input_tokens": 4200, "request_id": "synthetic-migration_planning-000526", "reversibility": 0.5935, "risk": 0.5872, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 4.184, "id": "worker_a", "p95_latency_ms": 52107, "pass_rate": 0.81, "provider_group": "provider_x", "quality": 0.9023, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.202, "id": "worker_b", "p95_latency_ms": 23901, "pass_rate": 0.5344, "provider_group": "provider_z", "quality": 0.7365, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.652, "id": "worker_c", "p95_latency_ms": 70164, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9884, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.974, "max_latency_ms": 51708, "minimum_quality": 0.7479}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000146", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.988, "estimated_latency_ms": 27229, "estimated_pass_rate": 0.3487, "estimated_quality": 0.6214, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.974, "max_latency_ms": 51708, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3487, "quality": 0.6214, "quality_target_met": false, "utility": 35.360692}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.974,\"max_latency_ms\":51708,\"minimum_quality\":0.7479},\"task\":{\"ambiguity\":0.5009,\"category\":\"analysis\",\"complexity\":0.7852,\"consequence\":0.2993,\"evidence_need\":0.685,\"family\":\"performance_tuning\",\"input_tokens\":3840,\"reversibility\":0.5732,\"risk\":0.2594,\"summary\":\"Diagnose latency variance in deployment rollback and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.988,\"id\":\"worker_a\",\"p95_latency_ms\":27229,\"pass_rate\":0.4966,\"provider_group\":\"provider_z\",\"quality\":0.7357,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.784,\"id\":\"worker_b\",\"p95_latency_ms\":75273,\"pass_rate\":0.7978,\"provider_group\":\"provider_x\",\"quality\":0.9242,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.44,\"id\":\"worker_c\",\"p95_latency_ms\":78732,\"pass_rate\":0.9725,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.988,\"estimated_latency_ms\":27229,\"estimated_pass_rate\":0.3487,\"estimated_quality\":0.6214,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.974,\"max_latency_ms\":51708,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5009, "category": "analysis", "complexity": 0.7852, "consequence": 0.2993, "evidence_need": 0.685, "family": "performance_tuning", "input_tokens": 3840, "request_id": "synthetic-performance_tuning-000146", "reversibility": 0.5732, "risk": 0.2594, "summary": "Diagnose latency variance in deployment rollback and rank the next experiments."}, "workers": [{"credits_per_call": 0.988, "id": "worker_a", "p95_latency_ms": 27229, "pass_rate": 0.4966, "provider_group": "provider_z", "quality": 0.7357, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.784, "id": "worker_b", "p95_latency_ms": 75273, "pass_rate": 0.7978, "provider_group": "provider_x", "quality": 0.9242, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.44, "id": "worker_c", "p95_latency_ms": 78732, "pass_rate": 0.9725, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.851, "max_latency_ms": 152120, "minimum_quality": 0.7813}, "family": "migration_planning", "id": "synthetic-migration_planning-001003", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.966, "estimated_latency_ms": 146071, "estimated_pass_rate": 0.7478, "estimated_quality": 0.9348, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.851, "max_latency_ms": 152120, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7478, "quality": 0.9348, "quality_target_met": true, "utility": 72.833384}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.851,\"max_latency_ms\":152120,\"minimum_quality\":0.7813},\"task\":{\"ambiguity\":0.3254,\"category\":\"coding\",\"complexity\":0.7489,\"consequence\":0.7804,\"evidence_need\":0.8233,\"family\":\"migration_planning\",\"input_tokens\":800,\"reversibility\":0.2738,\"risk\":0.4559,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":0.804,\"id\":\"worker_a\",\"p95_latency_ms\":20853,\"pass_rate\":0.5297,\"provider_group\":\"provider_y\",\"quality\":0.8019,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.238,\"id\":\"worker_b\",\"p95_latency_ms\":49677,\"pass_rate\":0.9704,\"provider_group\":\"provider_x\",\"quality\":0.9874,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.924,\"id\":\"worker_c\",\"p95_latency_ms\":75541,\"pass_rate\":0.8204,\"provider_group\":\"provider_z\",\"quality\":0.8774,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.966,\"estimated_latency_ms\":146071,\"estimated_pass_rate\":0.7478,\"estimated_quality\":0.9348,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.851,\"max_latency_ms\":152120,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3254, "category": "coding", "complexity": 0.7489, "consequence": 0.7804, "evidence_need": 0.8233, "family": "migration_planning", "input_tokens": 800, "request_id": "synthetic-migration_planning-001003", "reversibility": 0.2738, "risk": 0.4559, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 0.804, "id": "worker_a", "p95_latency_ms": 20853, "pass_rate": 0.5297, "provider_group": "provider_y", "quality": 0.8019, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.238, "id": "worker_b", "p95_latency_ms": 49677, "pass_rate": 0.9704, "provider_group": "provider_x", "quality": 0.9874, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.924, "id": "worker_c", "p95_latency_ms": 75541, "pass_rate": 0.8204, "provider_group": "provider_z", "quality": 0.8774, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.396, "max_latency_ms": 168308, "minimum_quality": 0.7189}, "family": "migration_planning", "id": "synthetic-migration_planning-001258", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.373, "estimated_latency_ms": 83743, "estimated_pass_rate": 0.7795, "estimated_quality": 0.8081, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.396, "max_latency_ms": 168308, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7795, "quality": 0.8081, "quality_target_met": true, "utility": 63.075995}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.396,\"max_latency_ms\":168308,\"minimum_quality\":0.7189},\"task\":{\"ambiguity\":0.3769,\"category\":\"coding\",\"complexity\":0.7958,\"consequence\":0.735,\"evidence_need\":0.4472,\"family\":\"migration_planning\",\"input_tokens\":1040,\"reversibility\":0.4596,\"risk\":0.5641,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.434,\"id\":\"worker_a\",\"p95_latency_ms\":52460,\"pass_rate\":0.8827,\"provider_group\":\"provider_x\",\"quality\":0.8861,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.986,\"id\":\"worker_b\",\"p95_latency_ms\":85405,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9575,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.939,\"id\":\"worker_c\",\"p95_latency_ms\":31283,\"pass_rate\":0.6174,\"provider_group\":\"provider_y\",\"quality\":0.6874,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.373,\"estimated_latency_ms\":83743,\"estimated_pass_rate\":0.7795,\"estimated_quality\":0.8081,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.396,\"max_latency_ms\":168308,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3769, "category": "coding", "complexity": 0.7958, "consequence": 0.735, "evidence_need": 0.4472, "family": "migration_planning", "input_tokens": 1040, "request_id": "synthetic-migration_planning-001258", "reversibility": 0.4596, "risk": 0.5641, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 3.434, "id": "worker_a", "p95_latency_ms": 52460, "pass_rate": 0.8827, "provider_group": "provider_x", "quality": 0.8861, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.986, "id": "worker_b", "p95_latency_ms": 85405, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9575, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.939, "id": "worker_c", "p95_latency_ms": 31283, "pass_rate": 0.6174, "provider_group": "provider_y", "quality": 0.6874, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.986, "max_latency_ms": 248055, "minimum_quality": 0.7859}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001023", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.309, "estimated_latency_ms": 96461, "estimated_pass_rate": 0.6052, "estimated_quality": 0.7891, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.986, "max_latency_ms": 248055, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6052, "quality": 0.7891, "quality_target_met": true, "utility": 61.485751}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.986,\"max_latency_ms\":248055,\"minimum_quality\":0.7859},\"task\":{\"ambiguity\":0.6766,\"category\":\"review\",\"complexity\":0.5692,\"consequence\":0.7103,\"evidence_need\":0.6357,\"family\":\"policy_compliance\",\"input_tokens\":3200,\"reversibility\":0.6952,\"risk\":0.7877,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.138,\"id\":\"worker_a\",\"p95_latency_ms\":64795,\"pass_rate\":0.7874,\"provider_group\":\"provider_y\",\"quality\":0.8839,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.64,\"id\":\"worker_b\",\"p95_latency_ms\":87754,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9553,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.171,\"id\":\"worker_c\",\"p95_latency_ms\":31666,\"pass_rate\":0.4792,\"provider_group\":\"provider_x\",\"quality\":0.7094,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.309,\"estimated_latency_ms\":96461,\"estimated_pass_rate\":0.6052,\"estimated_quality\":0.7891,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.986,\"max_latency_ms\":248055,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6766, "category": "review", "complexity": 0.5692, "consequence": 0.7103, "evidence_need": 0.6357, "family": "policy_compliance", "input_tokens": 3200, "request_id": "synthetic-policy_compliance-001023", "reversibility": 0.6952, "risk": 0.7877, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 4.138, "id": "worker_a", "p95_latency_ms": 64795, "pass_rate": 0.7874, "provider_group": "provider_y", "quality": 0.8839, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.64, "id": "worker_b", "p95_latency_ms": 87754, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9553, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.171, "id": "worker_c", "p95_latency_ms": 31666, "pass_rate": 0.4792, "provider_group": "provider_x", "quality": 0.7094, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.018, "max_latency_ms": 27521, "minimum_quality": 0.6093}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000165", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.214, "estimated_latency_ms": 17998, "estimated_pass_rate": 0.304, "estimated_quality": 0.6013, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.018, "max_latency_ms": 27521, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.304, "quality": 0.6013, "quality_target_met": false, "utility": 36.78262}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.018,\"max_latency_ms\":27521,\"minimum_quality\":0.6093},\"task\":{\"ambiguity\":0.5548,\"category\":\"review\",\"complexity\":0.4993,\"consequence\":0.5745,\"evidence_need\":0.8388,\"family\":\"policy_compliance\",\"input_tokens\":6800,\"reversibility\":0.3388,\"risk\":0.7467,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.214,\"id\":\"worker_a\",\"p95_latency_ms\":17998,\"pass_rate\":0.463,\"provider_group\":\"provider_y\",\"quality\":0.7026,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.355,\"id\":\"worker_b\",\"p95_latency_ms\":70017,\"pass_rate\":0.8981,\"provider_group\":\"provider_x\",\"quality\":0.929,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.1,\"id\":\"worker_c\",\"p95_latency_ms\":65451,\"pass_rate\":0.8234,\"provider_group\":\"provider_z\",\"quality\":0.8633,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.214,\"estimated_latency_ms\":17998,\"estimated_pass_rate\":0.304,\"estimated_quality\":0.6013,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.018,\"max_latency_ms\":27521,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5548, "category": "review", "complexity": 0.4993, "consequence": 0.5745, "evidence_need": 0.8388, "family": "policy_compliance", "input_tokens": 6800, "request_id": "synthetic-policy_compliance-000165", "reversibility": 0.3388, "risk": 0.7467, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 1.214, "id": "worker_a", "p95_latency_ms": 17998, "pass_rate": 0.463, "provider_group": "provider_y", "quality": 0.7026, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.355, "id": "worker_b", "p95_latency_ms": 70017, "pass_rate": 0.8981, "provider_group": "provider_x", "quality": 0.929, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.1, "id": "worker_c", "p95_latency_ms": 65451, "pass_rate": 0.8234, "provider_group": "provider_z", "quality": 0.8633, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.781, "max_latency_ms": 158385, "minimum_quality": 0.8175}, "family": "migration_planning", "id": "synthetic-migration_planning-000148", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.693, "estimated_latency_ms": 150625, "estimated_pass_rate": 0.6976, "estimated_quality": 0.9152, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.781, "max_latency_ms": 158385, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6976, "quality": 0.9152, "quality_target_met": true, "utility": 69.658458}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.781,\"max_latency_ms\":158385,\"minimum_quality\":0.8175},\"task\":{\"ambiguity\":0.4571,\"category\":\"coding\",\"complexity\":0.8205,\"consequence\":0.5641,\"evidence_need\":0.7055,\"family\":\"migration_planning\",\"input_tokens\":5520,\"reversibility\":0.5371,\"risk\":0.3726,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":13.08,\"id\":\"worker_a\",\"p95_latency_ms\":70332,\"pass_rate\":0.9265,\"provider_group\":\"provider_z\",\"quality\":0.9716,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.879,\"id\":\"worker_b\",\"p95_latency_ms\":29040,\"pass_rate\":0.4968,\"provider_group\":\"provider_x\",\"quality\":0.795,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.734,\"id\":\"worker_c\",\"p95_latency_ms\":51253,\"pass_rate\":0.7829,\"provider_group\":\"provider_y\",\"quality\":0.8691,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.693,\"estimated_latency_ms\":150625,\"estimated_pass_rate\":0.6976,\"estimated_quality\":0.9152,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.781,\"max_latency_ms\":158385,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4571, "category": "coding", "complexity": 0.8205, "consequence": 0.5641, "evidence_need": 0.7055, "family": "migration_planning", "input_tokens": 5520, "request_id": "synthetic-migration_planning-000148", "reversibility": 0.5371, "risk": 0.3726, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 13.08, "id": "worker_a", "p95_latency_ms": 70332, "pass_rate": 0.9265, "provider_group": "provider_z", "quality": 0.9716, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.879, "id": "worker_b", "p95_latency_ms": 29040, "pass_rate": 0.4968, "provider_group": "provider_x", "quality": 0.795, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.734, "id": "worker_c", "p95_latency_ms": 51253, "pass_rate": 0.7829, "provider_group": "provider_y", "quality": 0.8691, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.586, "max_latency_ms": 45499, "minimum_quality": 0.614}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000204", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.989, "estimated_latency_ms": 24137, "estimated_pass_rate": 0.3935, "estimated_quality": 0.653, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.586, "max_latency_ms": 45499, "policy": "lean", "reason_codes": ["high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3935, "quality": 0.653, "quality_target_met": true, "utility": 44.472164}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.586,\"max_latency_ms\":45499,\"minimum_quality\":0.614},\"task\":{\"ambiguity\":0.5873,\"category\":\"review\",\"complexity\":0.5351,\"consequence\":0.8369,\"evidence_need\":0.6109,\"family\":\"policy_compliance\",\"input_tokens\":4000,\"reversibility\":0.5796,\"risk\":0.5897,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":0.989,\"id\":\"worker_a\",\"p95_latency_ms\":24137,\"pass_rate\":0.5463,\"provider_group\":\"provider_x\",\"quality\":0.7517,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.741,\"id\":\"worker_b\",\"p95_latency_ms\":60619,\"pass_rate\":0.8266,\"provider_group\":\"provider_y\",\"quality\":0.9199,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.005,\"id\":\"worker_c\",\"p95_latency_ms\":54558,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.989,\"estimated_latency_ms\":24137,\"estimated_pass_rate\":0.3935,\"estimated_quality\":0.653,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.586,\"max_latency_ms\":45499,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5873, "category": "review", "complexity": 0.5351, "consequence": 0.8369, "evidence_need": 0.6109, "family": "policy_compliance", "input_tokens": 4000, "request_id": "synthetic-policy_compliance-000204", "reversibility": 0.5796, "risk": 0.5897, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 0.989, "id": "worker_a", "p95_latency_ms": 24137, "pass_rate": 0.5463, "provider_group": "provider_x", "quality": 0.7517, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.741, "id": "worker_b", "p95_latency_ms": 60619, "pass_rate": 0.8266, "provider_group": "provider_y", "quality": 0.9199, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 15.005, "id": "worker_c", "p95_latency_ms": 54558, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.648, "max_latency_ms": 225152, "minimum_quality": 0.8316}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000528", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.2, "estimated_latency_ms": 201460, "estimated_pass_rate": 0.7674, "estimated_quality": 0.9225, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.648, "max_latency_ms": 225152, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7674, "quality": 0.9225, "quality_target_met": true, "utility": 77.040175}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.648,\"max_latency_ms\":225152,\"minimum_quality\":0.8316},\"task\":{\"ambiguity\":0.5791,\"category\":\"review\",\"complexity\":0.5512,\"consequence\":0.8529,\"evidence_need\":0.5894,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.4068,\"risk\":0.8326,\"summary\":\"Review whether provider failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.865,\"id\":\"worker_a\",\"p95_latency_ms\":27987,\"pass_rate\":0.557,\"provider_group\":\"provider_x\",\"quality\":0.7619,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.326,\"id\":\"worker_b\",\"p95_latency_ms\":91633,\"pass_rate\":0.9241,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.009,\"id\":\"worker_c\",\"p95_latency_ms\":81840,\"pass_rate\":0.814,\"provider_group\":\"provider_y\",\"quality\":0.8939,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.2,\"estimated_latency_ms\":201460,\"estimated_pass_rate\":0.7674,\"estimated_quality\":0.9225,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.648,\"max_latency_ms\":225152,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5791, "category": "review", "complexity": 0.5512, "consequence": 0.8529, "evidence_need": 0.5894, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-000528", "reversibility": 0.4068, "risk": 0.8326, "summary": "Review whether provider failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.865, "id": "worker_a", "p95_latency_ms": 27987, "pass_rate": 0.557, "provider_group": "provider_x", "quality": 0.7619, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.326, "id": "worker_b", "p95_latency_ms": 91633, "pass_rate": 0.9241, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.009, "id": "worker_c", "p95_latency_ms": 81840, "pass_rate": 0.814, "provider_group": "provider_y", "quality": 0.8939, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.654, "max_latency_ms": 56075, "minimum_quality": 0.6291}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000492", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.215, "estimated_latency_ms": 30262, "estimated_pass_rate": 0.4074, "estimated_quality": 0.6457, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.654, "max_latency_ms": 56075, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4074, "quality": 0.6457, "quality_target_met": true, "utility": 43.770488}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.654,\"max_latency_ms\":56075,\"minimum_quality\":0.6291},\"task\":{\"ambiguity\":0.494,\"category\":\"review\",\"complexity\":0.5445,\"consequence\":0.6194,\"evidence_need\":0.7262,\"family\":\"policy_compliance\",\"input_tokens\":6560,\"reversibility\":0.6071,\"risk\":0.6528,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.083,\"id\":\"worker_a\",\"p95_latency_ms\":82528,\"pass_rate\":0.7401,\"provider_group\":\"provider_y\",\"quality\":0.8428,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.215,\"id\":\"worker_b\",\"p95_latency_ms\":30262,\"pass_rate\":0.5594,\"provider_group\":\"provider_z\",\"quality\":0.7443,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.283,\"id\":\"worker_c\",\"p95_latency_ms\":77493,\"pass_rate\":0.9621,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.215,\"estimated_latency_ms\":30262,\"estimated_pass_rate\":0.4074,\"estimated_quality\":0.6457,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.654,\"max_latency_ms\":56075,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.494, "category": "review", "complexity": 0.5445, "consequence": 0.6194, "evidence_need": 0.7262, "family": "policy_compliance", "input_tokens": 6560, "request_id": "synthetic-policy_compliance-000492", "reversibility": 0.6071, "risk": 0.6528, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 4.083, "id": "worker_a", "p95_latency_ms": 82528, "pass_rate": 0.7401, "provider_group": "provider_y", "quality": 0.8428, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.215, "id": "worker_b", "p95_latency_ms": 30262, "pass_rate": 0.5594, "provider_group": "provider_z", "quality": 0.7443, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.283, "id": "worker_c", "p95_latency_ms": 77493, "pass_rate": 0.9621, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.563, "max_latency_ms": 27648, "minimum_quality": 0.6442}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000692", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.918, "estimated_latency_ms": 25759, "estimated_pass_rate": 0.2872, "estimated_quality": 0.6307, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.563, "max_latency_ms": 27648, "policy": "lean", "reason_codes": ["best_feasible_below_target", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2872, "quality": 0.6307, "quality_target_met": false, "utility": 38.442236}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.563,\"max_latency_ms\":27648,\"minimum_quality\":0.6442},\"task\":{\"ambiguity\":0.616,\"category\":\"analysis\",\"complexity\":0.6259,\"consequence\":0.3256,\"evidence_need\":0.4442,\"family\":\"performance_tuning\",\"input_tokens\":2480,\"reversibility\":0.669,\"risk\":0.2698,\"summary\":\"Diagnose latency variance in database failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.918,\"id\":\"worker_a\",\"p95_latency_ms\":25759,\"pass_rate\":0.4266,\"provider_group\":\"provider_z\",\"quality\":0.7247,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.78,\"id\":\"worker_b\",\"p95_latency_ms\":72314,\"pass_rate\":0.8015,\"provider_group\":\"provider_y\",\"quality\":0.9608,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":9.918,\"id\":\"worker_c\",\"p95_latency_ms\":57389,\"pass_rate\":0.9706,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.918,\"estimated_latency_ms\":25759,\"estimated_pass_rate\":0.2872,\"estimated_quality\":0.6307,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.563,\"max_latency_ms\":27648,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.616, "category": "analysis", "complexity": 0.6259, "consequence": 0.3256, "evidence_need": 0.4442, "family": "performance_tuning", "input_tokens": 2480, "request_id": "synthetic-performance_tuning-000692", "reversibility": 0.669, "risk": 0.2698, "summary": "Diagnose latency variance in database failover and rank the next experiments."}, "workers": [{"credits_per_call": 0.918, "id": "worker_a", "p95_latency_ms": 25759, "pass_rate": 0.4266, "provider_group": "provider_z", "quality": 0.7247, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.78, "id": "worker_b", "p95_latency_ms": 72314, "pass_rate": 0.8015, "provider_group": "provider_y", "quality": 0.9608, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 9.918, "id": "worker_c", "p95_latency_ms": 57389, "pass_rate": 0.9706, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.198, "max_latency_ms": 199400, "minimum_quality": 0.8498}, "family": "migration_planning", "id": "synthetic-migration_planning-000904", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.816, "estimated_latency_ms": 144488, "estimated_pass_rate": 0.708, "estimated_quality": 0.8806, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.198, "max_latency_ms": 199400, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.708, "quality": 0.8806, "quality_target_met": true, "utility": 75.5972}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.198,\"max_latency_ms\":199400,\"minimum_quality\":0.8498},\"task\":{\"ambiguity\":0.6344,\"category\":\"coding\",\"complexity\":0.8629,\"consequence\":0.7245,\"evidence_need\":0.7812,\"family\":\"migration_planning\",\"input_tokens\":2680,\"reversibility\":0.3889,\"risk\":0.6201,\"summary\":\"Review a staged migration for observability and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.587,\"id\":\"worker_a\",\"p95_latency_ms\":66189,\"pass_rate\":0.7377,\"provider_group\":\"provider_y\",\"quality\":0.8562,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.87,\"id\":\"worker_b\",\"p95_latency_ms\":20699,\"pass_rate\":0.5487,\"provider_group\":\"provider_x\",\"quality\":0.6836,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.359,\"id\":\"worker_c\",\"p95_latency_ms\":57600,\"pass_rate\":0.9083,\"provider_group\":\"provider_z\",\"quality\":0.959,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.816,\"estimated_latency_ms\":144488,\"estimated_pass_rate\":0.708,\"estimated_quality\":0.8806,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.198,\"max_latency_ms\":199400,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6344, "category": "coding", "complexity": 0.8629, "consequence": 0.7245, "evidence_need": 0.7812, "family": "migration_planning", "input_tokens": 2680, "request_id": "synthetic-migration_planning-000904", "reversibility": 0.3889, "risk": 0.6201, "summary": "Review a staged migration for observability and identify irreversible steps."}, "workers": [{"credits_per_call": 3.587, "id": "worker_a", "p95_latency_ms": 66189, "pass_rate": 0.7377, "provider_group": "provider_y", "quality": 0.8562, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.87, "id": "worker_b", "p95_latency_ms": 20699, "pass_rate": 0.5487, "provider_group": "provider_x", "quality": 0.6836, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.359, "id": "worker_c", "p95_latency_ms": 57600, "pass_rate": 0.9083, "provider_group": "provider_z", "quality": 0.959, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.281, "max_latency_ms": 219497, "minimum_quality": 0.9191}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000495", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.07, "estimated_latency_ms": 159538, "estimated_pass_rate": 0.762, "estimated_quality": 0.9271, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.281, "max_latency_ms": 219497, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.762, "quality": 0.9271, "quality_target_met": true, "utility": 79.04916}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.281,\"max_latency_ms\":219497,\"minimum_quality\":0.9191},\"task\":{\"ambiguity\":0.5135,\"category\":\"review\",\"complexity\":0.7331,\"consequence\":0.6732,\"evidence_need\":0.8558,\"family\":\"policy_compliance\",\"input_tokens\":1720,\"reversibility\":0.5403,\"risk\":0.7239,\"summary\":\"Review whether webhook delivery satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.315,\"id\":\"worker_a\",\"p95_latency_ms\":60882,\"pass_rate\":0.7987,\"provider_group\":\"provider_z\",\"quality\":0.8827,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.717,\"id\":\"worker_b\",\"p95_latency_ms\":69448,\"pass_rate\":0.9185,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.038,\"id\":\"worker_c\",\"p95_latency_ms\":29208,\"pass_rate\":0.5985,\"provider_group\":\"provider_x\",\"quality\":0.7702,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.07,\"estimated_latency_ms\":159538,\"estimated_pass_rate\":0.762,\"estimated_quality\":0.9271,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.281,\"max_latency_ms\":219497,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5135, "category": "review", "complexity": 0.7331, "consequence": 0.6732, "evidence_need": 0.8558, "family": "policy_compliance", "input_tokens": 1720, "request_id": "synthetic-policy_compliance-000495", "reversibility": 0.5403, "risk": 0.7239, "summary": "Review whether webhook delivery satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.315, "id": "worker_a", "p95_latency_ms": 60882, "pass_rate": 0.7987, "provider_group": "provider_z", "quality": 0.8827, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.717, "id": "worker_b", "p95_latency_ms": 69448, "pass_rate": 0.9185, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.038, "id": "worker_c", "p95_latency_ms": 29208, "pass_rate": 0.5985, "provider_group": "provider_x", "quality": 0.7702, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.447, "max_latency_ms": 159972, "minimum_quality": 0.7652}, "family": "migration_planning", "id": "synthetic-migration_planning-000829", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.805, "estimated_latency_ms": 75810, "estimated_pass_rate": 0.6076, "estimated_quality": 0.804, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.447, "max_latency_ms": 159972, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6076, "quality": 0.804, "quality_target_met": true, "utility": 61.865013}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.447,\"max_latency_ms\":159972,\"minimum_quality\":0.7652},\"task\":{\"ambiguity\":0.5313,\"category\":\"coding\",\"complexity\":0.8658,\"consequence\":0.5685,\"evidence_need\":0.6142,\"family\":\"migration_planning\",\"input_tokens\":4040,\"reversibility\":0.2543,\"risk\":0.7493,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.036,\"id\":\"worker_a\",\"p95_latency_ms\":25922,\"pass_rate\":0.4504,\"provider_group\":\"provider_x\",\"quality\":0.7413,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.1,\"id\":\"worker_b\",\"p95_latency_ms\":54506,\"pass_rate\":0.9245,\"provider_group\":\"provider_y\",\"quality\":0.9216,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.769,\"id\":\"worker_c\",\"p95_latency_ms\":49888,\"pass_rate\":0.8716,\"provider_group\":\"provider_z\",\"quality\":0.8885,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.805,\"estimated_latency_ms\":75810,\"estimated_pass_rate\":0.6076,\"estimated_quality\":0.804,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.447,\"max_latency_ms\":159972,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5313, "category": "coding", "complexity": 0.8658, "consequence": 0.5685, "evidence_need": 0.6142, "family": "migration_planning", "input_tokens": 4040, "request_id": "synthetic-migration_planning-000829", "reversibility": 0.2543, "risk": 0.7493, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 1.036, "id": "worker_a", "p95_latency_ms": 25922, "pass_rate": 0.4504, "provider_group": "provider_x", "quality": 0.7413, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.1, "id": "worker_b", "p95_latency_ms": 54506, "pass_rate": 0.9245, "provider_group": "provider_y", "quality": 0.9216, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.769, "id": "worker_c", "p95_latency_ms": 49888, "pass_rate": 0.8716, "provider_group": "provider_z", "quality": 0.8885, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.983, "max_latency_ms": 257331, "minimum_quality": 0.8656}, "family": "migration_planning", "id": "synthetic-migration_planning-000136", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.694, "estimated_latency_ms": 189788, "estimated_pass_rate": 0.7685, "estimated_quality": 0.9133, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.983, "max_latency_ms": 257331, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7685, "quality": 0.9133, "quality_target_met": true, "utility": 74.85421}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.983,\"max_latency_ms\":257331,\"minimum_quality\":0.8656},\"task\":{\"ambiguity\":0.3717,\"category\":\"coding\",\"complexity\":0.6763,\"consequence\":0.7315,\"evidence_need\":0.6534,\"family\":\"migration_planning\",\"input_tokens\":3360,\"reversibility\":0.5062,\"risk\":0.7188,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":10.936,\"id\":\"worker_a\",\"p95_latency_ms\":82079,\"pass_rate\":0.9349,\"provider_group\":\"provider_x\",\"quality\":0.9775,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.025,\"id\":\"worker_b\",\"p95_latency_ms\":25018,\"pass_rate\":0.56,\"provider_group\":\"provider_y\",\"quality\":0.7353,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.733,\"id\":\"worker_c\",\"p95_latency_ms\":82691,\"pass_rate\":0.8247,\"provider_group\":\"provider_z\",\"quality\":0.8979,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.694,\"estimated_latency_ms\":189788,\"estimated_pass_rate\":0.7685,\"estimated_quality\":0.9133,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.983,\"max_latency_ms\":257331,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3717, "category": "coding", "complexity": 0.6763, "consequence": 0.7315, "evidence_need": 0.6534, "family": "migration_planning", "input_tokens": 3360, "request_id": "synthetic-migration_planning-000136", "reversibility": 0.5062, "risk": 0.7188, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 10.936, "id": "worker_a", "p95_latency_ms": 82079, "pass_rate": 0.9349, "provider_group": "provider_x", "quality": 0.9775, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.025, "id": "worker_b", "p95_latency_ms": 25018, "pass_rate": 0.56, "provider_group": "provider_y", "quality": 0.7353, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.733, "id": "worker_c", "p95_latency_ms": 82691, "pass_rate": 0.8247, "provider_group": "provider_z", "quality": 0.8979, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.235, "max_latency_ms": 40454, "minimum_quality": 0.6409}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000563", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.195, "estimated_latency_ms": 25918, "estimated_pass_rate": 0.3442, "estimated_quality": 0.5197, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.235, "max_latency_ms": 40454, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3442, "quality": 0.5197, "quality_target_met": false, "utility": 28.421818}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.235,\"max_latency_ms\":40454,\"minimum_quality\":0.6409},\"task\":{\"ambiguity\":0.5457,\"category\":\"analysis\",\"complexity\":0.7277,\"consequence\":0.3566,\"evidence_need\":0.6265,\"family\":\"performance_tuning\",\"input_tokens\":5320,\"reversibility\":0.5789,\"risk\":0.4082,\"summary\":\"Review a proposed optimization for access control for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.195,\"id\":\"worker_a\",\"p95_latency_ms\":25918,\"pass_rate\":0.501,\"provider_group\":\"provider_x\",\"quality\":0.6373,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.103,\"id\":\"worker_b\",\"p95_latency_ms\":84183,\"pass_rate\":0.9943,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.811,\"id\":\"worker_c\",\"p95_latency_ms\":69562,\"pass_rate\":0.9374,\"provider_group\":\"provider_y\",\"quality\":0.9402,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.195,\"estimated_latency_ms\":25918,\"estimated_pass_rate\":0.3442,\"estimated_quality\":0.5197,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.235,\"max_latency_ms\":40454,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5457, "category": "analysis", "complexity": 0.7277, "consequence": 0.3566, "evidence_need": 0.6265, "family": "performance_tuning", "input_tokens": 5320, "request_id": "synthetic-performance_tuning-000563", "reversibility": 0.5789, "risk": 0.4082, "summary": "Review a proposed optimization for access control for hidden regressions."}, "workers": [{"credits_per_call": 1.195, "id": "worker_a", "p95_latency_ms": 25918, "pass_rate": 0.501, "provider_group": "provider_x", "quality": 0.6373, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.103, "id": "worker_b", "p95_latency_ms": 84183, "pass_rate": 0.9943, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.811, "id": "worker_c", "p95_latency_ms": 69562, "pass_rate": 0.9374, "provider_group": "provider_y", "quality": 0.9402, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.722, "max_latency_ms": 25099, "minimum_quality": 0.7236}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000735", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.152, "estimated_latency_ms": 21746, "estimated_pass_rate": 0.321, "estimated_quality": 0.6453, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.722, "max_latency_ms": 25099, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.321, "quality": 0.6453, "quality_target_met": false, "utility": 22.523985}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.722,\"max_latency_ms\":25099,\"minimum_quality\":0.7236},\"task\":{\"ambiguity\":0.7554,\"category\":\"review\",\"complexity\":0.5602,\"consequence\":0.8507,\"evidence_need\":0.5559,\"family\":\"policy_compliance\",\"input_tokens\":3960,\"reversibility\":0.6837,\"risk\":0.7722,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.152,\"id\":\"worker_a\",\"p95_latency_ms\":21746,\"pass_rate\":0.5042,\"provider_group\":\"provider_x\",\"quality\":0.7729,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.791,\"id\":\"worker_b\",\"p95_latency_ms\":59572,\"pass_rate\":0.9868,\"provider_group\":\"provider_z\",\"quality\":0.9187,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.132,\"id\":\"worker_c\",\"p95_latency_ms\":53331,\"pass_rate\":0.7521,\"provider_group\":\"provider_y\",\"quality\":0.9056,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.152,\"estimated_latency_ms\":21746,\"estimated_pass_rate\":0.321,\"estimated_quality\":0.6453,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.722,\"max_latency_ms\":25099,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7554, "category": "review", "complexity": 0.5602, "consequence": 0.8507, "evidence_need": 0.5559, "family": "policy_compliance", "input_tokens": 3960, "request_id": "synthetic-policy_compliance-000735", "reversibility": 0.6837, "risk": 0.7722, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.152, "id": "worker_a", "p95_latency_ms": 21746, "pass_rate": 0.5042, "provider_group": "provider_x", "quality": 0.7729, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.791, "id": "worker_b", "p95_latency_ms": 59572, "pass_rate": 0.9868, "provider_group": "provider_z", "quality": 0.9187, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.132, "id": "worker_c", "p95_latency_ms": 53331, "pass_rate": 0.7521, "provider_group": "provider_y", "quality": 0.9056, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 6.556, "max_latency_ms": 104227, "minimum_quality": 0.7729}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000147", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.88, "estimated_latency_ms": 68754, "estimated_pass_rate": 0.6821, "estimated_quality": 0.7999, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.556, "max_latency_ms": 104227, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6821, "quality": 0.7999, "quality_target_met": true, "utility": 61.516451}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":6.556,\"max_latency_ms\":104227,\"minimum_quality\":0.7729},\"task\":{\"ambiguity\":0.7472,\"category\":\"review\",\"complexity\":0.5686,\"consequence\":0.6565,\"evidence_need\":0.752,\"family\":\"policy_compliance\",\"input_tokens\":120,\"reversibility\":0.58,\"risk\":0.8449,\"summary\":\"Review whether OAuth callbacks satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.285,\"id\":\"worker_a\",\"p95_latency_ms\":60288,\"pass_rate\":0.9312,\"provider_group\":\"provider_z\",\"quality\":0.9608,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.799,\"id\":\"worker_b\",\"p95_latency_ms\":22119,\"pass_rate\":0.5835,\"provider_group\":\"provider_x\",\"quality\":0.6755,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.081,\"id\":\"worker_c\",\"p95_latency_ms\":46635,\"pass_rate\":0.8717,\"provider_group\":\"provider_y\",\"quality\":0.8441,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.88,\"estimated_latency_ms\":68754,\"estimated_pass_rate\":0.6821,\"estimated_quality\":0.7999,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.556,\"max_latency_ms\":104227,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7472, "category": "review", "complexity": 0.5686, "consequence": 0.6565, "evidence_need": 0.752, "family": "policy_compliance", "input_tokens": 120, "request_id": "synthetic-policy_compliance-000147", "reversibility": 0.58, "risk": 0.8449, "summary": "Review whether OAuth callbacks satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.285, "id": "worker_a", "p95_latency_ms": 60288, "pass_rate": 0.9312, "provider_group": "provider_z", "quality": 0.9608, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.799, "id": "worker_b", "p95_latency_ms": 22119, "pass_rate": 0.5835, "provider_group": "provider_x", "quality": 0.6755, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.081, "id": "worker_c", "p95_latency_ms": 46635, "pass_rate": 0.8717, "provider_group": "provider_y", "quality": 0.8441, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.583, "max_latency_ms": 210666, "minimum_quality": 0.8313}, "family": "migration_planning", "id": "synthetic-migration_planning-001015", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.591, "estimated_latency_ms": 102842, "estimated_pass_rate": 0.8076, "estimated_quality": 0.8923, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.583, "max_latency_ms": 210666, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8076, "quality": 0.8923, "quality_target_met": true, "utility": 63.649235}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.583,\"max_latency_ms\":210666,\"minimum_quality\":0.8313},\"task\":{\"ambiguity\":0.4445,\"category\":\"coding\",\"complexity\":0.7661,\"consequence\":0.4867,\"evidence_need\":0.5696,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.3149,\"risk\":0.6309,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.413,\"id\":\"worker_a\",\"p95_latency_ms\":63720,\"pass_rate\":0.8392,\"provider_group\":\"provider_z\",\"quality\":0.8674,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.721,\"id\":\"worker_b\",\"p95_latency_ms\":74014,\"pass_rate\":0.9181,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.87,\"id\":\"worker_c\",\"p95_latency_ms\":28828,\"pass_rate\":0.5902,\"provider_group\":\"provider_x\",\"quality\":0.7596,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.591,\"estimated_latency_ms\":102842,\"estimated_pass_rate\":0.8076,\"estimated_quality\":0.8923,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.583,\"max_latency_ms\":210666,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4445, "category": "coding", "complexity": 0.7661, "consequence": 0.4867, "evidence_need": 0.5696, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-001015", "reversibility": 0.3149, "risk": 0.6309, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 3.413, "id": "worker_a", "p95_latency_ms": 63720, "pass_rate": 0.8392, "provider_group": "provider_z", "quality": 0.8674, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.721, "id": "worker_b", "p95_latency_ms": 74014, "pass_rate": 0.9181, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.87, "id": "worker_c", "p95_latency_ms": 28828, "pass_rate": 0.5902, "provider_group": "provider_x", "quality": 0.7596, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.908, "max_latency_ms": 42001, "minimum_quality": 0.5276}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000390", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.966, "estimated_latency_ms": 24210, "estimated_pass_rate": 0.4184, "estimated_quality": 0.6718, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.908, "max_latency_ms": 42001, "policy": "lean", "reason_codes": ["high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4184, "quality": 0.6718, "quality_target_met": true, "utility": 34.174045}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.908,\"max_latency_ms\":42001,\"minimum_quality\":0.5276},\"task\":{\"ambiguity\":0.5449,\"category\":\"review\",\"complexity\":0.5704,\"consequence\":0.7618,\"evidence_need\":0.5712,\"family\":\"policy_compliance\",\"input_tokens\":4640,\"reversibility\":0.4058,\"risk\":0.7422,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":10.816,\"id\":\"worker_a\",\"p95_latency_ms\":77472,\"pass_rate\":0.9858,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.966,\"id\":\"worker_b\",\"p95_latency_ms\":24210,\"pass_rate\":0.5844,\"provider_group\":\"provider_z\",\"quality\":0.785,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.057,\"id\":\"worker_c\",\"p95_latency_ms\":60525,\"pass_rate\":0.8115,\"provider_group\":\"provider_x\",\"quality\":0.8899,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.966,\"estimated_latency_ms\":24210,\"estimated_pass_rate\":0.4184,\"estimated_quality\":0.6718,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.908,\"max_latency_ms\":42001,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5449, "category": "review", "complexity": 0.5704, "consequence": 0.7618, "evidence_need": 0.5712, "family": "policy_compliance", "input_tokens": 4640, "request_id": "synthetic-policy_compliance-000390", "reversibility": 0.4058, "risk": 0.7422, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 10.816, "id": "worker_a", "p95_latency_ms": 77472, "pass_rate": 0.9858, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.966, "id": "worker_b", "p95_latency_ms": 24210, "pass_rate": 0.5844, "provider_group": "provider_z", "quality": 0.785, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.057, "id": "worker_c", "p95_latency_ms": 60525, "pass_rate": 0.8115, "provider_group": "provider_x", "quality": 0.8899, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.827, "max_latency_ms": 35514, "minimum_quality": 0.6553}, "family": "migration_planning", "id": "synthetic-migration_planning-000637", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.022, "estimated_latency_ms": 32269, "estimated_pass_rate": 0.2935, "estimated_quality": 0.5631, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.827, "max_latency_ms": 35514, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2935, "quality": 0.5631, "quality_target_met": false, "utility": 29.785767}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.827,\"max_latency_ms\":35514,\"minimum_quality\":0.6553},\"task\":{\"ambiguity\":0.4242,\"category\":\"coding\",\"complexity\":0.9013,\"consequence\":0.6008,\"evidence_need\":0.6114,\"family\":\"migration_planning\",\"input_tokens\":3000,\"reversibility\":0.5094,\"risk\":0.4603,\"summary\":\"Review a staged migration for streaming inference and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":12.57,\"id\":\"worker_a\",\"p95_latency_ms\":66186,\"pass_rate\":0.9106,\"provider_group\":\"provider_z\",\"quality\":0.9321,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.165,\"id\":\"worker_b\",\"p95_latency_ms\":80411,\"pass_rate\":0.8447,\"provider_group\":\"provider_y\",\"quality\":0.9071,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.022,\"id\":\"worker_c\",\"p95_latency_ms\":32269,\"pass_rate\":0.4651,\"provider_group\":\"provider_x\",\"quality\":0.7077,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.022,\"estimated_latency_ms\":32269,\"estimated_pass_rate\":0.2935,\"estimated_quality\":0.5631,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.827,\"max_latency_ms\":35514,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4242, "category": "coding", "complexity": 0.9013, "consequence": 0.6008, "evidence_need": 0.6114, "family": "migration_planning", "input_tokens": 3000, "request_id": "synthetic-migration_planning-000637", "reversibility": 0.5094, "risk": 0.4603, "summary": "Review a staged migration for streaming inference and identify irreversible steps."}, "workers": [{"credits_per_call": 12.57, "id": "worker_a", "p95_latency_ms": 66186, "pass_rate": 0.9106, "provider_group": "provider_z", "quality": 0.9321, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.165, "id": "worker_b", "p95_latency_ms": 80411, "pass_rate": 0.8447, "provider_group": "provider_y", "quality": 0.9071, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.022, "id": "worker_c", "p95_latency_ms": 32269, "pass_rate": 0.4651, "provider_group": "provider_x", "quality": 0.7077, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.641, "max_latency_ms": 187716, "minimum_quality": 0.9084}, "family": "migration_planning", "id": "synthetic-migration_planning-000211", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.014, "estimated_latency_ms": 143817, "estimated_pass_rate": 0.7135, "estimated_quality": 0.921, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.641, "max_latency_ms": 187716, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7135, "quality": 0.921, "quality_target_met": true, "utility": 71.104709}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.641,\"max_latency_ms\":187716,\"minimum_quality\":0.9084},\"task\":{\"ambiguity\":0.3375,\"category\":\"coding\",\"complexity\":0.8854,\"consequence\":0.6022,\"evidence_need\":0.5449,\"family\":\"migration_planning\",\"input_tokens\":6760,\"reversibility\":0.4861,\"risk\":0.421,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.447,\"id\":\"worker_a\",\"p95_latency_ms\":50640,\"pass_rate\":0.9367,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.98,\"id\":\"worker_b\",\"p95_latency_ms\":23293,\"pass_rate\":0.492,\"provider_group\":\"provider_z\",\"quality\":0.778,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.587,\"id\":\"worker_c\",\"p95_latency_ms\":69884,\"pass_rate\":0.8662,\"provider_group\":\"provider_y\",\"quality\":0.8416,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.014,\"estimated_latency_ms\":143817,\"estimated_pass_rate\":0.7135,\"estimated_quality\":0.921,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.641,\"max_latency_ms\":187716,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3375, "category": "coding", "complexity": 0.8854, "consequence": 0.6022, "evidence_need": 0.5449, "family": "migration_planning", "input_tokens": 6760, "request_id": "synthetic-migration_planning-000211", "reversibility": 0.4861, "risk": 0.421, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 11.447, "id": "worker_a", "p95_latency_ms": 50640, "pass_rate": 0.9367, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.98, "id": "worker_b", "p95_latency_ms": 23293, "pass_rate": 0.492, "provider_group": "provider_z", "quality": 0.778, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.587, "id": "worker_c", "p95_latency_ms": 69884, "pass_rate": 0.8662, "provider_group": "provider_y", "quality": 0.8416, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.817, "max_latency_ms": 205407, "minimum_quality": 0.85}, "family": "migration_planning", "id": "synthetic-migration_planning-001060", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.303, "estimated_latency_ms": 151941, "estimated_pass_rate": 0.7598, "estimated_quality": 0.8722, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.817, "max_latency_ms": 205407, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7598, "quality": 0.8722, "quality_target_met": true, "utility": 76.23579}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.817,\"max_latency_ms\":205407,\"minimum_quality\":0.85},\"task\":{\"ambiguity\":0.6911,\"category\":\"coding\",\"complexity\":0.897,\"consequence\":0.7446,\"evidence_need\":0.4554,\"family\":\"migration_planning\",\"input_tokens\":5360,\"reversibility\":0.5361,\"risk\":0.7277,\"summary\":\"Design compatibility testing for a credit reservation schema migration.\"},\"workers\":[{\"credits_per_call\":3.294,\"id\":\"worker_a\",\"p95_latency_ms\":58927,\"pass_rate\":0.8428,\"provider_group\":\"provider_y\",\"quality\":0.8215,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.19,\"id\":\"worker_b\",\"p95_latency_ms\":74799,\"pass_rate\":0.9417,\"provider_group\":\"provider_z\",\"quality\":0.9203,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.819,\"id\":\"worker_c\",\"p95_latency_ms\":18215,\"pass_rate\":0.6073,\"provider_group\":\"provider_x\",\"quality\":0.773,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.303,\"estimated_latency_ms\":151941,\"estimated_pass_rate\":0.7598,\"estimated_quality\":0.8722,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.817,\"max_latency_ms\":205407,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6911, "category": "coding", "complexity": 0.897, "consequence": 0.7446, "evidence_need": 0.4554, "family": "migration_planning", "input_tokens": 5360, "request_id": "synthetic-migration_planning-001060", "reversibility": 0.5361, "risk": 0.7277, "summary": "Design compatibility testing for a credit reservation schema migration."}, "workers": [{"credits_per_call": 3.294, "id": "worker_a", "p95_latency_ms": 58927, "pass_rate": 0.8428, "provider_group": "provider_y", "quality": 0.8215, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.19, "id": "worker_b", "p95_latency_ms": 74799, "pass_rate": 0.9417, "provider_group": "provider_z", "quality": 0.9203, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.819, "id": "worker_c", "p95_latency_ms": 18215, "pass_rate": 0.6073, "provider_group": "provider_x", "quality": 0.773, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.519, "max_latency_ms": 136704, "minimum_quality": 0.79}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001197", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.337, "estimated_latency_ms": 99051, "estimated_pass_rate": 0.5732, "estimated_quality": 0.8129, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.519, "max_latency_ms": 136704, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5732, "quality": 0.8129, "quality_target_met": true, "utility": 61.875442}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.519,\"max_latency_ms\":136704,\"minimum_quality\":0.79},\"task\":{\"ambiguity\":0.4727,\"category\":\"review\",\"complexity\":0.4893,\"consequence\":0.9156,\"evidence_need\":0.6521,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.4341,\"risk\":0.6138,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.36,\"id\":\"worker_a\",\"p95_latency_ms\":67146,\"pass_rate\":0.7456,\"provider_group\":\"provider_x\",\"quality\":0.8824,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.977,\"id\":\"worker_b\",\"p95_latency_ms\":31905,\"pass_rate\":0.4599,\"provider_group\":\"provider_z\",\"quality\":0.7479,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.462,\"id\":\"worker_c\",\"p95_latency_ms\":77854,\"pass_rate\":0.9147,\"provider_group\":\"provider_y\",\"quality\":0.9154,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.337,\"estimated_latency_ms\":99051,\"estimated_pass_rate\":0.5732,\"estimated_quality\":0.8129,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.519,\"max_latency_ms\":136704,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4727, "category": "review", "complexity": 0.4893, "consequence": 0.9156, "evidence_need": 0.6521, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-001197", "reversibility": 0.4341, "risk": 0.6138, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 3.36, "id": "worker_a", "p95_latency_ms": 67146, "pass_rate": 0.7456, "provider_group": "provider_x", "quality": 0.8824, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.977, "id": "worker_b", "p95_latency_ms": 31905, "pass_rate": 0.4599, "provider_group": "provider_z", "quality": 0.7479, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.462, "id": "worker_c", "p95_latency_ms": 77854, "pass_rate": 0.9147, "provider_group": "provider_y", "quality": 0.9154, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.154, "max_latency_ms": 35365, "minimum_quality": 0.67}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000186", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.088, "estimated_latency_ms": 20161, "estimated_pass_rate": 0.4916, "estimated_quality": 0.7083, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.154, "max_latency_ms": 35365, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4916, "quality": 0.7083, "quality_target_met": true, "utility": 50.241313}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.154,\"max_latency_ms\":35365,\"minimum_quality\":0.67},\"task\":{\"ambiguity\":0.3645,\"category\":\"review\",\"complexity\":0.456,\"consequence\":0.8043,\"evidence_need\":0.892,\"family\":\"policy_compliance\",\"input_tokens\":4560,\"reversibility\":0.3201,\"risk\":0.4729,\"summary\":\"Review whether schema evolution satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.761,\"id\":\"worker_a\",\"p95_latency_ms\":75824,\"pass_rate\":0.834,\"provider_group\":\"provider_x\",\"quality\":0.8927,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.088,\"id\":\"worker_b\",\"p95_latency_ms\":20161,\"pass_rate\":0.6092,\"provider_group\":\"provider_z\",\"quality\":0.7693,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.599,\"id\":\"worker_c\",\"p95_latency_ms\":87164,\"pass_rate\":0.994,\"provider_group\":\"provider_y\",\"quality\":0.9492,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.088,\"estimated_latency_ms\":20161,\"estimated_pass_rate\":0.4916,\"estimated_quality\":0.7083,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.154,\"max_latency_ms\":35365,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3645, "category": "review", "complexity": 0.456, "consequence": 0.8043, "evidence_need": 0.892, "family": "policy_compliance", "input_tokens": 4560, "request_id": "synthetic-policy_compliance-000186", "reversibility": 0.3201, "risk": 0.4729, "summary": "Review whether schema evolution satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.761, "id": "worker_a", "p95_latency_ms": 75824, "pass_rate": 0.834, "provider_group": "provider_x", "quality": 0.8927, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.088, "id": "worker_b", "p95_latency_ms": 20161, "pass_rate": 0.6092, "provider_group": "provider_z", "quality": 0.7693, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.599, "id": "worker_c", "p95_latency_ms": 87164, "pass_rate": 0.994, "provider_group": "provider_y", "quality": 0.9492, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.795, "max_latency_ms": 220962, "minimum_quality": 0.83}, "family": "migration_planning", "id": "synthetic-migration_planning-001009", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.659, "estimated_latency_ms": 109012, "estimated_pass_rate": 0.6368, "estimated_quality": 0.8186, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.795, "max_latency_ms": 220962, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6368, "quality": 0.8186, "quality_target_met": false, "utility": 64.126026}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.795,\"max_latency_ms\":220962,\"minimum_quality\":0.83},\"task\":{\"ambiguity\":0.6238,\"category\":\"coding\",\"complexity\":0.5978,\"consequence\":0.7374,\"evidence_need\":0.6811,\"family\":\"migration_planning\",\"input_tokens\":6240,\"reversibility\":0.2644,\"risk\":0.731,\"summary\":\"Design compatibility testing for a background jobs schema migration.\"},\"workers\":[{\"credits_per_call\":13.906,\"id\":\"worker_a\",\"p95_latency_ms\":62495,\"pass_rate\":0.9837,\"provider_group\":\"provider_x\",\"quality\":0.9741,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.809,\"id\":\"worker_b\",\"p95_latency_ms\":76886,\"pass_rate\":0.7765,\"provider_group\":\"provider_z\",\"quality\":0.8222,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.85,\"id\":\"worker_c\",\"p95_latency_ms\":32126,\"pass_rate\":0.5819,\"provider_group\":\"provider_y\",\"quality\":0.7932,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.659,\"estimated_latency_ms\":109012,\"estimated_pass_rate\":0.6368,\"estimated_quality\":0.8186,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.795,\"max_latency_ms\":220962,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6238, "category": "coding", "complexity": 0.5978, "consequence": 0.7374, "evidence_need": 0.6811, "family": "migration_planning", "input_tokens": 6240, "request_id": "synthetic-migration_planning-001009", "reversibility": 0.2644, "risk": 0.731, "summary": "Design compatibility testing for a background jobs schema migration."}, "workers": [{"credits_per_call": 13.906, "id": "worker_a", "p95_latency_ms": 62495, "pass_rate": 0.9837, "provider_group": "provider_x", "quality": 0.9741, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.809, "id": "worker_b", "p95_latency_ms": 76886, "pass_rate": 0.7765, "provider_group": "provider_z", "quality": 0.8222, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.85, "id": "worker_c", "p95_latency_ms": 32126, "pass_rate": 0.5819, "provider_group": "provider_y", "quality": 0.7932, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.824, "max_latency_ms": 228436, "minimum_quality": 0.9468}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001239", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.08, "estimated_latency_ms": 180384, "estimated_pass_rate": 0.7169, "estimated_quality": 0.9086, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.824, "max_latency_ms": 228436, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7169, "quality": 0.9086, "quality_target_met": false, "utility": 72.916772}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.824,\"max_latency_ms\":228436,\"minimum_quality\":0.9468},\"task\":{\"ambiguity\":0.5392,\"category\":\"review\",\"complexity\":0.7281,\"consequence\":0.7327,\"evidence_need\":0.6315,\"family\":\"policy_compliance\",\"input_tokens\":6320,\"reversibility\":0.3918,\"risk\":0.6382,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.787,\"id\":\"worker_a\",\"p95_latency_ms\":27602,\"pass_rate\":0.4907,\"provider_group\":\"provider_x\",\"quality\":0.7823,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.05,\"id\":\"worker_b\",\"p95_latency_ms\":61573,\"pass_rate\":0.7934,\"provider_group\":\"provider_z\",\"quality\":0.8536,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.243,\"id\":\"worker_c\",\"p95_latency_ms\":91209,\"pass_rate\":0.9362,\"provider_group\":\"provider_y\",\"quality\":0.9694,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.08,\"estimated_latency_ms\":180384,\"estimated_pass_rate\":0.7169,\"estimated_quality\":0.9086,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.824,\"max_latency_ms\":228436,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5392, "category": "review", "complexity": 0.7281, "consequence": 0.7327, "evidence_need": 0.6315, "family": "policy_compliance", "input_tokens": 6320, "request_id": "synthetic-policy_compliance-001239", "reversibility": 0.3918, "risk": 0.6382, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.787, "id": "worker_a", "p95_latency_ms": 27602, "pass_rate": 0.4907, "provider_group": "provider_x", "quality": 0.7823, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.05, "id": "worker_b", "p95_latency_ms": 61573, "pass_rate": 0.7934, "provider_group": "provider_z", "quality": 0.8536, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.243, "id": "worker_c", "p95_latency_ms": 91209, "pass_rate": 0.9362, "provider_group": "provider_y", "quality": 0.9694, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.331, "max_latency_ms": 130167, "minimum_quality": 0.7803}, "family": "migration_planning", "id": "synthetic-migration_planning-000553", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.975, "estimated_latency_ms": 94968, "estimated_pass_rate": 0.5762, "estimated_quality": 0.7858, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.331, "max_latency_ms": 130167, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5762, "quality": 0.7858, "quality_target_met": true, "utility": 58.348898}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.331,\"max_latency_ms\":130167,\"minimum_quality\":0.7803},\"task\":{\"ambiguity\":0.6306,\"category\":\"coding\",\"complexity\":0.8988,\"consequence\":0.7055,\"evidence_need\":0.8226,\"family\":\"migration_planning\",\"input_tokens\":7760,\"reversibility\":0.3595,\"risk\":0.5816,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":13.042,\"id\":\"worker_a\",\"p95_latency_ms\":72897,\"pass_rate\":0.9234,\"provider_group\":\"provider_z\",\"quality\":0.9803,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.825,\"id\":\"worker_b\",\"p95_latency_ms\":63340,\"pass_rate\":0.8605,\"provider_group\":\"provider_y\",\"quality\":0.8404,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.15,\"id\":\"worker_c\",\"p95_latency_ms\":31628,\"pass_rate\":0.4872,\"provider_group\":\"provider_x\",\"quality\":0.7064,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.975,\"estimated_latency_ms\":94968,\"estimated_pass_rate\":0.5762,\"estimated_quality\":0.7858,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.331,\"max_latency_ms\":130167,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6306, "category": "coding", "complexity": 0.8988, "consequence": 0.7055, "evidence_need": 0.8226, "family": "migration_planning", "input_tokens": 7760, "request_id": "synthetic-migration_planning-000553", "reversibility": 0.3595, "risk": 0.5816, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 13.042, "id": "worker_a", "p95_latency_ms": 72897, "pass_rate": 0.9234, "provider_group": "provider_z", "quality": 0.9803, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.825, "id": "worker_b", "p95_latency_ms": 63340, "pass_rate": 0.8605, "provider_group": "provider_y", "quality": 0.8404, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.15, "id": "worker_c", "p95_latency_ms": 31628, "pass_rate": 0.4872, "provider_group": "provider_x", "quality": 0.7064, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.274, "max_latency_ms": 151147, "minimum_quality": 0.825}, "family": "migration_planning", "id": "synthetic-migration_planning-000454", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.123, "estimated_latency_ms": 137225, "estimated_pass_rate": 0.6835, "estimated_quality": 0.9143, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.274, "max_latency_ms": 151147, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6835, "quality": 0.9143, "quality_target_met": true, "utility": 75.690628}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.274,\"max_latency_ms\":151147,\"minimum_quality\":0.825},\"task\":{\"ambiguity\":0.7486,\"category\":\"coding\",\"complexity\":0.9301,\"consequence\":0.436,\"evidence_need\":0.7313,\"family\":\"migration_planning\",\"input_tokens\":3480,\"reversibility\":0.4787,\"risk\":0.5754,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.341,\"id\":\"worker_a\",\"p95_latency_ms\":51267,\"pass_rate\":0.8947,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.745,\"id\":\"worker_b\",\"p95_latency_ms\":60839,\"pass_rate\":0.8434,\"provider_group\":\"provider_z\",\"quality\":0.8118,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.037,\"id\":\"worker_c\",\"p95_latency_ms\":25119,\"pass_rate\":0.4898,\"provider_group\":\"provider_y\",\"quality\":0.7747,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.123,\"estimated_latency_ms\":137225,\"estimated_pass_rate\":0.6835,\"estimated_quality\":0.9143,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.274,\"max_latency_ms\":151147,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7486, "category": "coding", "complexity": 0.9301, "consequence": 0.436, "evidence_need": 0.7313, "family": "migration_planning", "input_tokens": 3480, "request_id": "synthetic-migration_planning-000454", "reversibility": 0.4787, "risk": 0.5754, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 14.341, "id": "worker_a", "p95_latency_ms": 51267, "pass_rate": 0.8947, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.745, "id": "worker_b", "p95_latency_ms": 60839, "pass_rate": 0.8434, "provider_group": "provider_z", "quality": 0.8118, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.037, "id": "worker_c", "p95_latency_ms": 25119, "pass_rate": 0.4898, "provider_group": "provider_y", "quality": 0.7747, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.258, "max_latency_ms": 52890, "minimum_quality": 0.5107}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000522", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.881, "estimated_latency_ms": 27411, "estimated_pass_rate": 0.3418, "estimated_quality": 0.6132, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.258, "max_latency_ms": 52890, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3418, "quality": 0.6132, "quality_target_met": true, "utility": 40.816433}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.258,\"max_latency_ms\":52890,\"minimum_quality\":0.5107},\"task\":{\"ambiguity\":0.7135,\"category\":\"review\",\"complexity\":0.4885,\"consequence\":0.7791,\"evidence_need\":0.5546,\"family\":\"policy_compliance\",\"input_tokens\":2560,\"reversibility\":0.4627,\"risk\":0.6368,\"summary\":\"Identify evidence required before approving credit reservation for production.\"},\"workers\":[{\"credits_per_call\":14.546,\"id\":\"worker_a\",\"p95_latency_ms\":56981,\"pass_rate\":0.9078,\"provider_group\":\"provider_y\",\"quality\":0.9711,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.881,\"id\":\"worker_b\",\"p95_latency_ms\":27411,\"pass_rate\":0.5027,\"provider_group\":\"provider_x\",\"quality\":0.7153,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.62,\"id\":\"worker_c\",\"p95_latency_ms\":53436,\"pass_rate\":0.8362,\"provider_group\":\"provider_z\",\"quality\":0.8528,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.881,\"estimated_latency_ms\":27411,\"estimated_pass_rate\":0.3418,\"estimated_quality\":0.6132,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.258,\"max_latency_ms\":52890,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7135, "category": "review", "complexity": 0.4885, "consequence": 0.7791, "evidence_need": 0.5546, "family": "policy_compliance", "input_tokens": 2560, "request_id": "synthetic-policy_compliance-000522", "reversibility": 0.4627, "risk": 0.6368, "summary": "Identify evidence required before approving credit reservation for production."}, "workers": [{"credits_per_call": 14.546, "id": "worker_a", "p95_latency_ms": 56981, "pass_rate": 0.9078, "provider_group": "provider_y", "quality": 0.9711, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.881, "id": "worker_b", "p95_latency_ms": 27411, "pass_rate": 0.5027, "provider_group": "provider_x", "quality": 0.7153, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.62, "id": "worker_c", "p95_latency_ms": 53436, "pass_rate": 0.8362, "provider_group": "provider_z", "quality": 0.8528, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.532, "max_latency_ms": 39742, "minimum_quality": 0.7284}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000792", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.864, "estimated_latency_ms": 24629, "estimated_pass_rate": 0.38, "estimated_quality": 0.621, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.532, "max_latency_ms": 39742, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.38, "quality": 0.621, "quality_target_met": false, "utility": 25.500056}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.532,\"max_latency_ms\":39742,\"minimum_quality\":0.7284},\"task\":{\"ambiguity\":0.5125,\"category\":\"review\",\"complexity\":0.7217,\"consequence\":0.7385,\"evidence_need\":0.8369,\"family\":\"policy_compliance\",\"input_tokens\":5200,\"reversibility\":0.3415,\"risk\":0.6291,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.864,\"id\":\"worker_a\",\"p95_latency_ms\":24629,\"pass_rate\":0.5515,\"provider_group\":\"provider_x\",\"quality\":0.7512,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.956,\"id\":\"worker_b\",\"p95_latency_ms\":66625,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9514,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.869,\"id\":\"worker_c\",\"p95_latency_ms\":56751,\"pass_rate\":0.7405,\"provider_group\":\"provider_z\",\"quality\":0.9123,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.864,\"estimated_latency_ms\":24629,\"estimated_pass_rate\":0.38,\"estimated_quality\":0.621,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.532,\"max_latency_ms\":39742,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5125, "category": "review", "complexity": 0.7217, "consequence": 0.7385, "evidence_need": 0.8369, "family": "policy_compliance", "input_tokens": 5200, "request_id": "synthetic-policy_compliance-000792", "reversibility": 0.3415, "risk": 0.6291, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.864, "id": "worker_a", "p95_latency_ms": 24629, "pass_rate": 0.5515, "provider_group": "provider_x", "quality": 0.7512, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.956, "id": "worker_b", "p95_latency_ms": 66625, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9514, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.869, "id": "worker_c", "p95_latency_ms": 56751, "pass_rate": 0.7405, "provider_group": "provider_z", "quality": 0.9123, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.955, "max_latency_ms": 111835, "minimum_quality": 0.748}, "family": "migration_planning", "id": "synthetic-migration_planning-000808", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.337, "estimated_latency_ms": 97240, "estimated_pass_rate": 0.6119, "estimated_quality": 0.7623, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.955, "max_latency_ms": 111835, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6119, "quality": 0.7623, "quality_target_met": true, "utility": 56.456396}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.955,\"max_latency_ms\":111835,\"minimum_quality\":0.748},\"task\":{\"ambiguity\":0.7286,\"category\":\"coding\",\"complexity\":0.8777,\"consequence\":0.8419,\"evidence_need\":0.6189,\"family\":\"migration_planning\",\"input_tokens\":2440,\"reversibility\":0.4345,\"risk\":0.5196,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":3.322,\"id\":\"worker_a\",\"p95_latency_ms\":74849,\"pass_rate\":0.7503,\"provider_group\":\"provider_z\",\"quality\":0.8615,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.119,\"id\":\"worker_b\",\"p95_latency_ms\":90254,\"pass_rate\":0.886,\"provider_group\":\"provider_x\",\"quality\":0.9669,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.015,\"id\":\"worker_c\",\"p95_latency_ms\":22391,\"pass_rate\":0.604,\"provider_group\":\"provider_y\",\"quality\":0.7582,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.337,\"estimated_latency_ms\":97240,\"estimated_pass_rate\":0.6119,\"estimated_quality\":0.7623,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.955,\"max_latency_ms\":111835,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7286, "category": "coding", "complexity": 0.8777, "consequence": 0.8419, "evidence_need": 0.6189, "family": "migration_planning", "input_tokens": 2440, "request_id": "synthetic-migration_planning-000808", "reversibility": 0.4345, "risk": 0.5196, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 3.322, "id": "worker_a", "p95_latency_ms": 74849, "pass_rate": 0.7503, "provider_group": "provider_z", "quality": 0.8615, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.119, "id": "worker_b", "p95_latency_ms": 90254, "pass_rate": 0.886, "provider_group": "provider_x", "quality": 0.9669, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.015, "id": "worker_c", "p95_latency_ms": 22391, "pass_rate": 0.604, "provider_group": "provider_y", "quality": 0.7582, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.406, "max_latency_ms": 131779, "minimum_quality": 0.7578}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000851", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.635, "estimated_latency_ms": 130227, "estimated_pass_rate": 0.9049, "estimated_quality": 0.9589, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.406, "max_latency_ms": 131779, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9049, "quality": 0.9589, "quality_target_met": true, "utility": 70.200901}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.406,\"max_latency_ms\":131779,\"minimum_quality\":0.7578},\"task\":{\"ambiguity\":0.3809,\"category\":\"analysis\",\"complexity\":0.5525,\"consequence\":0.4881,\"evidence_need\":0.5414,\"family\":\"performance_tuning\",\"input_tokens\":6080,\"reversibility\":0.6033,\"risk\":0.4754,\"summary\":\"Review a proposed optimization for private credentials for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.963,\"id\":\"worker_a\",\"p95_latency_ms\":80384,\"pass_rate\":0.9275,\"provider_group\":\"provider_x\",\"quality\":0.9151,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.672,\"id\":\"worker_b\",\"p95_latency_ms\":49843,\"pass_rate\":0.9598,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.97,\"id\":\"worker_c\",\"p95_latency_ms\":18496,\"pass_rate\":0.4605,\"provider_group\":\"provider_z\",\"quality\":0.7057,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.635,\"estimated_latency_ms\":130227,\"estimated_pass_rate\":0.9049,\"estimated_quality\":0.9589,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.406,\"max_latency_ms\":131779,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3809, "category": "analysis", "complexity": 0.5525, "consequence": 0.4881, "evidence_need": 0.5414, "family": "performance_tuning", "input_tokens": 6080, "request_id": "synthetic-performance_tuning-000851", "reversibility": 0.6033, "risk": 0.4754, "summary": "Review a proposed optimization for private credentials for hidden regressions."}, "workers": [{"credits_per_call": 3.963, "id": "worker_a", "p95_latency_ms": 80384, "pass_rate": 0.9275, "provider_group": "provider_x", "quality": 0.9151, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.672, "id": "worker_b", "p95_latency_ms": 49843, "pass_rate": 0.9598, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.97, "id": "worker_c", "p95_latency_ms": 18496, "pass_rate": 0.4605, "provider_group": "provider_z", "quality": 0.7057, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.583, "max_latency_ms": 28565, "minimum_quality": 0.5903}, "family": "migration_planning", "id": "synthetic-migration_planning-000589", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.869, "estimated_latency_ms": 18116, "estimated_pass_rate": 0.3061, "estimated_quality": 0.6617, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.583, "max_latency_ms": 28565, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3061, "quality": 0.6617, "quality_target_met": true, "utility": 43.514049}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.583,\"max_latency_ms\":28565,\"minimum_quality\":0.5903},\"task\":{\"ambiguity\":0.3951,\"category\":\"coding\",\"complexity\":0.6312,\"consequence\":0.6297,\"evidence_need\":0.5806,\"family\":\"migration_planning\",\"input_tokens\":5000,\"reversibility\":0.3819,\"risk\":0.6536,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":2.999,\"id\":\"worker_a\",\"p95_latency_ms\":75796,\"pass_rate\":0.7447,\"provider_group\":\"provider_z\",\"quality\":0.916,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.869,\"id\":\"worker_b\",\"p95_latency_ms\":18116,\"pass_rate\":0.4605,\"provider_group\":\"provider_x\",\"quality\":0.7694,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.013,\"id\":\"worker_c\",\"p95_latency_ms\":74333,\"pass_rate\":0.9326,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.869,\"estimated_latency_ms\":18116,\"estimated_pass_rate\":0.3061,\"estimated_quality\":0.6617,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.583,\"max_latency_ms\":28565,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3951, "category": "coding", "complexity": 0.6312, "consequence": 0.6297, "evidence_need": 0.5806, "family": "migration_planning", "input_tokens": 5000, "request_id": "synthetic-migration_planning-000589", "reversibility": 0.3819, "risk": 0.6536, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 2.999, "id": "worker_a", "p95_latency_ms": 75796, "pass_rate": 0.7447, "provider_group": "provider_z", "quality": 0.916, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.869, "id": "worker_b", "p95_latency_ms": 18116, "pass_rate": 0.4605, "provider_group": "provider_x", "quality": 0.7694, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.013, "id": "worker_c", "p95_latency_ms": 74333, "pass_rate": 0.9326, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.784, "max_latency_ms": 148255, "minimum_quality": 0.8041}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000713", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.779, "estimated_latency_ms": 144192, "estimated_pass_rate": 0.7167, "estimated_quality": 0.9109, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.784, "max_latency_ms": 148255, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7167, "quality": 0.9109, "quality_target_met": true, "utility": 68.593057}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.784,\"max_latency_ms\":148255,\"minimum_quality\":0.8041},\"task\":{\"ambiguity\":0.6628,\"category\":\"analysis\",\"complexity\":0.7664,\"consequence\":0.5621,\"evidence_need\":0.472,\"family\":\"performance_tuning\",\"input_tokens\":4240,\"reversibility\":0.6212,\"risk\":0.2724,\"summary\":\"Review a proposed optimization for tenant-scoped caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.066,\"id\":\"worker_a\",\"p95_latency_ms\":24231,\"pass_rate\":0.503,\"provider_group\":\"provider_z\",\"quality\":0.7115,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.03,\"id\":\"worker_b\",\"p95_latency_ms\":53094,\"pass_rate\":0.7958,\"provider_group\":\"provider_y\",\"quality\":0.9287,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.683,\"id\":\"worker_c\",\"p95_latency_ms\":66867,\"pass_rate\":0.9674,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.779,\"estimated_latency_ms\":144192,\"estimated_pass_rate\":0.7167,\"estimated_quality\":0.9109,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.784,\"max_latency_ms\":148255,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6628, "category": "analysis", "complexity": 0.7664, "consequence": 0.5621, "evidence_need": 0.472, "family": "performance_tuning", "input_tokens": 4240, "request_id": "synthetic-performance_tuning-000713", "reversibility": 0.6212, "risk": 0.2724, "summary": "Review a proposed optimization for tenant-scoped caching for hidden regressions."}, "workers": [{"credits_per_call": 1.066, "id": "worker_a", "p95_latency_ms": 24231, "pass_rate": 0.503, "provider_group": "provider_z", "quality": 0.7115, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.03, "id": "worker_b", "p95_latency_ms": 53094, "pass_rate": 0.7958, "provider_group": "provider_y", "quality": 0.9287, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.683, "id": "worker_c", "p95_latency_ms": 66867, "pass_rate": 0.9674, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.819, "max_latency_ms": 170154, "minimum_quality": 0.8899}, "family": "migration_planning", "id": "synthetic-migration_planning-000193", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.654, "estimated_latency_ms": 71362, "estimated_pass_rate": 0.6026, "estimated_quality": 0.799, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.819, "max_latency_ms": 170154, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6026, "quality": 0.799, "quality_target_met": false, "utility": 55.525921}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.819,\"max_latency_ms\":170154,\"minimum_quality\":0.8899},\"task\":{\"ambiguity\":0.5495,\"category\":\"coding\",\"complexity\":0.6426,\"consequence\":0.6195,\"evidence_need\":0.5438,\"family\":\"migration_planning\",\"input_tokens\":7240,\"reversibility\":0.6311,\"risk\":0.3537,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.368,\"id\":\"worker_a\",\"p95_latency_ms\":63543,\"pass_rate\":0.916,\"provider_group\":\"provider_y\",\"quality\":0.9163,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.139,\"id\":\"worker_b\",\"p95_latency_ms\":20037,\"pass_rate\":0.5273,\"provider_group\":\"provider_z\",\"quality\":0.7435,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.515,\"id\":\"worker_c\",\"p95_latency_ms\":51325,\"pass_rate\":0.8843,\"provider_group\":\"provider_x\",\"quality\":0.8368,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.654,\"estimated_latency_ms\":71362,\"estimated_pass_rate\":0.6026,\"estimated_quality\":0.799,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.819,\"max_latency_ms\":170154,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5495, "category": "coding", "complexity": 0.6426, "consequence": 0.6195, "evidence_need": 0.5438, "family": "migration_planning", "input_tokens": 7240, "request_id": "synthetic-migration_planning-000193", "reversibility": 0.6311, "risk": 0.3537, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 14.368, "id": "worker_a", "p95_latency_ms": 63543, "pass_rate": 0.916, "provider_group": "provider_y", "quality": 0.9163, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.139, "id": "worker_b", "p95_latency_ms": 20037, "pass_rate": 0.5273, "provider_group": "provider_z", "quality": 0.7435, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.515, "id": "worker_c", "p95_latency_ms": 51325, "pass_rate": 0.8843, "provider_group": "provider_x", "quality": 0.8368, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.624, "max_latency_ms": 95964, "minimum_quality": 0.7829}, "family": "migration_planning", "id": "synthetic-migration_planning-000010", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.672, "estimated_latency_ms": 84576, "estimated_pass_rate": 0.7594, "estimated_quality": 0.8134, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.624, "max_latency_ms": 95964, "policy": "balanced", "reason_codes": ["independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7594, "quality": 0.8134, "quality_target_met": true, "utility": 57.451832}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.624,\"max_latency_ms\":95964,\"minimum_quality\":0.7829},\"task\":{\"ambiguity\":0.7186,\"category\":\"coding\",\"complexity\":0.7004,\"consequence\":0.4823,\"evidence_need\":0.5098,\"family\":\"migration_planning\",\"input_tokens\":1680,\"reversibility\":0.5276,\"risk\":0.643,\"summary\":\"Plan a zero-downtime migration of access control with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.737,\"id\":\"worker_a\",\"p95_latency_ms\":62690,\"pass_rate\":0.8817,\"provider_group\":\"provider_y\",\"quality\":0.9098,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.935,\"id\":\"worker_b\",\"p95_latency_ms\":21886,\"pass_rate\":0.5609,\"provider_group\":\"provider_x\",\"quality\":0.7061,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.877,\"id\":\"worker_c\",\"p95_latency_ms\":71914,\"pass_rate\":0.9093,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.672,\"estimated_latency_ms\":84576,\"estimated_pass_rate\":0.7594,\"estimated_quality\":0.8134,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.624,\"max_latency_ms\":95964,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7186, "category": "coding", "complexity": 0.7004, "consequence": 0.4823, "evidence_need": 0.5098, "family": "migration_planning", "input_tokens": 1680, "request_id": "synthetic-migration_planning-000010", "reversibility": 0.5276, "risk": 0.643, "summary": "Plan a zero-downtime migration of access control with rollback checkpoints."}, "workers": [{"credits_per_call": 3.737, "id": "worker_a", "p95_latency_ms": 62690, "pass_rate": 0.8817, "provider_group": "provider_y", "quality": 0.9098, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.935, "id": "worker_b", "p95_latency_ms": 21886, "pass_rate": 0.5609, "provider_group": "provider_x", "quality": 0.7061, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.877, "id": "worker_c", "p95_latency_ms": 71914, "pass_rate": 0.9093, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.931, "max_latency_ms": 35886, "minimum_quality": 0.524}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000228", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.14, "estimated_latency_ms": 32787, "estimated_pass_rate": 0.4201, "estimated_quality": 0.6144, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.931, "max_latency_ms": 35886, "policy": "lean", "reason_codes": ["high_consequence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4201, "quality": 0.6144, "quality_target_met": true, "utility": 21.447483}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.931,\"max_latency_ms\":35886,\"minimum_quality\":0.524},\"task\":{\"ambiguity\":0.6879,\"category\":\"review\",\"complexity\":0.6401,\"consequence\":0.9115,\"evidence_need\":0.6391,\"family\":\"policy_compliance\",\"input_tokens\":3240,\"reversibility\":0.5064,\"risk\":0.845,\"summary\":\"Review whether model routing satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.646,\"id\":\"worker_a\",\"p95_latency_ms\":70402,\"pass_rate\":0.7688,\"provider_group\":\"provider_y\",\"quality\":0.8708,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.027,\"id\":\"worker_b\",\"p95_latency_ms\":77417,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9566,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.14,\"id\":\"worker_c\",\"p95_latency_ms\":32787,\"pass_rate\":0.6131,\"provider_group\":\"provider_x\",\"quality\":0.7572,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.14,\"estimated_latency_ms\":32787,\"estimated_pass_rate\":0.4201,\"estimated_quality\":0.6144,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.931,\"max_latency_ms\":35886,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6879, "category": "review", "complexity": 0.6401, "consequence": 0.9115, "evidence_need": 0.6391, "family": "policy_compliance", "input_tokens": 3240, "request_id": "synthetic-policy_compliance-000228", "reversibility": 0.5064, "risk": 0.845, "summary": "Review whether model routing satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.646, "id": "worker_a", "p95_latency_ms": 70402, "pass_rate": 0.7688, "provider_group": "provider_y", "quality": 0.8708, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.027, "id": "worker_b", "p95_latency_ms": 77417, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9566, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.14, "id": "worker_c", "p95_latency_ms": 32787, "pass_rate": 0.6131, "provider_group": "provider_x", "quality": 0.7572, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.438, "max_latency_ms": 129503, "minimum_quality": 0.798}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001215", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.326, "estimated_latency_ms": 83762, "estimated_pass_rate": 0.5586, "estimated_quality": 0.8172, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.438, "max_latency_ms": 129503, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5586, "quality": 0.8172, "quality_target_met": true, "utility": 61.786162}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.438,\"max_latency_ms\":129503,\"minimum_quality\":0.798},\"task\":{\"ambiguity\":0.6712,\"category\":\"review\",\"complexity\":0.6372,\"consequence\":0.7531,\"evidence_need\":0.7539,\"family\":\"policy_compliance\",\"input_tokens\":2240,\"reversibility\":0.5378,\"risk\":0.6948,\"summary\":\"Identify evidence required before approving private credentials for production.\"},\"workers\":[{\"credits_per_call\":3.211,\"id\":\"worker_a\",\"p95_latency_ms\":60211,\"pass_rate\":0.7655,\"provider_group\":\"provider_y\",\"quality\":0.8211,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.115,\"id\":\"worker_b\",\"p95_latency_ms\":23551,\"pass_rate\":0.4746,\"provider_group\":\"provider_x\",\"quality\":0.7952,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.638,\"id\":\"worker_c\",\"p95_latency_ms\":76658,\"pass_rate\":0.9525,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.326,\"estimated_latency_ms\":83762,\"estimated_pass_rate\":0.5586,\"estimated_quality\":0.8172,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.438,\"max_latency_ms\":129503,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6712, "category": "review", "complexity": 0.6372, "consequence": 0.7531, "evidence_need": 0.7539, "family": "policy_compliance", "input_tokens": 2240, "request_id": "synthetic-policy_compliance-001215", "reversibility": 0.5378, "risk": 0.6948, "summary": "Identify evidence required before approving private credentials for production."}, "workers": [{"credits_per_call": 3.211, "id": "worker_a", "p95_latency_ms": 60211, "pass_rate": 0.7655, "provider_group": "provider_y", "quality": 0.8211, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.115, "id": "worker_b", "p95_latency_ms": 23551, "pass_rate": 0.4746, "provider_group": "provider_x", "quality": 0.7952, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.638, "id": "worker_c", "p95_latency_ms": 76658, "pass_rate": 0.9525, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.93, "max_latency_ms": 80807, "minimum_quality": 0.7036}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000753", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.113, "estimated_latency_ms": 76715, "estimated_pass_rate": 0.7846, "estimated_quality": 0.8598, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.93, "max_latency_ms": 80807, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7846, "quality": 0.8598, "quality_target_met": true, "utility": 64.412893}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.93,\"max_latency_ms\":80807,\"minimum_quality\":0.7036},\"task\":{\"ambiguity\":0.5355,\"category\":\"review\",\"complexity\":0.5589,\"consequence\":0.8588,\"evidence_need\":0.7542,\"family\":\"policy_compliance\",\"input_tokens\":3520,\"reversibility\":0.6407,\"risk\":0.5235,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.31,\"id\":\"worker_a\",\"p95_latency_ms\":52032,\"pass_rate\":0.8833,\"provider_group\":\"provider_z\",\"quality\":0.922,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.803,\"id\":\"worker_b\",\"p95_latency_ms\":24683,\"pass_rate\":0.5359,\"provider_group\":\"provider_x\",\"quality\":0.7083,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.629,\"id\":\"worker_c\",\"p95_latency_ms\":89975,\"pass_rate\":0.8878,\"provider_group\":\"provider_y\",\"quality\":0.9841,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.113,\"estimated_latency_ms\":76715,\"estimated_pass_rate\":0.7846,\"estimated_quality\":0.8598,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.93,\"max_latency_ms\":80807,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5355, "category": "review", "complexity": 0.5589, "consequence": 0.8588, "evidence_need": 0.7542, "family": "policy_compliance", "input_tokens": 3520, "request_id": "synthetic-policy_compliance-000753", "reversibility": 0.6407, "risk": 0.5235, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.31, "id": "worker_a", "p95_latency_ms": 52032, "pass_rate": 0.8833, "provider_group": "provider_z", "quality": 0.922, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.803, "id": "worker_b", "p95_latency_ms": 24683, "pass_rate": 0.5359, "provider_group": "provider_x", "quality": 0.7083, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.629, "id": "worker_c", "p95_latency_ms": 89975, "pass_rate": 0.8878, "provider_group": "provider_y", "quality": 0.9841, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.665, "max_latency_ms": 45301, "minimum_quality": 0.6235}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000360", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.964, "estimated_latency_ms": 24877, "estimated_pass_rate": 0.2654, "estimated_quality": 0.5472, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.665, "max_latency_ms": 45301, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2654, "quality": 0.5472, "quality_target_met": false, "utility": 31.822747}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.665,\"max_latency_ms\":45301,\"minimum_quality\":0.6235},\"task\":{\"ambiguity\":0.7756,\"category\":\"review\",\"complexity\":0.8127,\"consequence\":0.6721,\"evidence_need\":0.692,\"family\":\"policy_compliance\",\"input_tokens\":4320,\"reversibility\":0.3843,\"risk\":0.4599,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":15.328,\"id\":\"worker_a\",\"p95_latency_ms\":54309,\"pass_rate\":0.9217,\"provider_group\":\"provider_y\",\"quality\":0.9518,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.964,\"id\":\"worker_b\",\"p95_latency_ms\":24877,\"pass_rate\":0.4534,\"provider_group\":\"provider_z\",\"quality\":0.6994,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.177,\"id\":\"worker_c\",\"p95_latency_ms\":70261,\"pass_rate\":0.7417,\"provider_group\":\"provider_x\",\"quality\":0.9021,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.964,\"estimated_latency_ms\":24877,\"estimated_pass_rate\":0.2654,\"estimated_quality\":0.5472,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.665,\"max_latency_ms\":45301,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7756, "category": "review", "complexity": 0.8127, "consequence": 0.6721, "evidence_need": 0.692, "family": "policy_compliance", "input_tokens": 4320, "request_id": "synthetic-policy_compliance-000360", "reversibility": 0.3843, "risk": 0.4599, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 15.328, "id": "worker_a", "p95_latency_ms": 54309, "pass_rate": 0.9217, "provider_group": "provider_y", "quality": 0.9518, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.964, "id": "worker_b", "p95_latency_ms": 24877, "pass_rate": 0.4534, "provider_group": "provider_z", "quality": 0.6994, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.177, "id": "worker_c", "p95_latency_ms": 70261, "pass_rate": 0.7417, "provider_group": "provider_x", "quality": 0.9021, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.897, "max_latency_ms": 32982, "minimum_quality": 0.6093}, "family": "migration_planning", "id": "synthetic-migration_planning-000691", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.219, "estimated_latency_ms": 22958, "estimated_pass_rate": 0.2607, "estimated_quality": 0.6147, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.897, "max_latency_ms": 32982, "policy": "lean", "reason_codes": ["complex_task", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2607, "quality": 0.6147, "quality_target_met": true, "utility": 38.029694}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.897,\"max_latency_ms\":32982,\"minimum_quality\":0.6093},\"task\":{\"ambiguity\":0.5834,\"category\":\"coding\",\"complexity\":0.7925,\"consequence\":0.477,\"evidence_need\":0.4705,\"family\":\"migration_planning\",\"input_tokens\":760,\"reversibility\":0.3464,\"risk\":0.7416,\"summary\":\"Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.452,\"id\":\"worker_a\",\"p95_latency_ms\":69665,\"pass_rate\":0.8504,\"provider_group\":\"provider_y\",\"quality\":0.9181,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.072,\"id\":\"worker_b\",\"p95_latency_ms\":65353,\"pass_rate\":0.9473,\"provider_group\":\"provider_x\",\"quality\":0.9113,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.219,\"id\":\"worker_c\",\"p95_latency_ms\":22958,\"pass_rate\":0.4547,\"provider_group\":\"provider_z\",\"quality\":0.7706,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.219,\"estimated_latency_ms\":22958,\"estimated_pass_rate\":0.2607,\"estimated_quality\":0.6147,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.897,\"max_latency_ms\":32982,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5834, "category": "coding", "complexity": 0.7925, "consequence": 0.477, "evidence_need": 0.4705, "family": "migration_planning", "input_tokens": 760, "request_id": "synthetic-migration_planning-000691", "reversibility": 0.3464, "risk": 0.7416, "summary": "Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints."}, "workers": [{"credits_per_call": 3.452, "id": "worker_a", "p95_latency_ms": 69665, "pass_rate": 0.8504, "provider_group": "provider_y", "quality": 0.9181, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.072, "id": "worker_b", "p95_latency_ms": 65353, "pass_rate": 0.9473, "provider_group": "provider_x", "quality": 0.9113, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.219, "id": "worker_c", "p95_latency_ms": 22958, "pass_rate": 0.4547, "provider_group": "provider_z", "quality": 0.7706, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.649, "max_latency_ms": 222128, "minimum_quality": 0.8215}, "family": "migration_planning", "id": "synthetic-migration_planning-000451", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.762, "estimated_latency_ms": 156134, "estimated_pass_rate": 0.7665, "estimated_quality": 0.92, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 22.649, "max_latency_ms": 222128, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7665, "quality": 0.92, "quality_target_met": true, "utility": 77.693619}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.649,\"max_latency_ms\":222128,\"minimum_quality\":0.8215},\"task\":{\"ambiguity\":0.4873,\"category\":\"coding\",\"complexity\":0.7997,\"consequence\":0.8352,\"evidence_need\":0.7732,\"family\":\"migration_planning\",\"input_tokens\":2440,\"reversibility\":0.5509,\"risk\":0.5631,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.118,\"id\":\"worker_a\",\"p95_latency_ms\":23710,\"pass_rate\":0.6192,\"provider_group\":\"provider_z\",\"quality\":0.7488,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.434,\"id\":\"worker_b\",\"p95_latency_ms\":61449,\"pass_rate\":0.9493,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.21,\"id\":\"worker_c\",\"p95_latency_ms\":70975,\"pass_rate\":0.8606,\"provider_group\":\"provider_y\",\"quality\":0.7959,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.762,\"estimated_latency_ms\":156134,\"estimated_pass_rate\":0.7665,\"estimated_quality\":0.92,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.649,\"max_latency_ms\":222128,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4873, "category": "coding", "complexity": 0.7997, "consequence": 0.8352, "evidence_need": 0.7732, "family": "migration_planning", "input_tokens": 2440, "request_id": "synthetic-migration_planning-000451", "reversibility": 0.5509, "risk": 0.5631, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 1.118, "id": "worker_a", "p95_latency_ms": 23710, "pass_rate": 0.6192, "provider_group": "provider_z", "quality": 0.7488, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.434, "id": "worker_b", "p95_latency_ms": 61449, "pass_rate": 0.9493, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.21, "id": "worker_c", "p95_latency_ms": 70975, "pass_rate": 0.8606, "provider_group": "provider_y", "quality": 0.7959, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.997, "max_latency_ms": 35885, "minimum_quality": 0.6767}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000162", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.094, "estimated_latency_ms": 25076, "estimated_pass_rate": 0.3206, "estimated_quality": 0.6983, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.997, "max_latency_ms": 35885, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3206, "quality": 0.6983, "quality_target_met": true, "utility": 43.689103}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.997,\"max_latency_ms\":35885,\"minimum_quality\":0.6767},\"task\":{\"ambiguity\":0.5775,\"category\":\"review\",\"complexity\":0.4304,\"consequence\":0.563,\"evidence_need\":0.8357,\"family\":\"policy_compliance\",\"input_tokens\":6280,\"reversibility\":0.4344,\"risk\":0.7699,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.094,\"id\":\"worker_a\",\"p95_latency_ms\":25076,\"pass_rate\":0.4754,\"provider_group\":\"provider_x\",\"quality\":0.7904,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.19,\"id\":\"worker_b\",\"p95_latency_ms\":87484,\"pass_rate\":0.9656,\"provider_group\":\"provider_z\",\"quality\":0.9395,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.584,\"id\":\"worker_c\",\"p95_latency_ms\":49803,\"pass_rate\":0.7448,\"provider_group\":\"provider_y\",\"quality\":0.8619,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.094,\"estimated_latency_ms\":25076,\"estimated_pass_rate\":0.3206,\"estimated_quality\":0.6983,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.997,\"max_latency_ms\":35885,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5775, "category": "review", "complexity": 0.4304, "consequence": 0.563, "evidence_need": 0.8357, "family": "policy_compliance", "input_tokens": 6280, "request_id": "synthetic-policy_compliance-000162", "reversibility": 0.4344, "risk": 0.7699, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 1.094, "id": "worker_a", "p95_latency_ms": 25076, "pass_rate": 0.4754, "provider_group": "provider_x", "quality": 0.7904, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.19, "id": "worker_b", "p95_latency_ms": 87484, "pass_rate": 0.9656, "provider_group": "provider_z", "quality": 0.9395, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.584, "id": "worker_c", "p95_latency_ms": 49803, "pass_rate": 0.7448, "provider_group": "provider_y", "quality": 0.8619, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.075, "max_latency_ms": 187769, "minimum_quality": 0.7087}, "family": "migration_planning", "id": "synthetic-migration_planning-000340", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.321, "estimated_latency_ms": 87464, "estimated_pass_rate": 0.5712, "estimated_quality": 0.7392, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.075, "max_latency_ms": 187769, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5712, "quality": 0.7392, "quality_target_met": true, "utility": 54.638676}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.075,\"max_latency_ms\":187769,\"minimum_quality\":0.7087},\"task\":{\"ambiguity\":0.5683,\"category\":\"coding\",\"complexity\":0.7598,\"consequence\":0.4934,\"evidence_need\":0.7058,\"family\":\"migration_planning\",\"input_tokens\":5920,\"reversibility\":0.6413,\"risk\":0.4417,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":3.43,\"id\":\"worker_a\",\"p95_latency_ms\":66296,\"pass_rate\":0.7494,\"provider_group\":\"provider_x\",\"quality\":0.8046,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.284,\"id\":\"worker_b\",\"p95_latency_ms\":68163,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.891,\"id\":\"worker_c\",\"p95_latency_ms\":21168,\"pass_rate\":0.5473,\"provider_group\":\"provider_y\",\"quality\":0.695,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.321,\"estimated_latency_ms\":87464,\"estimated_pass_rate\":0.5712,\"estimated_quality\":0.7392,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.075,\"max_latency_ms\":187769,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5683, "category": "coding", "complexity": 0.7598, "consequence": 0.4934, "evidence_need": 0.7058, "family": "migration_planning", "input_tokens": 5920, "request_id": "synthetic-migration_planning-000340", "reversibility": 0.6413, "risk": 0.4417, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 3.43, "id": "worker_a", "p95_latency_ms": 66296, "pass_rate": 0.7494, "provider_group": "provider_x", "quality": 0.8046, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.284, "id": "worker_b", "p95_latency_ms": 68163, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.891, "id": "worker_c", "p95_latency_ms": 21168, "pass_rate": 0.5473, "provider_group": "provider_y", "quality": 0.695, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.338, "max_latency_ms": 205077, "minimum_quality": 0.8647}, "family": "migration_planning", "id": "synthetic-migration_planning-000994", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.023, "estimated_latency_ms": 149324, "estimated_pass_rate": 0.7492, "estimated_quality": 0.9016, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.338, "max_latency_ms": 205077, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7492, "quality": 0.9016, "quality_target_met": true, "utility": 74.428692}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.338,\"max_latency_ms\":205077,\"minimum_quality\":0.8647},\"task\":{\"ambiguity\":0.5105,\"category\":\"coding\",\"complexity\":0.7861,\"consequence\":0.7901,\"evidence_need\":0.6096,\"family\":\"migration_planning\",\"input_tokens\":3160,\"reversibility\":0.5585,\"risk\":0.6237,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.266,\"id\":\"worker_a\",\"p95_latency_ms\":52795,\"pass_rate\":0.8242,\"provider_group\":\"provider_y\",\"quality\":0.9116,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.043,\"id\":\"worker_b\",\"p95_latency_ms\":31069,\"pass_rate\":0.4746,\"provider_group\":\"provider_x\",\"quality\":0.7511,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.714,\"id\":\"worker_c\",\"p95_latency_ms\":65460,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9499,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.023,\"estimated_latency_ms\":149324,\"estimated_pass_rate\":0.7492,\"estimated_quality\":0.9016,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.338,\"max_latency_ms\":205077,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5105, "category": "coding", "complexity": 0.7861, "consequence": 0.7901, "evidence_need": 0.6096, "family": "migration_planning", "input_tokens": 3160, "request_id": "synthetic-migration_planning-000994", "reversibility": 0.5585, "risk": 0.6237, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 4.266, "id": "worker_a", "p95_latency_ms": 52795, "pass_rate": 0.8242, "provider_group": "provider_y", "quality": 0.9116, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.043, "id": "worker_b", "p95_latency_ms": 31069, "pass_rate": 0.4746, "provider_group": "provider_x", "quality": 0.7511, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.714, "id": "worker_c", "p95_latency_ms": 65460, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9499, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.737, "max_latency_ms": 27176, "minimum_quality": 0.6296}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000119", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.853, "estimated_latency_ms": 17771, "estimated_pass_rate": 0.2685, "estimated_quality": 0.5036, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.737, "max_latency_ms": 27176, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2685, "quality": 0.5036, "quality_target_met": false, "utility": 25.726512}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.737,\"max_latency_ms\":27176,\"minimum_quality\":0.6296},\"task\":{\"ambiguity\":0.4691,\"category\":\"analysis\",\"complexity\":0.7419,\"consequence\":0.6587,\"evidence_need\":0.6373,\"family\":\"performance_tuning\",\"input_tokens\":4240,\"reversibility\":0.8874,\"risk\":0.499,\"summary\":\"Diagnose latency variance in access control and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.853,\"id\":\"worker_a\",\"p95_latency_ms\":17771,\"pass_rate\":0.4285,\"provider_group\":\"provider_z\",\"quality\":0.6252,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.889,\"id\":\"worker_b\",\"p95_latency_ms\":59896,\"pass_rate\":0.9424,\"provider_group\":\"provider_x\",\"quality\":0.8949,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.092,\"id\":\"worker_c\",\"p95_latency_ms\":71892,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.853,\"estimated_latency_ms\":17771,\"estimated_pass_rate\":0.2685,\"estimated_quality\":0.5036,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.737,\"max_latency_ms\":27176,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4691, "category": "analysis", "complexity": 0.7419, "consequence": 0.6587, "evidence_need": 0.6373, "family": "performance_tuning", "input_tokens": 4240, "request_id": "synthetic-performance_tuning-000119", "reversibility": 0.8874, "risk": 0.499, "summary": "Diagnose latency variance in access control and rank the next experiments."}, "workers": [{"credits_per_call": 0.853, "id": "worker_a", "p95_latency_ms": 17771, "pass_rate": 0.4285, "provider_group": "provider_z", "quality": 0.6252, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.889, "id": "worker_b", "p95_latency_ms": 59896, "pass_rate": 0.9424, "provider_group": "provider_x", "quality": 0.8949, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.092, "id": "worker_c", "p95_latency_ms": 71892, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.894, "max_latency_ms": 210095, "minimum_quality": 0.9505}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000189", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.023, "estimated_latency_ms": 186506, "estimated_pass_rate": 0.7787, "estimated_quality": 0.9341, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.894, "max_latency_ms": 210095, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7787, "quality": 0.9341, "quality_target_met": false, "utility": 77.909624}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.894,\"max_latency_ms\":210095,\"minimum_quality\":0.9505},\"task\":{\"ambiguity\":0.6857,\"category\":\"review\",\"complexity\":0.5375,\"consequence\":0.6058,\"evidence_need\":0.9145,\"family\":\"policy_compliance\",\"input_tokens\":2000,\"reversibility\":0.3654,\"risk\":0.8191,\"summary\":\"Identify evidence required before approving usage metering for production.\"},\"workers\":[{\"credits_per_call\":13.133,\"id\":\"worker_a\",\"p95_latency_ms\":88056,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.015,\"id\":\"worker_b\",\"p95_latency_ms\":32696,\"pass_rate\":0.5359,\"provider_group\":\"provider_z\",\"quality\":0.7651,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.875,\"id\":\"worker_c\",\"p95_latency_ms\":65754,\"pass_rate\":0.8762,\"provider_group\":\"provider_y\",\"quality\":0.8056,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.023,\"estimated_latency_ms\":186506,\"estimated_pass_rate\":0.7787,\"estimated_quality\":0.9341,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.894,\"max_latency_ms\":210095,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6857, "category": "review", "complexity": 0.5375, "consequence": 0.6058, "evidence_need": 0.9145, "family": "policy_compliance", "input_tokens": 2000, "request_id": "synthetic-policy_compliance-000189", "reversibility": 0.3654, "risk": 0.8191, "summary": "Identify evidence required before approving usage metering for production."}, "workers": [{"credits_per_call": 13.133, "id": "worker_a", "p95_latency_ms": 88056, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.015, "id": "worker_b", "p95_latency_ms": 32696, "pass_rate": 0.5359, "provider_group": "provider_z", "quality": 0.7651, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.875, "id": "worker_c", "p95_latency_ms": 65754, "pass_rate": 0.8762, "provider_group": "provider_y", "quality": 0.8056, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.866, "max_latency_ms": 192351, "minimum_quality": 0.8883}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001080", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.241, "estimated_latency_ms": 132631, "estimated_pass_rate": 0.8996, "estimated_quality": 0.8979, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.866, "max_latency_ms": 192351, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8996, "quality": 0.8979, "quality_target_met": true, "utility": 69.965979}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.866,\"max_latency_ms\":192351,\"minimum_quality\":0.8883},\"task\":{\"ambiguity\":0.4656,\"category\":\"review\",\"complexity\":0.5714,\"consequence\":0.6814,\"evidence_need\":0.7562,\"family\":\"policy_compliance\",\"input_tokens\":520,\"reversibility\":0.3316,\"risk\":0.6228,\"summary\":\"Review whether deployment rollback satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.063,\"id\":\"worker_a\",\"p95_latency_ms\":78845,\"pass_rate\":0.7526,\"provider_group\":\"provider_y\",\"quality\":0.8501,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.178,\"id\":\"worker_b\",\"p95_latency_ms\":53786,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9152,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.039,\"id\":\"worker_c\",\"p95_latency_ms\":20855,\"pass_rate\":0.4802,\"provider_group\":\"provider_x\",\"quality\":0.7638,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.241,\"estimated_latency_ms\":132631,\"estimated_pass_rate\":0.8996,\"estimated_quality\":0.8979,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.866,\"max_latency_ms\":192351,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4656, "category": "review", "complexity": 0.5714, "consequence": 0.6814, "evidence_need": 0.7562, "family": "policy_compliance", "input_tokens": 520, "request_id": "synthetic-policy_compliance-001080", "reversibility": 0.3316, "risk": 0.6228, "summary": "Review whether deployment rollback satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.063, "id": "worker_a", "p95_latency_ms": 78845, "pass_rate": 0.7526, "provider_group": "provider_y", "quality": 0.8501, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.178, "id": "worker_b", "p95_latency_ms": 53786, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9152, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.039, "id": "worker_c", "p95_latency_ms": 20855, "pass_rate": 0.4802, "provider_group": "provider_x", "quality": 0.7638, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.441, "max_latency_ms": 129860, "minimum_quality": 0.8397}, "family": "migration_planning", "id": "synthetic-migration_planning-000571", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.508, "estimated_latency_ms": 88579, "estimated_pass_rate": 0.7786, "estimated_quality": 0.8223, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.441, "max_latency_ms": 129860, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7786, "quality": 0.8223, "quality_target_met": false, "utility": 59.284277}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.441,\"max_latency_ms\":129860,\"minimum_quality\":0.8397},\"task\":{\"ambiguity\":0.3786,\"category\":\"coding\",\"complexity\":0.7001,\"consequence\":0.6209,\"evidence_need\":0.7359,\"family\":\"migration_planning\",\"input_tokens\":4120,\"reversibility\":0.5198,\"risk\":0.5201,\"summary\":\"Review a staged migration for credit reservation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.6,\"id\":\"worker_a\",\"p95_latency_ms\":64979,\"pass_rate\":0.8666,\"provider_group\":\"provider_z\",\"quality\":0.8669,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.908,\"id\":\"worker_b\",\"p95_latency_ms\":23600,\"pass_rate\":0.6025,\"provider_group\":\"provider_y\",\"quality\":0.7251,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.533,\"id\":\"worker_c\",\"p95_latency_ms\":53803,\"pass_rate\":0.9506,\"provider_group\":\"provider_x\",\"quality\":0.9241,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.508,\"estimated_latency_ms\":88579,\"estimated_pass_rate\":0.7786,\"estimated_quality\":0.8223,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.441,\"max_latency_ms\":129860,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3786, "category": "coding", "complexity": 0.7001, "consequence": 0.6209, "evidence_need": 0.7359, "family": "migration_planning", "input_tokens": 4120, "request_id": "synthetic-migration_planning-000571", "reversibility": 0.5198, "risk": 0.5201, "summary": "Review a staged migration for credit reservation and identify irreversible steps."}, "workers": [{"credits_per_call": 3.6, "id": "worker_a", "p95_latency_ms": 64979, "pass_rate": 0.8666, "provider_group": "provider_z", "quality": 0.8669, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.908, "id": "worker_b", "p95_latency_ms": 23600, "pass_rate": 0.6025, "provider_group": "provider_y", "quality": 0.7251, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.533, "id": "worker_c", "p95_latency_ms": 53803, "pass_rate": 0.9506, "provider_group": "provider_x", "quality": 0.9241, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.715, "max_latency_ms": 201372, "minimum_quality": 0.8504}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000951", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.0, "estimated_latency_ms": 151431, "estimated_pass_rate": 0.7601, "estimated_quality": 0.8847, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.715, "max_latency_ms": 201372, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7601, "quality": 0.8847, "quality_target_met": true, "utility": 75.348171}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.715,\"max_latency_ms\":201372,\"minimum_quality\":0.8504},\"task\":{\"ambiguity\":0.3662,\"category\":\"review\",\"complexity\":0.6726,\"consequence\":0.9279,\"evidence_need\":0.7504,\"family\":\"policy_compliance\",\"input_tokens\":7520,\"reversibility\":0.4372,\"risk\":0.7391,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":0.99,\"id\":\"worker_a\",\"p95_latency_ms\":32985,\"pass_rate\":0.521,\"provider_group\":\"provider_y\",\"quality\":0.6962,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.894,\"id\":\"worker_b\",\"p95_latency_ms\":53213,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9359,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.116,\"id\":\"worker_c\",\"p95_latency_ms\":65233,\"pass_rate\":0.746,\"provider_group\":\"provider_x\",\"quality\":0.8853,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.0,\"estimated_latency_ms\":151431,\"estimated_pass_rate\":0.7601,\"estimated_quality\":0.8847,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.715,\"max_latency_ms\":201372,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3662, "category": "review", "complexity": 0.6726, "consequence": 0.9279, "evidence_need": 0.7504, "family": "policy_compliance", "input_tokens": 7520, "request_id": "synthetic-policy_compliance-000951", "reversibility": 0.4372, "risk": 0.7391, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 0.99, "id": "worker_a", "p95_latency_ms": 32985, "pass_rate": 0.521, "provider_group": "provider_y", "quality": 0.6962, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.894, "id": "worker_b", "p95_latency_ms": 53213, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9359, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.116, "id": "worker_c", "p95_latency_ms": 65233, "pass_rate": 0.746, "provider_group": "provider_x", "quality": 0.8853, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.74, "max_latency_ms": 127801, "minimum_quality": 0.7909}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000666", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.089, "estimated_latency_ms": 74048, "estimated_pass_rate": 0.6531, "estimated_quality": 0.8651, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.74, "max_latency_ms": 127801, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6531, "quality": 0.8651, "quality_target_met": true, "utility": 69.694649}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.74,\"max_latency_ms\":127801,\"minimum_quality\":0.7909},\"task\":{\"ambiguity\":0.5647,\"category\":\"review\",\"complexity\":0.4454,\"consequence\":0.8108,\"evidence_need\":0.9197,\"family\":\"policy_compliance\",\"input_tokens\":3360,\"reversibility\":0.6766,\"risk\":0.7707,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.173,\"id\":\"worker_a\",\"p95_latency_ms\":21378,\"pass_rate\":0.5276,\"provider_group\":\"provider_z\",\"quality\":0.7689,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.166,\"id\":\"worker_b\",\"p95_latency_ms\":80026,\"pass_rate\":0.886,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.916,\"id\":\"worker_c\",\"p95_latency_ms\":52670,\"pass_rate\":0.8136,\"provider_group\":\"provider_y\",\"quality\":0.8861,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.089,\"estimated_latency_ms\":74048,\"estimated_pass_rate\":0.6531,\"estimated_quality\":0.8651,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.74,\"max_latency_ms\":127801,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5647, "category": "review", "complexity": 0.4454, "consequence": 0.8108, "evidence_need": 0.9197, "family": "policy_compliance", "input_tokens": 3360, "request_id": "synthetic-policy_compliance-000666", "reversibility": 0.6766, "risk": 0.7707, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.173, "id": "worker_a", "p95_latency_ms": 21378, "pass_rate": 0.5276, "provider_group": "provider_z", "quality": 0.7689, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.166, "id": "worker_b", "p95_latency_ms": 80026, "pass_rate": 0.886, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.916, "id": "worker_c", "p95_latency_ms": 52670, "pass_rate": 0.8136, "provider_group": "provider_y", "quality": 0.8861, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.027, "max_latency_ms": 127432, "minimum_quality": 0.7231}, "family": "migration_planning", "id": "synthetic-migration_planning-001156", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.321, "estimated_latency_ms": 107463, "estimated_pass_rate": 0.6986, "estimated_quality": 0.761, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.027, "max_latency_ms": 127432, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6986, "quality": 0.761, "quality_target_met": true, "utility": 52.811871}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.027,\"max_latency_ms\":127432,\"minimum_quality\":0.7231},\"task\":{\"ambiguity\":0.5622,\"category\":\"coding\",\"complexity\":0.8063,\"consequence\":0.688,\"evidence_need\":0.4211,\"family\":\"migration_planning\",\"input_tokens\":2160,\"reversibility\":0.5153,\"risk\":0.5872,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":0.787,\"id\":\"worker_a\",\"p95_latency_ms\":26273,\"pass_rate\":0.5624,\"provider_group\":\"provider_y\",\"quality\":0.73,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.534,\"id\":\"worker_b\",\"p95_latency_ms\":81190,\"pass_rate\":0.7961,\"provider_group\":\"provider_z\",\"quality\":0.8436,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.228,\"id\":\"worker_c\",\"p95_latency_ms\":61350,\"pass_rate\":0.9061,\"provider_group\":\"provider_x\",\"quality\":0.921,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.321,\"estimated_latency_ms\":107463,\"estimated_pass_rate\":0.6986,\"estimated_quality\":0.761,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.027,\"max_latency_ms\":127432,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5622, "category": "coding", "complexity": 0.8063, "consequence": 0.688, "evidence_need": 0.4211, "family": "migration_planning", "input_tokens": 2160, "request_id": "synthetic-migration_planning-001156", "reversibility": 0.5153, "risk": 0.5872, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 0.787, "id": "worker_a", "p95_latency_ms": 26273, "pass_rate": 0.5624, "provider_group": "provider_y", "quality": 0.73, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.534, "id": "worker_b", "p95_latency_ms": 81190, "pass_rate": 0.7961, "provider_group": "provider_z", "quality": 0.8436, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.228, "id": "worker_c", "p95_latency_ms": 61350, "pass_rate": 0.9061, "provider_group": "provider_x", "quality": 0.921, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.773, "max_latency_ms": 201690, "minimum_quality": 0.7967}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001142", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.577, "estimated_latency_ms": 173757, "estimated_pass_rate": 0.8511, "estimated_quality": 0.9246, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.773, "max_latency_ms": 201690, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8511, "quality": 0.9246, "quality_target_met": true, "utility": 70.696527}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.773,\"max_latency_ms\":201690,\"minimum_quality\":0.7967},\"task\":{\"ambiguity\":0.3002,\"category\":\"analysis\",\"complexity\":0.8827,\"consequence\":0.3292,\"evidence_need\":0.5072,\"family\":\"performance_tuning\",\"input_tokens\":7800,\"reversibility\":0.7528,\"risk\":0.3762,\"summary\":\"Diagnose latency variance in access control and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.752,\"id\":\"worker_a\",\"p95_latency_ms\":69520,\"pass_rate\":0.9677,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.922,\"id\":\"worker_b\",\"p95_latency_ms\":21651,\"pass_rate\":0.4775,\"provider_group\":\"provider_z\",\"quality\":0.6758,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.903,\"id\":\"worker_c\",\"p95_latency_ms\":82586,\"pass_rate\":0.8995,\"provider_group\":\"provider_y\",\"quality\":0.9034,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.577,\"estimated_latency_ms\":173757,\"estimated_pass_rate\":0.8511,\"estimated_quality\":0.9246,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.773,\"max_latency_ms\":201690,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3002, "category": "analysis", "complexity": 0.8827, "consequence": 0.3292, "evidence_need": 0.5072, "family": "performance_tuning", "input_tokens": 7800, "request_id": "synthetic-performance_tuning-001142", "reversibility": 0.7528, "risk": 0.3762, "summary": "Diagnose latency variance in access control and rank the next experiments."}, "workers": [{"credits_per_call": 11.752, "id": "worker_a", "p95_latency_ms": 69520, "pass_rate": 0.9677, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.922, "id": "worker_b", "p95_latency_ms": 21651, "pass_rate": 0.4775, "provider_group": "provider_z", "quality": 0.6758, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.903, "id": "worker_c", "p95_latency_ms": 82586, "pass_rate": 0.8995, "provider_group": "provider_y", "quality": 0.9034, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.652, "max_latency_ms": 160909, "minimum_quality": 0.7704}, "family": "migration_planning", "id": "synthetic-migration_planning-000775", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.831, "estimated_latency_ms": 84317, "estimated_pass_rate": 0.6548, "estimated_quality": 0.812, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.652, "max_latency_ms": 160909, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6548, "quality": 0.812, "quality_target_met": true, "utility": 64.630602}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.652,\"max_latency_ms\":160909,\"minimum_quality\":0.7704},\"task\":{\"ambiguity\":0.6472,\"category\":\"coding\",\"complexity\":0.5657,\"consequence\":0.5899,\"evidence_need\":0.8351,\"family\":\"migration_planning\",\"input_tokens\":4360,\"reversibility\":0.4564,\"risk\":0.7499,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.861,\"id\":\"worker_a\",\"p95_latency_ms\":61525,\"pass_rate\":0.7959,\"provider_group\":\"provider_z\",\"quality\":0.8609,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.288,\"id\":\"worker_b\",\"p95_latency_ms\":71303,\"pass_rate\":0.9838,\"provider_group\":\"provider_y\",\"quality\":0.936,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.97,\"id\":\"worker_c\",\"p95_latency_ms\":22792,\"pass_rate\":0.5786,\"provider_group\":\"provider_x\",\"quality\":0.7192,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.831,\"estimated_latency_ms\":84317,\"estimated_pass_rate\":0.6548,\"estimated_quality\":0.812,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.652,\"max_latency_ms\":160909,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6472, "category": "coding", "complexity": 0.5657, "consequence": 0.5899, "evidence_need": 0.8351, "family": "migration_planning", "input_tokens": 4360, "request_id": "synthetic-migration_planning-000775", "reversibility": 0.4564, "risk": 0.7499, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 2.861, "id": "worker_a", "p95_latency_ms": 61525, "pass_rate": 0.7959, "provider_group": "provider_z", "quality": 0.8609, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.288, "id": "worker_b", "p95_latency_ms": 71303, "pass_rate": 0.9838, "provider_group": "provider_y", "quality": 0.936, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.97, "id": "worker_c", "p95_latency_ms": 22792, "pass_rate": 0.5786, "provider_group": "provider_x", "quality": 0.7192, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.22, "max_latency_ms": 176230, "minimum_quality": 0.9178}, "family": "migration_planning", "id": "synthetic-migration_planning-001093", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.086, "estimated_latency_ms": 168925, "estimated_pass_rate": 0.6928, "estimated_quality": 0.8866, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.22, "max_latency_ms": 176230, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6928, "quality": 0.8866, "quality_target_met": false, "utility": 69.637124}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.22,\"max_latency_ms\":176230,\"minimum_quality\":0.9178},\"task\":{\"ambiguity\":0.6683,\"category\":\"coding\",\"complexity\":0.8678,\"consequence\":0.4764,\"evidence_need\":0.7291,\"family\":\"migration_planning\",\"input_tokens\":6760,\"reversibility\":0.6099,\"risk\":0.5186,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.843,\"id\":\"worker_a\",\"p95_latency_ms\":21319,\"pass_rate\":0.4771,\"provider_group\":\"provider_y\",\"quality\":0.7176,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.201,\"id\":\"worker_b\",\"p95_latency_ms\":73794,\"pass_rate\":0.7838,\"provider_group\":\"provider_z\",\"quality\":0.8412,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.042,\"id\":\"worker_c\",\"p95_latency_ms\":73812,\"pass_rate\":0.9456,\"provider_group\":\"provider_x\",\"quality\":0.9506,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.086,\"estimated_latency_ms\":168925,\"estimated_pass_rate\":0.6928,\"estimated_quality\":0.8866,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.22,\"max_latency_ms\":176230,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6683, "category": "coding", "complexity": 0.8678, "consequence": 0.4764, "evidence_need": 0.7291, "family": "migration_planning", "input_tokens": 6760, "request_id": "synthetic-migration_planning-001093", "reversibility": 0.6099, "risk": 0.5186, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 0.843, "id": "worker_a", "p95_latency_ms": 21319, "pass_rate": 0.4771, "provider_group": "provider_y", "quality": 0.7176, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.201, "id": "worker_b", "p95_latency_ms": 73794, "pass_rate": 0.7838, "provider_group": "provider_z", "quality": 0.8412, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.042, "id": "worker_c", "p95_latency_ms": 73812, "pass_rate": 0.9456, "provider_group": "provider_x", "quality": 0.9506, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.603, "max_latency_ms": 226836, "minimum_quality": 0.9209}, "family": "migration_planning", "id": "synthetic-migration_planning-000442", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.633, "estimated_latency_ms": 163499, "estimated_pass_rate": 0.743, "estimated_quality": 0.891, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.603, "max_latency_ms": 226836, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.743, "quality": 0.891, "quality_target_met": false, "utility": 71.761853}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.603,\"max_latency_ms\":226836,\"minimum_quality\":0.9209},\"task\":{\"ambiguity\":0.5942,\"category\":\"coding\",\"complexity\":0.9201,\"consequence\":0.431,\"evidence_need\":0.4239,\"family\":\"migration_planning\",\"input_tokens\":8000,\"reversibility\":0.4582,\"risk\":0.643,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.892,\"id\":\"worker_a\",\"p95_latency_ms\":67066,\"pass_rate\":0.8098,\"provider_group\":\"provider_z\",\"quality\":0.8452,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.548,\"id\":\"worker_b\",\"p95_latency_ms\":70828,\"pass_rate\":0.9459,\"provider_group\":\"provider_x\",\"quality\":0.9683,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.193,\"id\":\"worker_c\",\"p95_latency_ms\":25605,\"pass_rate\":0.5711,\"provider_group\":\"provider_y\",\"quality\":0.7509,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.633,\"estimated_latency_ms\":163499,\"estimated_pass_rate\":0.743,\"estimated_quality\":0.891,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.603,\"max_latency_ms\":226836,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5942, "category": "coding", "complexity": 0.9201, "consequence": 0.431, "evidence_need": 0.4239, "family": "migration_planning", "input_tokens": 8000, "request_id": "synthetic-migration_planning-000442", "reversibility": 0.4582, "risk": 0.643, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 2.892, "id": "worker_a", "p95_latency_ms": 67066, "pass_rate": 0.8098, "provider_group": "provider_z", "quality": 0.8452, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.548, "id": "worker_b", "p95_latency_ms": 70828, "pass_rate": 0.9459, "provider_group": "provider_x", "quality": 0.9683, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.193, "id": "worker_c", "p95_latency_ms": 25605, "pass_rate": 0.5711, "provider_group": "provider_y", "quality": 0.7509, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.747, "max_latency_ms": 146424, "minimum_quality": 0.7634}, "family": "migration_planning", "id": "synthetic-migration_planning-000772", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.761, "estimated_latency_ms": 102590, "estimated_pass_rate": 0.5702, "estimated_quality": 0.7867, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.747, "max_latency_ms": 146424, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5702, "quality": 0.7867, "quality_target_met": true, "utility": 55.315301}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.747,\"max_latency_ms\":146424,\"minimum_quality\":0.7634},\"task\":{\"ambiguity\":0.7095,\"category\":\"coding\",\"complexity\":0.7758,\"consequence\":0.5833,\"evidence_need\":0.4714,\"family\":\"migration_planning\",\"input_tokens\":760,\"reversibility\":0.3245,\"risk\":0.7268,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.911,\"id\":\"worker_a\",\"p95_latency_ms\":23259,\"pass_rate\":0.4759,\"provider_group\":\"provider_z\",\"quality\":0.7904,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.09,\"id\":\"worker_b\",\"p95_latency_ms\":65640,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9474,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.85,\"id\":\"worker_c\",\"p95_latency_ms\":79331,\"pass_rate\":0.747,\"provider_group\":\"provider_y\",\"quality\":0.8985,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.761,\"estimated_latency_ms\":102590,\"estimated_pass_rate\":0.5702,\"estimated_quality\":0.7867,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.747,\"max_latency_ms\":146424,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7095, "category": "coding", "complexity": 0.7758, "consequence": 0.5833, "evidence_need": 0.4714, "family": "migration_planning", "input_tokens": 760, "request_id": "synthetic-migration_planning-000772", "reversibility": 0.3245, "risk": 0.7268, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 0.911, "id": "worker_a", "p95_latency_ms": 23259, "pass_rate": 0.4759, "provider_group": "provider_z", "quality": 0.7904, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.09, "id": "worker_b", "p95_latency_ms": 65640, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9474, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.85, "id": "worker_c", "p95_latency_ms": 79331, "pass_rate": 0.747, "provider_group": "provider_y", "quality": 0.8985, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.978, "max_latency_ms": 214438, "minimum_quality": 0.8291}, "family": "migration_planning", "id": "synthetic-migration_planning-001192", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.015, "estimated_latency_ms": 119089, "estimated_pass_rate": 0.8755, "estimated_quality": 0.9386, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.978, "max_latency_ms": 214438, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8755, "quality": 0.9386, "quality_target_met": true, "utility": 71.545715}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.978,\"max_latency_ms\":214438,\"minimum_quality\":0.8291},\"task\":{\"ambiguity\":0.3099,\"category\":\"coding\",\"complexity\":0.5779,\"consequence\":0.6824,\"evidence_need\":0.6384,\"family\":\"migration_planning\",\"input_tokens\":1960,\"reversibility\":0.4936,\"risk\":0.4366,\"summary\":\"Plan a zero-downtime migration of prompt caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.01,\"id\":\"worker_a\",\"p95_latency_ms\":59955,\"pass_rate\":0.8589,\"provider_group\":\"provider_z\",\"quality\":0.882,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.005,\"id\":\"worker_b\",\"p95_latency_ms\":59134,\"pass_rate\":0.923,\"provider_group\":\"provider_y\",\"quality\":0.9597,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.952,\"id\":\"worker_c\",\"p95_latency_ms\":27937,\"pass_rate\":0.5734,\"provider_group\":\"provider_x\",\"quality\":0.6845,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.015,\"estimated_latency_ms\":119089,\"estimated_pass_rate\":0.8755,\"estimated_quality\":0.9386,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.978,\"max_latency_ms\":214438,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3099, "category": "coding", "complexity": 0.5779, "consequence": 0.6824, "evidence_need": 0.6384, "family": "migration_planning", "input_tokens": 1960, "request_id": "synthetic-migration_planning-001192", "reversibility": 0.4936, "risk": 0.4366, "summary": "Plan a zero-downtime migration of prompt caching with rollback checkpoints."}, "workers": [{"credits_per_call": 4.01, "id": "worker_a", "p95_latency_ms": 59955, "pass_rate": 0.8589, "provider_group": "provider_z", "quality": 0.882, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.005, "id": "worker_b", "p95_latency_ms": 59134, "pass_rate": 0.923, "provider_group": "provider_y", "quality": 0.9597, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.952, "id": "worker_c", "p95_latency_ms": 27937, "pass_rate": 0.5734, "provider_group": "provider_x", "quality": 0.6845, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.969, "max_latency_ms": 36022, "minimum_quality": 0.5791}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000465", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.139, "estimated_latency_ms": 20244, "estimated_pass_rate": 0.4077, "estimated_quality": 0.6551, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.969, "max_latency_ms": 36022, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4077, "quality": 0.6551, "quality_target_met": true, "utility": 30.067925}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.969,\"max_latency_ms\":36022,\"minimum_quality\":0.5791},\"task\":{\"ambiguity\":0.7864,\"category\":\"review\",\"complexity\":0.5282,\"consequence\":0.671,\"evidence_need\":0.939,\"family\":\"policy_compliance\",\"input_tokens\":4960,\"reversibility\":0.3133,\"risk\":0.8051,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":12.125,\"id\":\"worker_a\",\"p95_latency_ms\":56203,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.443,\"id\":\"worker_b\",\"p95_latency_ms\":52457,\"pass_rate\":0.8077,\"provider_group\":\"provider_z\",\"quality\":0.8778,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.139,\"id\":\"worker_c\",\"p95_latency_ms\":20244,\"pass_rate\":0.5923,\"provider_group\":\"provider_x\",\"quality\":0.7814,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.139,\"estimated_latency_ms\":20244,\"estimated_pass_rate\":0.4077,\"estimated_quality\":0.6551,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.969,\"max_latency_ms\":36022,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7864, "category": "review", "complexity": 0.5282, "consequence": 0.671, "evidence_need": 0.939, "family": "policy_compliance", "input_tokens": 4960, "request_id": "synthetic-policy_compliance-000465", "reversibility": 0.3133, "risk": 0.8051, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 12.125, "id": "worker_a", "p95_latency_ms": 56203, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.443, "id": "worker_b", "p95_latency_ms": 52457, "pass_rate": 0.8077, "provider_group": "provider_z", "quality": 0.8778, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.139, "id": "worker_c", "p95_latency_ms": 20244, "pass_rate": 0.5923, "provider_group": "provider_x", "quality": 0.7814, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 22.501, "max_latency_ms": 107874, "minimum_quality": 0.7773}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001032", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.839, "estimated_latency_ms": 98500, "estimated_pass_rate": 0.6492, "estimated_quality": 0.8393, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 22.501, "max_latency_ms": 107874, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6492, "quality": 0.8393, "quality_target_met": true, "utility": 61.343817}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":22.501,\"max_latency_ms\":107874,\"minimum_quality\":0.7773},\"task\":{\"ambiguity\":0.796,\"category\":\"review\",\"complexity\":0.6402,\"consequence\":0.8411,\"evidence_need\":0.7973,\"family\":\"policy_compliance\",\"input_tokens\":1120,\"reversibility\":0.5096,\"risk\":0.5016,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.178,\"id\":\"worker_a\",\"p95_latency_ms\":20725,\"pass_rate\":0.519,\"provider_group\":\"provider_z\",\"quality\":0.694,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.661,\"id\":\"worker_b\",\"p95_latency_ms\":77775,\"pass_rate\":0.9266,\"provider_group\":\"provider_x\",\"quality\":0.9439,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.161,\"id\":\"worker_c\",\"p95_latency_ms\":76207,\"pass_rate\":0.867,\"provider_group\":\"provider_y\",\"quality\":0.8727,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.839,\"estimated_latency_ms\":98500,\"estimated_pass_rate\":0.6492,\"estimated_quality\":0.8393,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":22.501,\"max_latency_ms\":107874,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.796, "category": "review", "complexity": 0.6402, "consequence": 0.8411, "evidence_need": 0.7973, "family": "policy_compliance", "input_tokens": 1120, "request_id": "synthetic-policy_compliance-001032", "reversibility": 0.5096, "risk": 0.5016, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.178, "id": "worker_a", "p95_latency_ms": 20725, "pass_rate": 0.519, "provider_group": "provider_z", "quality": 0.694, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.661, "id": "worker_b", "p95_latency_ms": 77775, "pass_rate": 0.9266, "provider_group": "provider_x", "quality": 0.9439, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.161, "id": "worker_c", "p95_latency_ms": 76207, "pass_rate": 0.867, "provider_group": "provider_y", "quality": 0.8727, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.608, "max_latency_ms": 199856, "minimum_quality": 0.8962}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000696", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.364, "estimated_latency_ms": 176004, "estimated_pass_rate": 0.7431, "estimated_quality": 0.9062, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.608, "max_latency_ms": 199856, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7431, "quality": 0.9062, "quality_target_met": true, "utility": 80.169444}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.608,\"max_latency_ms\":199856,\"minimum_quality\":0.8962},\"task\":{\"ambiguity\":0.7856,\"category\":\"review\",\"complexity\":0.673,\"consequence\":0.9132,\"evidence_need\":0.6942,\"family\":\"policy_compliance\",\"input_tokens\":7240,\"reversibility\":0.4431,\"risk\":0.7951,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.539,\"id\":\"worker_a\",\"p95_latency_ms\":70731,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.009,\"id\":\"worker_b\",\"p95_latency_ms\":28863,\"pass_rate\":0.5056,\"provider_group\":\"provider_x\",\"quality\":0.6927,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.816,\"id\":\"worker_c\",\"p95_latency_ms\":76410,\"pass_rate\":0.7437,\"provider_group\":\"provider_y\",\"quality\":0.8723,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.364,\"estimated_latency_ms\":176004,\"estimated_pass_rate\":0.7431,\"estimated_quality\":0.9062,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.608,\"max_latency_ms\":199856,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7856, "category": "review", "complexity": 0.673, "consequence": 0.9132, "evidence_need": 0.6942, "family": "policy_compliance", "input_tokens": 7240, "request_id": "synthetic-policy_compliance-000696", "reversibility": 0.4431, "risk": 0.7951, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 13.539, "id": "worker_a", "p95_latency_ms": 70731, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.009, "id": "worker_b", "p95_latency_ms": 28863, "pass_rate": 0.5056, "provider_group": "provider_x", "quality": 0.6927, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.816, "id": "worker_c", "p95_latency_ms": 76410, "pass_rate": 0.7437, "provider_group": "provider_y", "quality": 0.8723, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.698, "max_latency_ms": 86943, "minimum_quality": 0.8085}, "family": "migration_planning", "id": "synthetic-migration_planning-000142", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.18, "estimated_latency_ms": 78157, "estimated_pass_rate": 0.647, "estimated_quality": 0.8278, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.698, "max_latency_ms": 86943, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.647, "quality": 0.8278, "quality_target_met": true, "utility": 63.543158}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.698,\"max_latency_ms\":86943,\"minimum_quality\":0.8085},\"task\":{\"ambiguity\":0.377,\"category\":\"coding\",\"complexity\":0.7713,\"consequence\":0.7578,\"evidence_need\":0.7767,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.3929,\"risk\":0.7214,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":3.069,\"id\":\"worker_a\",\"p95_latency_ms\":49459,\"pass_rate\":0.8532,\"provider_group\":\"provider_y\",\"quality\":0.9078,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.111,\"id\":\"worker_b\",\"p95_latency_ms\":28698,\"pass_rate\":0.5104,\"provider_group\":\"provider_z\",\"quality\":0.7242,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.12,\"id\":\"worker_c\",\"p95_latency_ms\":89905,\"pass_rate\":0.9498,\"provider_group\":\"provider_x\",\"quality\":0.9128,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.18,\"estimated_latency_ms\":78157,\"estimated_pass_rate\":0.647,\"estimated_quality\":0.8278,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.698,\"max_latency_ms\":86943,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.377, "category": "coding", "complexity": 0.7713, "consequence": 0.7578, "evidence_need": 0.7767, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-000142", "reversibility": 0.3929, "risk": 0.7214, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 3.069, "id": "worker_a", "p95_latency_ms": 49459, "pass_rate": 0.8532, "provider_group": "provider_y", "quality": 0.9078, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.111, "id": "worker_b", "p95_latency_ms": 28698, "pass_rate": 0.5104, "provider_group": "provider_z", "quality": 0.7242, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.12, "id": "worker_c", "p95_latency_ms": 89905, "pass_rate": 0.9498, "provider_group": "provider_x", "quality": 0.9128, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.305, "max_latency_ms": 31294, "minimum_quality": 0.5019}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000060", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.184, "estimated_latency_ms": 27150, "estimated_pass_rate": 0.3643, "estimated_quality": 0.6352, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.305, "max_latency_ms": 31294, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3643, "quality": 0.6352, "quality_target_met": true, "utility": 25.067444}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.305,\"max_latency_ms\":31294,\"minimum_quality\":0.5019},\"task\":{\"ambiguity\":0.3964,\"category\":\"review\",\"complexity\":0.524,\"consequence\":0.921,\"evidence_need\":0.8587,\"family\":\"policy_compliance\",\"input_tokens\":1880,\"reversibility\":0.4916,\"risk\":0.6292,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.57,\"id\":\"worker_a\",\"p95_latency_ms\":54850,\"pass_rate\":0.7967,\"provider_group\":\"provider_z\",\"quality\":0.7961,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.567,\"id\":\"worker_b\",\"p95_latency_ms\":78078,\"pass_rate\":0.963,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.184,\"id\":\"worker_c\",\"p95_latency_ms\":27150,\"pass_rate\":0.5047,\"provider_group\":\"provider_y\",\"quality\":0.7219,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.184,\"estimated_latency_ms\":27150,\"estimated_pass_rate\":0.3643,\"estimated_quality\":0.6352,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.305,\"max_latency_ms\":31294,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3964, "category": "review", "complexity": 0.524, "consequence": 0.921, "evidence_need": 0.8587, "family": "policy_compliance", "input_tokens": 1880, "request_id": "synthetic-policy_compliance-000060", "reversibility": 0.4916, "risk": 0.6292, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 3.57, "id": "worker_a", "p95_latency_ms": 54850, "pass_rate": 0.7967, "provider_group": "provider_z", "quality": 0.7961, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.567, "id": "worker_b", "p95_latency_ms": 78078, "pass_rate": 0.963, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.184, "id": "worker_c", "p95_latency_ms": 27150, "pass_rate": 0.5047, "provider_group": "provider_y", "quality": 0.7219, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.648, "max_latency_ms": 144970, "minimum_quality": 0.8012}, "family": "migration_planning", "id": "synthetic-migration_planning-001120", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.824, "estimated_latency_ms": 128847, "estimated_pass_rate": 0.7155, "estimated_quality": 0.9171, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.648, "max_latency_ms": 144970, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7155, "quality": 0.9171, "quality_target_met": true, "utility": 74.564619}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.648,\"max_latency_ms\":144970,\"minimum_quality\":0.8012},\"task\":{\"ambiguity\":0.5621,\"category\":\"coding\",\"complexity\":0.7996,\"consequence\":0.7358,\"evidence_need\":0.7621,\"family\":\"migration_planning\",\"input_tokens\":5040,\"reversibility\":0.3207,\"risk\":0.5606,\"summary\":\"Design compatibility testing for a prompt caching schema migration.\"},\"workers\":[{\"credits_per_call\":3.361,\"id\":\"worker_a\",\"p95_latency_ms\":45680,\"pass_rate\":0.8022,\"provider_group\":\"provider_y\",\"quality\":0.8031,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.035,\"id\":\"worker_b\",\"p95_latency_ms\":25069,\"pass_rate\":0.4855,\"provider_group\":\"provider_x\",\"quality\":0.7494,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.428,\"id\":\"worker_c\",\"p95_latency_ms\":58098,\"pass_rate\":0.9822,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.824,\"estimated_latency_ms\":128847,\"estimated_pass_rate\":0.7155,\"estimated_quality\":0.9171,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.648,\"max_latency_ms\":144970,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5621, "category": "coding", "complexity": 0.7996, "consequence": 0.7358, "evidence_need": 0.7621, "family": "migration_planning", "input_tokens": 5040, "request_id": "synthetic-migration_planning-001120", "reversibility": 0.3207, "risk": 0.5606, "summary": "Design compatibility testing for a prompt caching schema migration."}, "workers": [{"credits_per_call": 3.361, "id": "worker_a", "p95_latency_ms": 45680, "pass_rate": 0.8022, "provider_group": "provider_y", "quality": 0.8031, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.035, "id": "worker_b", "p95_latency_ms": 25069, "pass_rate": 0.4855, "provider_group": "provider_x", "quality": 0.7494, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.428, "id": "worker_c", "p95_latency_ms": 58098, "pass_rate": 0.9822, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.436, "max_latency_ms": 31659, "minimum_quality": 0.6018}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000788", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.104, "estimated_latency_ms": 18799, "estimated_pass_rate": 0.2836, "estimated_quality": 0.512, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.436, "max_latency_ms": 31659, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2836, "quality": 0.512, "quality_target_met": false, "utility": 26.441501}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.436,\"max_latency_ms\":31659,\"minimum_quality\":0.6018},\"task\":{\"ambiguity\":0.6861,\"category\":\"analysis\",\"complexity\":0.5992,\"consequence\":0.4056,\"evidence_need\":0.6813,\"family\":\"performance_tuning\",\"input_tokens\":3400,\"reversibility\":0.6257,\"risk\":0.571,\"summary\":\"Review a proposed optimization for tenant-scoped caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.104,\"id\":\"worker_a\",\"p95_latency_ms\":18799,\"pass_rate\":0.4497,\"provider_group\":\"provider_y\",\"quality\":0.6276,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.325,\"id\":\"worker_b\",\"p95_latency_ms\":91858,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.828,\"id\":\"worker_c\",\"p95_latency_ms\":75101,\"pass_rate\":0.8411,\"provider_group\":\"provider_z\",\"quality\":0.8993,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.104,\"estimated_latency_ms\":18799,\"estimated_pass_rate\":0.2836,\"estimated_quality\":0.512,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.436,\"max_latency_ms\":31659,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6861, "category": "analysis", "complexity": 0.5992, "consequence": 0.4056, "evidence_need": 0.6813, "family": "performance_tuning", "input_tokens": 3400, "request_id": "synthetic-performance_tuning-000788", "reversibility": 0.6257, "risk": 0.571, "summary": "Review a proposed optimization for tenant-scoped caching for hidden regressions."}, "workers": [{"credits_per_call": 1.104, "id": "worker_a", "p95_latency_ms": 18799, "pass_rate": 0.4497, "provider_group": "provider_y", "quality": 0.6276, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.325, "id": "worker_b", "p95_latency_ms": 91858, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.828, "id": "worker_c", "p95_latency_ms": 75101, "pass_rate": 0.8411, "provider_group": "provider_z", "quality": 0.8993, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.53, "max_latency_ms": 259530, "minimum_quality": 0.6969}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001137", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.173, "estimated_latency_ms": 99324, "estimated_pass_rate": 0.6967, "estimated_quality": 0.8346, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.53, "max_latency_ms": 259530, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6967, "quality": 0.8346, "quality_target_met": true, "utility": 65.498688}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.53,\"max_latency_ms\":259530,\"minimum_quality\":0.6969},\"task\":{\"ambiguity\":0.4495,\"category\":\"review\",\"complexity\":0.7253,\"consequence\":0.5628,\"evidence_need\":0.6205,\"family\":\"policy_compliance\",\"input_tokens\":7280,\"reversibility\":0.6645,\"risk\":0.8484,\"summary\":\"Audit a workflow for OAuth callbacks against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.758,\"id\":\"worker_a\",\"p95_latency_ms\":91961,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9401,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.923,\"id\":\"worker_b\",\"p95_latency_ms\":28983,\"pass_rate\":0.5614,\"provider_group\":\"provider_z\",\"quality\":0.7585,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.25,\"id\":\"worker_c\",\"p95_latency_ms\":70341,\"pass_rate\":0.878,\"provider_group\":\"provider_y\",\"quality\":0.9028,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.173,\"estimated_latency_ms\":99324,\"estimated_pass_rate\":0.6967,\"estimated_quality\":0.8346,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.53,\"max_latency_ms\":259530,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4495, "category": "review", "complexity": 0.7253, "consequence": 0.5628, "evidence_need": 0.6205, "family": "policy_compliance", "input_tokens": 7280, "request_id": "synthetic-policy_compliance-001137", "reversibility": 0.6645, "risk": 0.8484, "summary": "Audit a workflow for OAuth callbacks against explicit access requirements."}, "workers": [{"credits_per_call": 13.758, "id": "worker_a", "p95_latency_ms": 91961, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9401, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.923, "id": "worker_b", "p95_latency_ms": 28983, "pass_rate": 0.5614, "provider_group": "provider_z", "quality": 0.7585, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.25, "id": "worker_c", "p95_latency_ms": 70341, "pass_rate": 0.878, "provider_group": "provider_y", "quality": 0.9028, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.454, "max_latency_ms": 170576, "minimum_quality": 0.8242}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000116", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.597, "estimated_latency_ms": 144620, "estimated_pass_rate": 0.8424, "estimated_quality": 0.9416, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.454, "max_latency_ms": 170576, "policy": "strict", "reason_codes": ["independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8424, "quality": 0.9416, "quality_target_met": true, "utility": 71.504294}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.454,\"max_latency_ms\":170576,\"minimum_quality\":0.8242},\"task\":{\"ambiguity\":0.4337,\"category\":\"analysis\",\"complexity\":0.6775,\"consequence\":0.4456,\"evidence_need\":0.3232,\"family\":\"performance_tuning\",\"input_tokens\":5520,\"reversibility\":0.5653,\"risk\":0.4887,\"summary\":\"Diagnose latency variance in tenant-scoped caching and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.545,\"id\":\"worker_a\",\"p95_latency_ms\":59921,\"pass_rate\":0.8394,\"provider_group\":\"provider_z\",\"quality\":0.9428,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.035,\"id\":\"worker_b\",\"p95_latency_ms\":32084,\"pass_rate\":0.5344,\"provider_group\":\"provider_x\",\"quality\":0.7301,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":12.017,\"id\":\"worker_c\",\"p95_latency_ms\":52615,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.597,\"estimated_latency_ms\":144620,\"estimated_pass_rate\":0.8424,\"estimated_quality\":0.9416,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.454,\"max_latency_ms\":170576,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4337, "category": "analysis", "complexity": 0.6775, "consequence": 0.4456, "evidence_need": 0.3232, "family": "performance_tuning", "input_tokens": 5520, "request_id": "synthetic-performance_tuning-000116", "reversibility": 0.5653, "risk": 0.4887, "summary": "Diagnose latency variance in tenant-scoped caching and rank the next experiments."}, "workers": [{"credits_per_call": 3.545, "id": "worker_a", "p95_latency_ms": 59921, "pass_rate": 0.8394, "provider_group": "provider_z", "quality": 0.9428, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.035, "id": "worker_b", "p95_latency_ms": 32084, "pass_rate": 0.5344, "provider_group": "provider_x", "quality": 0.7301, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 12.017, "id": "worker_c", "p95_latency_ms": 52615, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.301, "max_latency_ms": 43164, "minimum_quality": 0.5896}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000524", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.112, "estimated_latency_ms": 29111, "estimated_pass_rate": 0.3775, "estimated_quality": 0.5688, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.301, "max_latency_ms": 43164, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3775, "quality": 0.5688, "quality_target_met": false, "utility": 37.681502}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.301,\"max_latency_ms\":43164,\"minimum_quality\":0.5896},\"task\":{\"ambiguity\":0.598,\"category\":\"analysis\",\"complexity\":0.5909,\"consequence\":0.3922,\"evidence_need\":0.7076,\"family\":\"performance_tuning\",\"input_tokens\":6360,\"reversibility\":0.5862,\"risk\":0.2651,\"summary\":\"Review a proposed optimization for OAuth callbacks for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.112,\"id\":\"worker_a\",\"p95_latency_ms\":29111,\"pass_rate\":0.5112,\"provider_group\":\"provider_x\",\"quality\":0.6549,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.184,\"id\":\"worker_b\",\"p95_latency_ms\":80914,\"pass_rate\":0.7928,\"provider_group\":\"provider_z\",\"quality\":0.8662,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.287,\"id\":\"worker_c\",\"p95_latency_ms\":65830,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9879,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.112,\"estimated_latency_ms\":29111,\"estimated_pass_rate\":0.3775,\"estimated_quality\":0.5688,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.301,\"max_latency_ms\":43164,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.598, "category": "analysis", "complexity": 0.5909, "consequence": 0.3922, "evidence_need": 0.7076, "family": "performance_tuning", "input_tokens": 6360, "request_id": "synthetic-performance_tuning-000524", "reversibility": 0.5862, "risk": 0.2651, "summary": "Review a proposed optimization for OAuth callbacks for hidden regressions."}, "workers": [{"credits_per_call": 1.112, "id": "worker_a", "p95_latency_ms": 29111, "pass_rate": 0.5112, "provider_group": "provider_x", "quality": 0.6549, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.184, "id": "worker_b", "p95_latency_ms": 80914, "pass_rate": 0.7928, "provider_group": "provider_z", "quality": 0.8662, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.287, "id": "worker_c", "p95_latency_ms": 65830, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9879, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.483, "max_latency_ms": 196653, "minimum_quality": 0.803}, "family": "migration_planning", "id": "synthetic-migration_planning-000268", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.021, "estimated_latency_ms": 90615, "estimated_pass_rate": 0.6147, "estimated_quality": 0.7705, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.483, "max_latency_ms": 196653, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6147, "quality": 0.7705, "quality_target_met": false, "utility": 59.196652}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.483,\"max_latency_ms\":196653,\"minimum_quality\":0.803},\"task\":{\"ambiguity\":0.4379,\"category\":\"coding\",\"complexity\":0.9032,\"consequence\":0.6662,\"evidence_need\":0.7876,\"family\":\"migration_planning\",\"input_tokens\":4640,\"reversibility\":0.5244,\"risk\":0.7184,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":10.74,\"id\":\"worker_a\",\"p95_latency_ms\":63477,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9355,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.061,\"id\":\"worker_b\",\"p95_latency_ms\":21922,\"pass_rate\":0.5508,\"provider_group\":\"provider_y\",\"quality\":0.7057,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.96,\"id\":\"worker_c\",\"p95_latency_ms\":68693,\"pass_rate\":0.8347,\"provider_group\":\"provider_z\",\"quality\":0.7967,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.021,\"estimated_latency_ms\":90615,\"estimated_pass_rate\":0.6147,\"estimated_quality\":0.7705,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.483,\"max_latency_ms\":196653,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4379, "category": "coding", "complexity": 0.9032, "consequence": 0.6662, "evidence_need": 0.7876, "family": "migration_planning", "input_tokens": 4640, "request_id": "synthetic-migration_planning-000268", "reversibility": 0.5244, "risk": 0.7184, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 10.74, "id": "worker_a", "p95_latency_ms": 63477, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9355, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.061, "id": "worker_b", "p95_latency_ms": 21922, "pass_rate": 0.5508, "provider_group": "provider_y", "quality": 0.7057, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.96, "id": "worker_c", "p95_latency_ms": 68693, "pass_rate": 0.8347, "provider_group": "provider_z", "quality": 0.7967, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.099, "max_latency_ms": 168897, "minimum_quality": 0.8102}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000233", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.149, "estimated_latency_ms": 151573, "estimated_pass_rate": 0.7269, "estimated_quality": 0.9215, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.099, "max_latency_ms": 168897, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7269, "quality": 0.9215, "quality_target_met": true, "utility": 71.079068}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.099,\"max_latency_ms\":168897,\"minimum_quality\":0.8102},\"task\":{\"ambiguity\":0.6674,\"category\":\"analysis\",\"complexity\":0.8232,\"consequence\":0.59,\"evidence_need\":0.6446,\"family\":\"performance_tuning\",\"input_tokens\":3480,\"reversibility\":0.7957,\"risk\":0.2734,\"summary\":\"Choose between throughput and tail-latency changes for usage metering.\"},\"workers\":[{\"credits_per_call\":4.018,\"id\":\"worker_a\",\"p95_latency_ms\":53981,\"pass_rate\":0.813,\"provider_group\":\"provider_y\",\"quality\":0.9257,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.044,\"id\":\"worker_b\",\"p95_latency_ms\":69420,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.087,\"id\":\"worker_c\",\"p95_latency_ms\":28172,\"pass_rate\":0.5028,\"provider_group\":\"provider_x\",\"quality\":0.7282,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.149,\"estimated_latency_ms\":151573,\"estimated_pass_rate\":0.7269,\"estimated_quality\":0.9215,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.099,\"max_latency_ms\":168897,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6674, "category": "analysis", "complexity": 0.8232, "consequence": 0.59, "evidence_need": 0.6446, "family": "performance_tuning", "input_tokens": 3480, "request_id": "synthetic-performance_tuning-000233", "reversibility": 0.7957, "risk": 0.2734, "summary": "Choose between throughput and tail-latency changes for usage metering."}, "workers": [{"credits_per_call": 4.018, "id": "worker_a", "p95_latency_ms": 53981, "pass_rate": 0.813, "provider_group": "provider_y", "quality": 0.9257, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.044, "id": "worker_b", "p95_latency_ms": 69420, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.087, "id": "worker_c", "p95_latency_ms": 28172, "pass_rate": 0.5028, "provider_group": "provider_x", "quality": 0.7282, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.885, "max_latency_ms": 21596, "minimum_quality": 0.724}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000138", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.014, "estimated_latency_ms": 19575, "estimated_pass_rate": 0.4216, "estimated_quality": 0.5342, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.885, "max_latency_ms": 21596, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4216, "quality": 0.5342, "quality_target_met": false, "utility": 27.001202}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.885,\"max_latency_ms\":21596,\"minimum_quality\":0.724},\"task\":{\"ambiguity\":0.5348,\"category\":\"review\",\"complexity\":0.7972,\"consequence\":0.5101,\"evidence_need\":0.8245,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.4567,\"risk\":0.8295,\"summary\":\"Audit a workflow for provider failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.734,\"id\":\"worker_a\",\"p95_latency_ms\":60962,\"pass_rate\":0.9825,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.208,\"id\":\"worker_b\",\"p95_latency_ms\":63584,\"pass_rate\":0.7354,\"provider_group\":\"provider_z\",\"quality\":0.8906,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.014,\"id\":\"worker_c\",\"p95_latency_ms\":19575,\"pass_rate\":0.6196,\"provider_group\":\"provider_y\",\"quality\":0.694,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.014,\"estimated_latency_ms\":19575,\"estimated_pass_rate\":0.4216,\"estimated_quality\":0.5342,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.885,\"max_latency_ms\":21596,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5348, "category": "review", "complexity": 0.7972, "consequence": 0.5101, "evidence_need": 0.8245, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-000138", "reversibility": 0.4567, "risk": 0.8295, "summary": "Audit a workflow for provider failover against explicit access requirements."}, "workers": [{"credits_per_call": 12.734, "id": "worker_a", "p95_latency_ms": 60962, "pass_rate": 0.9825, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.208, "id": "worker_b", "p95_latency_ms": 63584, "pass_rate": 0.7354, "provider_group": "provider_z", "quality": 0.8906, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.014, "id": "worker_c", "p95_latency_ms": 19575, "pass_rate": 0.6196, "provider_group": "provider_y", "quality": 0.694, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.931, "max_latency_ms": 25060, "minimum_quality": 0.5292}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000288", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.873, "estimated_latency_ms": 23826, "estimated_pass_rate": 0.3891, "estimated_quality": 0.5966, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.931, "max_latency_ms": 25060, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3891, "quality": 0.5966, "quality_target_met": true, "utility": 26.070403}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.931,\"max_latency_ms\":25060,\"minimum_quality\":0.5292},\"task\":{\"ambiguity\":0.7898,\"category\":\"review\",\"complexity\":0.6674,\"consequence\":0.8634,\"evidence_need\":0.7068,\"family\":\"policy_compliance\",\"input_tokens\":6760,\"reversibility\":0.647,\"risk\":0.6688,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.36,\"id\":\"worker_a\",\"p95_latency_ms\":76617,\"pass_rate\":0.9521,\"provider_group\":\"provider_x\",\"quality\":0.955,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.873,\"id\":\"worker_b\",\"p95_latency_ms\":23826,\"pass_rate\":0.5787,\"provider_group\":\"provider_z\",\"quality\":0.7385,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.289,\"id\":\"worker_c\",\"p95_latency_ms\":45089,\"pass_rate\":0.7427,\"provider_group\":\"provider_y\",\"quality\":0.8557,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.873,\"estimated_latency_ms\":23826,\"estimated_pass_rate\":0.3891,\"estimated_quality\":0.5966,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.931,\"max_latency_ms\":25060,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7898, "category": "review", "complexity": 0.6674, "consequence": 0.8634, "evidence_need": 0.7068, "family": "policy_compliance", "input_tokens": 6760, "request_id": "synthetic-policy_compliance-000288", "reversibility": 0.647, "risk": 0.6688, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 11.36, "id": "worker_a", "p95_latency_ms": 76617, "pass_rate": 0.9521, "provider_group": "provider_x", "quality": 0.955, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.873, "id": "worker_b", "p95_latency_ms": 23826, "pass_rate": 0.5787, "provider_group": "provider_z", "quality": 0.7385, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.289, "id": "worker_c", "p95_latency_ms": 45089, "pass_rate": 0.7427, "provider_group": "provider_y", "quality": 0.8557, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.453, "max_latency_ms": 251082, "minimum_quality": 0.8659}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000840", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.964, "estimated_latency_ms": 181899, "estimated_pass_rate": 0.7482, "estimated_quality": 0.8916, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.453, "max_latency_ms": 251082, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7482, "quality": 0.8916, "quality_target_met": true, "utility": 73.020494}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.453,\"max_latency_ms\":251082,\"minimum_quality\":0.8659},\"task\":{\"ambiguity\":0.6947,\"category\":\"review\",\"complexity\":0.4745,\"consequence\":0.5483,\"evidence_need\":0.8724,\"family\":\"policy_compliance\",\"input_tokens\":240,\"reversibility\":0.4513,\"risk\":0.7357,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":15.063,\"id\":\"worker_a\",\"p95_latency_ms\":85970,\"pass_rate\":0.8938,\"provider_group\":\"provider_x\",\"quality\":0.931,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.056,\"id\":\"worker_b\",\"p95_latency_ms\":70748,\"pass_rate\":0.7806,\"provider_group\":\"provider_z\",\"quality\":0.8578,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.845,\"id\":\"worker_c\",\"p95_latency_ms\":25181,\"pass_rate\":0.5909,\"provider_group\":\"provider_y\",\"quality\":0.7277,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.964,\"estimated_latency_ms\":181899,\"estimated_pass_rate\":0.7482,\"estimated_quality\":0.8916,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.453,\"max_latency_ms\":251082,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6947, "category": "review", "complexity": 0.4745, "consequence": 0.5483, "evidence_need": 0.8724, "family": "policy_compliance", "input_tokens": 240, "request_id": "synthetic-policy_compliance-000840", "reversibility": 0.4513, "risk": 0.7357, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 15.063, "id": "worker_a", "p95_latency_ms": 85970, "pass_rate": 0.8938, "provider_group": "provider_x", "quality": 0.931, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.056, "id": "worker_b", "p95_latency_ms": 70748, "pass_rate": 0.7806, "provider_group": "provider_z", "quality": 0.8578, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.845, "id": "worker_c", "p95_latency_ms": 25181, "pass_rate": 0.5909, "provider_group": "provider_y", "quality": 0.7277, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.014, "max_latency_ms": 180704, "minimum_quality": 0.8634}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000414", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.249, "estimated_latency_ms": 163878, "estimated_pass_rate": 0.7298, "estimated_quality": 0.871, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.014, "max_latency_ms": 180704, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7298, "quality": 0.871, "quality_target_met": true, "utility": 70.51366}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.014,\"max_latency_ms\":180704,\"minimum_quality\":0.8634},\"task\":{\"ambiguity\":0.7983,\"category\":\"review\",\"complexity\":0.7989,\"consequence\":0.599,\"evidence_need\":0.5873,\"family\":\"policy_compliance\",\"input_tokens\":1040,\"reversibility\":0.3968,\"risk\":0.459,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.427,\"id\":\"worker_a\",\"p95_latency_ms\":86599,\"pass_rate\":0.9238,\"provider_group\":\"provider_z\",\"quality\":0.924,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.835,\"id\":\"worker_b\",\"p95_latency_ms\":54364,\"pass_rate\":0.8756,\"provider_group\":\"provider_x\",\"quality\":0.8065,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.987,\"id\":\"worker_c\",\"p95_latency_ms\":22915,\"pass_rate\":0.5771,\"provider_group\":\"provider_y\",\"quality\":0.7075,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.249,\"estimated_latency_ms\":163878,\"estimated_pass_rate\":0.7298,\"estimated_quality\":0.871,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.014,\"max_latency_ms\":180704,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7983, "category": "review", "complexity": 0.7989, "consequence": 0.599, "evidence_need": 0.5873, "family": "policy_compliance", "input_tokens": 1040, "request_id": "synthetic-policy_compliance-000414", "reversibility": 0.3968, "risk": 0.459, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.427, "id": "worker_a", "p95_latency_ms": 86599, "pass_rate": 0.9238, "provider_group": "provider_z", "quality": 0.924, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.835, "id": "worker_b", "p95_latency_ms": 54364, "pass_rate": 0.8756, "provider_group": "provider_x", "quality": 0.8065, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.987, "id": "worker_c", "p95_latency_ms": 22915, "pass_rate": 0.5771, "provider_group": "provider_y", "quality": 0.7075, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.559, "max_latency_ms": 232051, "minimum_quality": 0.812}, "family": "migration_planning", "id": "synthetic-migration_planning-000862", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.866, "estimated_latency_ms": 145904, "estimated_pass_rate": 0.8898, "estimated_quality": 0.8964, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.559, "max_latency_ms": 232051, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8898, "quality": 0.8964, "quality_target_met": true, "utility": 70.970596}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.559,\"max_latency_ms\":232051,\"minimum_quality\":0.812},\"task\":{\"ambiguity\":0.4958,\"category\":\"coding\",\"complexity\":0.8536,\"consequence\":0.7813,\"evidence_need\":0.6856,\"family\":\"migration_planning\",\"input_tokens\":7360,\"reversibility\":0.4885,\"risk\":0.6936,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":11.981,\"id\":\"worker_a\",\"p95_latency_ms\":83922,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9415,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.821,\"id\":\"worker_b\",\"p95_latency_ms\":24845,\"pass_rate\":0.5379,\"provider_group\":\"provider_z\",\"quality\":0.7315,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.885,\"id\":\"worker_c\",\"p95_latency_ms\":61982,\"pass_rate\":0.8387,\"provider_group\":\"provider_x\",\"quality\":0.8822,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.866,\"estimated_latency_ms\":145904,\"estimated_pass_rate\":0.8898,\"estimated_quality\":0.8964,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.559,\"max_latency_ms\":232051,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4958, "category": "coding", "complexity": 0.8536, "consequence": 0.7813, "evidence_need": 0.6856, "family": "migration_planning", "input_tokens": 7360, "request_id": "synthetic-migration_planning-000862", "reversibility": 0.4885, "risk": 0.6936, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 11.981, "id": "worker_a", "p95_latency_ms": 83922, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9415, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.821, "id": "worker_b", "p95_latency_ms": 24845, "pass_rate": 0.5379, "provider_group": "provider_z", "quality": 0.7315, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.885, "id": "worker_c", "p95_latency_ms": 61982, "pass_rate": 0.8387, "provider_group": "provider_x", "quality": 0.8822, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.588, "max_latency_ms": 184509, "minimum_quality": 0.8802}, "family": "migration_planning", "id": "synthetic-migration_planning-000685", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.798, "estimated_latency_ms": 173237, "estimated_pass_rate": 0.749, "estimated_quality": 0.8925, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.588, "max_latency_ms": 184509, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.749, "quality": 0.8925, "quality_target_met": true, "utility": 70.073001}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.588,\"max_latency_ms\":184509,\"minimum_quality\":0.8802},\"task\":{\"ambiguity\":0.5842,\"category\":\"coding\",\"complexity\":0.6023,\"consequence\":0.534,\"evidence_need\":0.6011,\"family\":\"migration_planning\",\"input_tokens\":4920,\"reversibility\":0.5651,\"risk\":0.6366,\"summary\":\"Design compatibility testing for a credit reservation schema migration.\"},\"workers\":[{\"credits_per_call\":10.116,\"id\":\"worker_a\",\"p95_latency_ms\":88384,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9636,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.793,\"id\":\"worker_b\",\"p95_latency_ms\":26865,\"pass_rate\":0.5251,\"provider_group\":\"provider_x\",\"quality\":0.6998,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.889,\"id\":\"worker_c\",\"p95_latency_ms\":57988,\"pass_rate\":0.7769,\"provider_group\":\"provider_y\",\"quality\":0.8317,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.798,\"estimated_latency_ms\":173237,\"estimated_pass_rate\":0.749,\"estimated_quality\":0.8925,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.588,\"max_latency_ms\":184509,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5842, "category": "coding", "complexity": 0.6023, "consequence": 0.534, "evidence_need": 0.6011, "family": "migration_planning", "input_tokens": 4920, "request_id": "synthetic-migration_planning-000685", "reversibility": 0.5651, "risk": 0.6366, "summary": "Design compatibility testing for a credit reservation schema migration."}, "workers": [{"credits_per_call": 10.116, "id": "worker_a", "p95_latency_ms": 88384, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9636, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.793, "id": "worker_b", "p95_latency_ms": 26865, "pass_rate": 0.5251, "provider_group": "provider_x", "quality": 0.6998, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.889, "id": "worker_c", "p95_latency_ms": 57988, "pass_rate": 0.7769, "provider_group": "provider_y", "quality": 0.8317, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.487, "max_latency_ms": 262682, "minimum_quality": 0.91}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000558", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.914, "estimated_latency_ms": 188792, "estimated_pass_rate": 0.7675, "estimated_quality": 0.8857, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.487, "max_latency_ms": 262682, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7675, "quality": 0.8857, "quality_target_met": false, "utility": 73.612234}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.487,\"max_latency_ms\":262682,\"minimum_quality\":0.91},\"task\":{\"ambiguity\":0.386,\"category\":\"review\",\"complexity\":0.7314,\"consequence\":0.8762,\"evidence_need\":0.587,\"family\":\"policy_compliance\",\"input_tokens\":2320,\"reversibility\":0.4557,\"risk\":0.6992,\"summary\":\"Identify evidence required before approving API-key rotation for production.\"},\"workers\":[{\"credits_per_call\":0.959,\"id\":\"worker_a\",\"p95_latency_ms\":30295,\"pass_rate\":0.556,\"provider_group\":\"provider_y\",\"quality\":0.791,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.569,\"id\":\"worker_b\",\"p95_latency_ms\":68960,\"pass_rate\":0.7928,\"provider_group\":\"provider_x\",\"quality\":0.832,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.386,\"id\":\"worker_c\",\"p95_latency_ms\":89537,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9156,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.914,\"estimated_latency_ms\":188792,\"estimated_pass_rate\":0.7675,\"estimated_quality\":0.8857,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.487,\"max_latency_ms\":262682,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.386, "category": "review", "complexity": 0.7314, "consequence": 0.8762, "evidence_need": 0.587, "family": "policy_compliance", "input_tokens": 2320, "request_id": "synthetic-policy_compliance-000558", "reversibility": 0.4557, "risk": 0.6992, "summary": "Identify evidence required before approving API-key rotation for production."}, "workers": [{"credits_per_call": 0.959, "id": "worker_a", "p95_latency_ms": 30295, "pass_rate": 0.556, "provider_group": "provider_y", "quality": 0.791, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.569, "id": "worker_b", "p95_latency_ms": 68960, "pass_rate": 0.7928, "provider_group": "provider_x", "quality": 0.832, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.386, "id": "worker_c", "p95_latency_ms": 89537, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9156, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.645, "max_latency_ms": 136716, "minimum_quality": 0.6833}, "family": "migration_planning", "id": "synthetic-migration_planning-000400", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.027, "estimated_latency_ms": 71188, "estimated_pass_rate": 0.5593, "estimated_quality": 0.7761, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.645, "max_latency_ms": 136716, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.5593, "quality": 0.7761, "quality_target_met": true, "utility": 59.787209}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.645,\"max_latency_ms\":136716,\"minimum_quality\":0.6833},\"task\":{\"ambiguity\":0.5765,\"category\":\"coding\",\"complexity\":0.9357,\"consequence\":0.8294,\"evidence_need\":0.5437,\"family\":\"migration_planning\",\"input_tokens\":200,\"reversibility\":0.5025,\"risk\":0.6609,\"summary\":\"Plan a zero-downtime migration of schema evolution with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.193,\"id\":\"worker_a\",\"p95_latency_ms\":49113,\"pass_rate\":0.7364,\"provider_group\":\"provider_x\",\"quality\":0.8824,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.87,\"id\":\"worker_b\",\"p95_latency_ms\":83158,\"pass_rate\":0.8852,\"provider_group\":\"provider_y\",\"quality\":0.918,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.834,\"id\":\"worker_c\",\"p95_latency_ms\":22075,\"pass_rate\":0.4934,\"provider_group\":\"provider_z\",\"quality\":0.7903,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.027,\"estimated_latency_ms\":71188,\"estimated_pass_rate\":0.5593,\"estimated_quality\":0.7761,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.645,\"max_latency_ms\":136716,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5765, "category": "coding", "complexity": 0.9357, "consequence": 0.8294, "evidence_need": 0.5437, "family": "migration_planning", "input_tokens": 200, "request_id": "synthetic-migration_planning-000400", "reversibility": 0.5025, "risk": 0.6609, "summary": "Plan a zero-downtime migration of schema evolution with rollback checkpoints."}, "workers": [{"credits_per_call": 4.193, "id": "worker_a", "p95_latency_ms": 49113, "pass_rate": 0.7364, "provider_group": "provider_x", "quality": 0.8824, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.87, "id": "worker_b", "p95_latency_ms": 83158, "pass_rate": 0.8852, "provider_group": "provider_y", "quality": 0.918, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.834, "id": "worker_c", "p95_latency_ms": 22075, "pass_rate": 0.4934, "provider_group": "provider_z", "quality": 0.7903, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.657, "max_latency_ms": 156539, "minimum_quality": 0.7541}, "family": "migration_planning", "id": "synthetic-migration_planning-001147", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.782, "estimated_latency_ms": 79519, "estimated_pass_rate": 0.6364, "estimated_quality": 0.8038, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.657, "max_latency_ms": 156539, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6364, "quality": 0.8038, "quality_target_met": true, "utility": 61.08634}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.657,\"max_latency_ms\":156539,\"minimum_quality\":0.7541},\"task\":{\"ambiguity\":0.4038,\"category\":\"coding\",\"complexity\":0.814,\"consequence\":0.4552,\"evidence_need\":0.4227,\"family\":\"migration_planning\",\"input_tokens\":1920,\"reversibility\":0.6007,\"risk\":0.6161,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":11.137,\"id\":\"worker_a\",\"p95_latency_ms\":75743,\"pass_rate\":0.8955,\"provider_group\":\"provider_x\",\"quality\":0.9659,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.99,\"id\":\"worker_b\",\"p95_latency_ms\":59078,\"pass_rate\":0.8666,\"provider_group\":\"provider_y\",\"quality\":0.8961,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.792,\"id\":\"worker_c\",\"p95_latency_ms\":20441,\"pass_rate\":0.5214,\"provider_group\":\"provider_z\",\"quality\":0.7738,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.782,\"estimated_latency_ms\":79519,\"estimated_pass_rate\":0.6364,\"estimated_quality\":0.8038,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.657,\"max_latency_ms\":156539,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4038, "category": "coding", "complexity": 0.814, "consequence": 0.4552, "evidence_need": 0.4227, "family": "migration_planning", "input_tokens": 1920, "request_id": "synthetic-migration_planning-001147", "reversibility": 0.6007, "risk": 0.6161, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 11.137, "id": "worker_a", "p95_latency_ms": 75743, "pass_rate": 0.8955, "provider_group": "provider_x", "quality": 0.9659, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.99, "id": "worker_b", "p95_latency_ms": 59078, "pass_rate": 0.8666, "provider_group": "provider_y", "quality": 0.8961, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.792, "id": "worker_c", "p95_latency_ms": 20441, "pass_rate": 0.5214, "provider_group": "provider_z", "quality": 0.7738, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.331, "max_latency_ms": 35687, "minimum_quality": 0.657}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000225", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.958, "estimated_latency_ms": 30900, "estimated_pass_rate": 0.3732, "estimated_quality": 0.5912, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.331, "max_latency_ms": 35687, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3732, "quality": 0.5912, "quality_target_met": false, "utility": 16.008796}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.331,\"max_latency_ms\":35687,\"minimum_quality\":0.657},\"task\":{\"ambiguity\":0.7907,\"category\":\"review\",\"complexity\":0.802,\"consequence\":0.8606,\"evidence_need\":0.9435,\"family\":\"policy_compliance\",\"input_tokens\":320,\"reversibility\":0.6971,\"risk\":0.7465,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.958,\"id\":\"worker_a\",\"p95_latency_ms\":30900,\"pass_rate\":0.5844,\"provider_group\":\"provider_y\",\"quality\":0.7631,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.467,\"id\":\"worker_b\",\"p95_latency_ms\":58254,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.918,\"id\":\"worker_c\",\"p95_latency_ms\":70692,\"pass_rate\":0.8353,\"provider_group\":\"provider_z\",\"quality\":0.838,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.958,\"estimated_latency_ms\":30900,\"estimated_pass_rate\":0.3732,\"estimated_quality\":0.5912,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.331,\"max_latency_ms\":35687,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7907, "category": "review", "complexity": 0.802, "consequence": 0.8606, "evidence_need": 0.9435, "family": "policy_compliance", "input_tokens": 320, "request_id": "synthetic-policy_compliance-000225", "reversibility": 0.6971, "risk": 0.7465, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 0.958, "id": "worker_a", "p95_latency_ms": 30900, "pass_rate": 0.5844, "provider_group": "provider_y", "quality": 0.7631, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.467, "id": "worker_b", "p95_latency_ms": 58254, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.918, "id": "worker_c", "p95_latency_ms": 70692, "pass_rate": 0.8353, "provider_group": "provider_z", "quality": 0.838, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 4.45, "max_latency_ms": 191960, "minimum_quality": 0.7486}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000921", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.744, "estimated_latency_ms": 78724, "estimated_pass_rate": 0.6687, "estimated_quality": 0.8199, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.45, "max_latency_ms": 191960, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6687, "quality": 0.8199, "quality_target_met": true, "utility": 59.498852}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":4.45,\"max_latency_ms\":191960,\"minimum_quality\":0.7486},\"task\":{\"ambiguity\":0.3893,\"category\":\"review\",\"complexity\":0.6748,\"consequence\":0.8037,\"evidence_need\":0.6676,\"family\":\"policy_compliance\",\"input_tokens\":3840,\"reversibility\":0.6462,\"risk\":0.6525,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":11.624,\"id\":\"worker_a\",\"p95_latency_ms\":69390,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9277,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.895,\"id\":\"worker_b\",\"p95_latency_ms\":26315,\"pass_rate\":0.554,\"provider_group\":\"provider_y\",\"quality\":0.7013,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.849,\"id\":\"worker_c\",\"p95_latency_ms\":52409,\"pass_rate\":0.7363,\"provider_group\":\"provider_x\",\"quality\":0.8835,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.744,\"estimated_latency_ms\":78724,\"estimated_pass_rate\":0.6687,\"estimated_quality\":0.8199,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.45,\"max_latency_ms\":191960,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3893, "category": "review", "complexity": 0.6748, "consequence": 0.8037, "evidence_need": 0.6676, "family": "policy_compliance", "input_tokens": 3840, "request_id": "synthetic-policy_compliance-000921", "reversibility": 0.6462, "risk": 0.6525, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 11.624, "id": "worker_a", "p95_latency_ms": 69390, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9277, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.895, "id": "worker_b", "p95_latency_ms": 26315, "pass_rate": 0.554, "provider_group": "provider_y", "quality": 0.7013, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.849, "id": "worker_c", "p95_latency_ms": 52409, "pass_rate": 0.7363, "provider_group": "provider_x", "quality": 0.8835, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.135, "max_latency_ms": 41482, "minimum_quality": 0.5359}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000680", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.919, "estimated_latency_ms": 31724, "estimated_pass_rate": 0.3732, "estimated_quality": 0.5951, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.135, "max_latency_ms": 41482, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3732, "quality": 0.5951, "quality_target_met": true, "utility": 36.370514}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.135,\"max_latency_ms\":41482,\"minimum_quality\":0.5359},\"task\":{\"ambiguity\":0.3551,\"category\":\"analysis\",\"complexity\":0.5056,\"consequence\":0.5223,\"evidence_need\":0.3711,\"family\":\"performance_tuning\",\"input_tokens\":4760,\"reversibility\":0.6349,\"risk\":0.4324,\"summary\":\"Diagnose latency variance in deployment rollback and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.919,\"id\":\"worker_a\",\"p95_latency_ms\":31724,\"pass_rate\":0.4924,\"provider_group\":\"provider_z\",\"quality\":0.6615,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.743,\"id\":\"worker_b\",\"p95_latency_ms\":58784,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.265,\"id\":\"worker_c\",\"p95_latency_ms\":56594,\"pass_rate\":0.7867,\"provider_group\":\"provider_x\",\"quality\":0.9865,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.919,\"estimated_latency_ms\":31724,\"estimated_pass_rate\":0.3732,\"estimated_quality\":0.5951,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.135,\"max_latency_ms\":41482,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3551, "category": "analysis", "complexity": 0.5056, "consequence": 0.5223, "evidence_need": 0.3711, "family": "performance_tuning", "input_tokens": 4760, "request_id": "synthetic-performance_tuning-000680", "reversibility": 0.6349, "risk": 0.4324, "summary": "Diagnose latency variance in deployment rollback and rank the next experiments."}, "workers": [{"credits_per_call": 0.919, "id": "worker_a", "p95_latency_ms": 31724, "pass_rate": 0.4924, "provider_group": "provider_z", "quality": 0.6615, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.743, "id": "worker_b", "p95_latency_ms": 58784, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.265, "id": "worker_c", "p95_latency_ms": 56594, "pass_rate": 0.7867, "provider_group": "provider_x", "quality": 0.9865, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.237, "max_latency_ms": 34836, "minimum_quality": 0.5402}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000086", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.936, "estimated_latency_ms": 27374, "estimated_pass_rate": 0.3703, "estimated_quality": 0.659, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.237, "max_latency_ms": 34836, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3703, "quality": 0.659, "quality_target_met": true, "utility": 40.542131}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.237,\"max_latency_ms\":34836,\"minimum_quality\":0.5402},\"task\":{\"ambiguity\":0.334,\"category\":\"analysis\",\"complexity\":0.5073,\"consequence\":0.5634,\"evidence_need\":0.4808,\"family\":\"performance_tuning\",\"input_tokens\":2280,\"reversibility\":0.8136,\"risk\":0.5205,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":11.125,\"id\":\"worker_a\",\"p95_latency_ms\":84298,\"pass_rate\":0.9574,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.936,\"id\":\"worker_b\",\"p95_latency_ms\":27374,\"pass_rate\":0.4952,\"provider_group\":\"provider_x\",\"quality\":0.7306,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.994,\"id\":\"worker_c\",\"p95_latency_ms\":52740,\"pass_rate\":0.8588,\"provider_group\":\"provider_y\",\"quality\":0.9498,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.936,\"estimated_latency_ms\":27374,\"estimated_pass_rate\":0.3703,\"estimated_quality\":0.659,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.237,\"max_latency_ms\":34836,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.334, "category": "analysis", "complexity": 0.5073, "consequence": 0.5634, "evidence_need": 0.4808, "family": "performance_tuning", "input_tokens": 2280, "request_id": "synthetic-performance_tuning-000086", "reversibility": 0.8136, "risk": 0.5205, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 11.125, "id": "worker_a", "p95_latency_ms": 84298, "pass_rate": 0.9574, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.936, "id": "worker_b", "p95_latency_ms": 27374, "pass_rate": 0.4952, "provider_group": "provider_x", "quality": 0.7306, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.994, "id": "worker_c", "p95_latency_ms": 52740, "pass_rate": 0.8588, "provider_group": "provider_y", "quality": 0.9498, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.636, "max_latency_ms": 244098, "minimum_quality": 0.7936}, "family": "migration_planning", "id": "synthetic-migration_planning-000730", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.689, "estimated_latency_ms": 104797, "estimated_pass_rate": 0.6519, "estimated_quality": 0.7595, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.636, "max_latency_ms": 244098, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6519, "quality": 0.7595, "quality_target_met": false, "utility": 52.133218}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.636,\"max_latency_ms\":244098,\"minimum_quality\":0.7936},\"task\":{\"ambiguity\":0.6086,\"category\":\"coding\",\"complexity\":0.7983,\"consequence\":0.7718,\"evidence_need\":0.754,\"family\":\"migration_planning\",\"input_tokens\":4880,\"reversibility\":0.5212,\"risk\":0.4152,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.209,\"id\":\"worker_a\",\"p95_latency_ms\":31910,\"pass_rate\":0.4585,\"provider_group\":\"provider_x\",\"quality\":0.6823,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.258,\"id\":\"worker_b\",\"p95_latency_ms\":53322,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.48,\"id\":\"worker_c\",\"p95_latency_ms\":72887,\"pass_rate\":0.7651,\"provider_group\":\"provider_z\",\"quality\":0.8492,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.689,\"estimated_latency_ms\":104797,\"estimated_pass_rate\":0.6519,\"estimated_quality\":0.7595,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.636,\"max_latency_ms\":244098,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6086, "category": "coding", "complexity": 0.7983, "consequence": 0.7718, "evidence_need": 0.754, "family": "migration_planning", "input_tokens": 4880, "request_id": "synthetic-migration_planning-000730", "reversibility": 0.5212, "risk": 0.4152, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 1.209, "id": "worker_a", "p95_latency_ms": 31910, "pass_rate": 0.4585, "provider_group": "provider_x", "quality": 0.6823, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.258, "id": "worker_b", "p95_latency_ms": 53322, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.48, "id": "worker_c", "p95_latency_ms": 72887, "pass_rate": 0.7651, "provider_group": "provider_z", "quality": 0.8492, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.436, "max_latency_ms": 207663, "minimum_quality": 0.896}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000497", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.15, "estimated_latency_ms": 164902, "estimated_pass_rate": 0.7608, "estimated_quality": 0.9109, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.436, "max_latency_ms": 207663, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7608, "quality": 0.9109, "quality_target_met": true, "utility": 69.451095}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.436,\"max_latency_ms\":207663,\"minimum_quality\":0.896},\"task\":{\"ambiguity\":0.7157,\"category\":\"analysis\",\"complexity\":0.5745,\"consequence\":0.261,\"evidence_need\":0.4576,\"family\":\"performance_tuning\",\"input_tokens\":1120,\"reversibility\":0.7971,\"risk\":0.5086,\"summary\":\"Diagnose latency variance in streaming inference and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.364,\"id\":\"worker_a\",\"p95_latency_ms\":55096,\"pass_rate\":0.7924,\"provider_group\":\"provider_x\",\"quality\":0.9135,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.824,\"id\":\"worker_b\",\"p95_latency_ms\":21422,\"pass_rate\":0.5387,\"provider_group\":\"provider_z\",\"quality\":0.7011,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.962,\"id\":\"worker_c\",\"p95_latency_ms\":88384,\"pass_rate\":0.9943,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.15,\"estimated_latency_ms\":164902,\"estimated_pass_rate\":0.7608,\"estimated_quality\":0.9109,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.436,\"max_latency_ms\":207663,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7157, "category": "analysis", "complexity": 0.5745, "consequence": 0.261, "evidence_need": 0.4576, "family": "performance_tuning", "input_tokens": 1120, "request_id": "synthetic-performance_tuning-000497", "reversibility": 0.7971, "risk": 0.5086, "summary": "Diagnose latency variance in streaming inference and rank the next experiments."}, "workers": [{"credits_per_call": 3.364, "id": "worker_a", "p95_latency_ms": 55096, "pass_rate": 0.7924, "provider_group": "provider_x", "quality": 0.9135, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.824, "id": "worker_b", "p95_latency_ms": 21422, "pass_rate": 0.5387, "provider_group": "provider_z", "quality": 0.7011, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.962, "id": "worker_c", "p95_latency_ms": 88384, "pass_rate": 0.9943, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.088, "max_latency_ms": 175104, "minimum_quality": 0.7984}, "family": "migration_planning", "id": "synthetic-migration_planning-001261", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.41, "estimated_latency_ms": 104570, "estimated_pass_rate": 0.6505, "estimated_quality": 0.7778, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.088, "max_latency_ms": 175104, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6505, "quality": 0.7778, "quality_target_met": false, "utility": 59.541412}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.088,\"max_latency_ms\":175104,\"minimum_quality\":0.7984},\"task\":{\"ambiguity\":0.4808,\"category\":\"coding\",\"complexity\":0.8799,\"consequence\":0.5347,\"evidence_need\":0.4832,\"family\":\"migration_planning\",\"input_tokens\":6680,\"reversibility\":0.359,\"risk\":0.7311,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":0.93,\"id\":\"worker_a\",\"p95_latency_ms\":32670,\"pass_rate\":0.5336,\"provider_group\":\"provider_z\",\"quality\":0.721,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.34,\"id\":\"worker_b\",\"p95_latency_ms\":50178,\"pass_rate\":0.9603,\"provider_group\":\"provider_y\",\"quality\":0.9106,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.48,\"id\":\"worker_c\",\"p95_latency_ms\":71900,\"pass_rate\":0.8849,\"provider_group\":\"provider_x\",\"quality\":0.8769,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.41,\"estimated_latency_ms\":104570,\"estimated_pass_rate\":0.6505,\"estimated_quality\":0.7778,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.088,\"max_latency_ms\":175104,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4808, "category": "coding", "complexity": 0.8799, "consequence": 0.5347, "evidence_need": 0.4832, "family": "migration_planning", "input_tokens": 6680, "request_id": "synthetic-migration_planning-001261", "reversibility": 0.359, "risk": 0.7311, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 0.93, "id": "worker_a", "p95_latency_ms": 32670, "pass_rate": 0.5336, "provider_group": "provider_z", "quality": 0.721, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.34, "id": "worker_b", "p95_latency_ms": 50178, "pass_rate": 0.9603, "provider_group": "provider_y", "quality": 0.9106, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.48, "id": "worker_c", "p95_latency_ms": 71900, "pass_rate": 0.8849, "provider_group": "provider_x", "quality": 0.8769, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.714, "max_latency_ms": 49290, "minimum_quality": 0.7629}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000039", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.144, "estimated_latency_ms": 26456, "estimated_pass_rate": 0.3348, "estimated_quality": 0.6137, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.714, "max_latency_ms": 49290, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3348, "quality": 0.6137, "quality_target_met": false, "utility": 32.39595}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.714,\"max_latency_ms\":49290,\"minimum_quality\":0.7629},\"task\":{\"ambiguity\":0.6696,\"category\":\"review\",\"complexity\":0.6136,\"consequence\":0.5233,\"evidence_need\":0.8262,\"family\":\"policy_compliance\",\"input_tokens\":1360,\"reversibility\":0.5566,\"risk\":0.5339,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":3.362,\"id\":\"worker_a\",\"p95_latency_ms\":69680,\"pass_rate\":0.7346,\"provider_group\":\"provider_z\",\"quality\":0.8018,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.95,\"id\":\"worker_b\",\"p95_latency_ms\":71861,\"pass_rate\":0.9286,\"provider_group\":\"provider_y\",\"quality\":0.9918,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.144,\"id\":\"worker_c\",\"p95_latency_ms\":26456,\"pass_rate\":0.4982,\"provider_group\":\"provider_x\",\"quality\":0.7281,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.144,\"estimated_latency_ms\":26456,\"estimated_pass_rate\":0.3348,\"estimated_quality\":0.6137,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.714,\"max_latency_ms\":49290,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6696, "category": "review", "complexity": 0.6136, "consequence": 0.5233, "evidence_need": 0.8262, "family": "policy_compliance", "input_tokens": 1360, "request_id": "synthetic-policy_compliance-000039", "reversibility": 0.5566, "risk": 0.5339, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 3.362, "id": "worker_a", "p95_latency_ms": 69680, "pass_rate": 0.7346, "provider_group": "provider_z", "quality": 0.8018, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.95, "id": "worker_b", "p95_latency_ms": 71861, "pass_rate": 0.9286, "provider_group": "provider_y", "quality": 0.9918, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.144, "id": "worker_c", "p95_latency_ms": 26456, "pass_rate": 0.4982, "provider_group": "provider_x", "quality": 0.7281, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.837, "max_latency_ms": 127114, "minimum_quality": 0.7814}, "family": "migration_planning", "id": "synthetic-migration_planning-000850", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.253, "estimated_latency_ms": 81140, "estimated_pass_rate": 0.6675, "estimated_quality": 0.797, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.837, "max_latency_ms": 127114, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6675, "quality": 0.797, "quality_target_met": true, "utility": 59.339245}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.837,\"max_latency_ms\":127114,\"minimum_quality\":0.7814},\"task\":{\"ambiguity\":0.5539,\"category\":\"coding\",\"complexity\":0.6689,\"consequence\":0.8121,\"evidence_need\":0.6748,\"family\":\"migration_planning\",\"input_tokens\":6160,\"reversibility\":0.6236,\"risk\":0.6739,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.999,\"id\":\"worker_a\",\"p95_latency_ms\":28429,\"pass_rate\":0.5842,\"provider_group\":\"provider_z\",\"quality\":0.7225,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.045,\"id\":\"worker_b\",\"p95_latency_ms\":60107,\"pass_rate\":0.9222,\"provider_group\":\"provider_y\",\"quality\":0.9157,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.254,\"id\":\"worker_c\",\"p95_latency_ms\":52711,\"pass_rate\":0.7351,\"provider_group\":\"provider_x\",\"quality\":0.8523,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.253,\"estimated_latency_ms\":81140,\"estimated_pass_rate\":0.6675,\"estimated_quality\":0.797,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.837,\"max_latency_ms\":127114,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5539, "category": "coding", "complexity": 0.6689, "consequence": 0.8121, "evidence_need": 0.6748, "family": "migration_planning", "input_tokens": 6160, "request_id": "synthetic-migration_planning-000850", "reversibility": 0.6236, "risk": 0.6739, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 0.999, "id": "worker_a", "p95_latency_ms": 28429, "pass_rate": 0.5842, "provider_group": "provider_z", "quality": 0.7225, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.045, "id": "worker_b", "p95_latency_ms": 60107, "pass_rate": 0.9222, "provider_group": "provider_y", "quality": 0.9157, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.254, "id": "worker_c", "p95_latency_ms": 52711, "pass_rate": 0.7351, "provider_group": "provider_x", "quality": 0.8523, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.816, "max_latency_ms": 35674, "minimum_quality": 0.5663}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000728", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.822, "estimated_latency_ms": 33088, "estimated_pass_rate": 0.2753, "estimated_quality": 0.5934, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.816, "max_latency_ms": 35674, "policy": "lean", "reason_codes": ["independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2753, "quality": 0.5934, "quality_target_met": true, "utility": 37.441042}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.816,\"max_latency_ms\":35674,\"minimum_quality\":0.5663},\"task\":{\"ambiguity\":0.456,\"category\":\"analysis\",\"complexity\":0.6164,\"consequence\":0.4251,\"evidence_need\":0.6765,\"family\":\"performance_tuning\",\"input_tokens\":7600,\"reversibility\":0.8686,\"risk\":0.2826,\"summary\":\"Review a proposed optimization for API-key rotation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":0.822,\"id\":\"worker_a\",\"p95_latency_ms\":33088,\"pass_rate\":0.4026,\"provider_group\":\"provider_y\",\"quality\":0.6758,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.506,\"id\":\"worker_b\",\"p95_latency_ms\":61878,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.167,\"id\":\"worker_c\",\"p95_latency_ms\":65034,\"pass_rate\":0.7981,\"provider_group\":\"provider_x\",\"quality\":0.8704,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.822,\"estimated_latency_ms\":33088,\"estimated_pass_rate\":0.2753,\"estimated_quality\":0.5934,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.816,\"max_latency_ms\":35674,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.456, "category": "analysis", "complexity": 0.6164, "consequence": 0.4251, "evidence_need": 0.6765, "family": "performance_tuning", "input_tokens": 7600, "request_id": "synthetic-performance_tuning-000728", "reversibility": 0.8686, "risk": 0.2826, "summary": "Review a proposed optimization for API-key rotation for hidden regressions."}, "workers": [{"credits_per_call": 0.822, "id": "worker_a", "p95_latency_ms": 33088, "pass_rate": 0.4026, "provider_group": "provider_y", "quality": 0.6758, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.506, "id": "worker_b", "p95_latency_ms": 61878, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.167, "id": "worker_c", "p95_latency_ms": 65034, "pass_rate": 0.7981, "provider_group": "provider_x", "quality": 0.8704, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.827, "max_latency_ms": 224524, "minimum_quality": 0.7976}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000312", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.46, "estimated_latency_ms": 117468, "estimated_pass_rate": 0.8907, "estimated_quality": 0.9328, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.827, "max_latency_ms": 224524, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8907, "quality": 0.9328, "quality_target_met": true, "utility": 74.984916}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.827,\"max_latency_ms\":224524,\"minimum_quality\":0.7976},\"task\":{\"ambiguity\":0.4087,\"category\":\"review\",\"complexity\":0.4162,\"consequence\":0.8089,\"evidence_need\":0.602,\"family\":\"policy_compliance\",\"input_tokens\":3560,\"reversibility\":0.4854,\"risk\":0.7945,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.764,\"id\":\"worker_a\",\"p95_latency_ms\":66927,\"pass_rate\":0.8625,\"provider_group\":\"provider_z\",\"quality\":0.8325,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.696,\"id\":\"worker_b\",\"p95_latency_ms\":50541,\"pass_rate\":0.9447,\"provider_group\":\"provider_x\",\"quality\":0.9487,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.197,\"id\":\"worker_c\",\"p95_latency_ms\":29456,\"pass_rate\":0.4821,\"provider_group\":\"provider_y\",\"quality\":0.6854,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.46,\"estimated_latency_ms\":117468,\"estimated_pass_rate\":0.8907,\"estimated_quality\":0.9328,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.827,\"max_latency_ms\":224524,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4087, "category": "review", "complexity": 0.4162, "consequence": 0.8089, "evidence_need": 0.602, "family": "policy_compliance", "input_tokens": 3560, "request_id": "synthetic-policy_compliance-000312", "reversibility": 0.4854, "risk": 0.7945, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 2.764, "id": "worker_a", "p95_latency_ms": 66927, "pass_rate": 0.8625, "provider_group": "provider_z", "quality": 0.8325, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.696, "id": "worker_b", "p95_latency_ms": 50541, "pass_rate": 0.9447, "provider_group": "provider_x", "quality": 0.9487, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.197, "id": "worker_c", "p95_latency_ms": 29456, "pass_rate": 0.4821, "provider_group": "provider_y", "quality": 0.6854, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.148, "max_latency_ms": 27274, "minimum_quality": 0.7303}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000231", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.973, "estimated_latency_ms": 23118, "estimated_pass_rate": 0.3589, "estimated_quality": 0.6079, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.148, "max_latency_ms": 27274, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3589, "quality": 0.6079, "quality_target_met": false, "utility": 33.818551}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.148,\"max_latency_ms\":27274,\"minimum_quality\":0.7303},\"task\":{\"ambiguity\":0.6487,\"category\":\"review\",\"complexity\":0.6519,\"consequence\":0.511,\"evidence_need\":0.775,\"family\":\"policy_compliance\",\"input_tokens\":4520,\"reversibility\":0.5849,\"risk\":0.7269,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.687,\"id\":\"worker_a\",\"p95_latency_ms\":70373,\"pass_rate\":0.7431,\"provider_group\":\"provider_y\",\"quality\":0.8609,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.973,\"id\":\"worker_b\",\"p95_latency_ms\":23118,\"pass_rate\":0.5407,\"provider_group\":\"provider_x\",\"quality\":0.7417,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.835,\"id\":\"worker_c\",\"p95_latency_ms\":87413,\"pass_rate\":0.9814,\"provider_group\":\"provider_z\",\"quality\":0.9915,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.973,\"estimated_latency_ms\":23118,\"estimated_pass_rate\":0.3589,\"estimated_quality\":0.6079,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.148,\"max_latency_ms\":27274,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6487, "category": "review", "complexity": 0.6519, "consequence": 0.511, "evidence_need": 0.775, "family": "policy_compliance", "input_tokens": 4520, "request_id": "synthetic-policy_compliance-000231", "reversibility": 0.5849, "risk": 0.7269, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.687, "id": "worker_a", "p95_latency_ms": 70373, "pass_rate": 0.7431, "provider_group": "provider_y", "quality": 0.8609, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.973, "id": "worker_b", "p95_latency_ms": 23118, "pass_rate": 0.5407, "provider_group": "provider_x", "quality": 0.7417, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.835, "id": "worker_c", "p95_latency_ms": 87413, "pass_rate": 0.9814, "provider_group": "provider_z", "quality": 0.9915, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.418, "max_latency_ms": 198123, "minimum_quality": 0.9466}, "family": "migration_planning", "id": "synthetic-migration_planning-000220", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.156, "estimated_latency_ms": 168812, "estimated_pass_rate": 0.6992, "estimated_quality": 0.877, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.418, "max_latency_ms": 198123, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6992, "quality": 0.877, "quality_target_met": false, "utility": 66.947614}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.418,\"max_latency_ms\":198123,\"minimum_quality\":0.9466},\"task\":{\"ambiguity\":0.6865,\"category\":\"coding\",\"complexity\":0.7435,\"consequence\":0.7285,\"evidence_need\":0.7371,\"family\":\"migration_planning\",\"input_tokens\":2920,\"reversibility\":0.6487,\"risk\":0.4271,\"summary\":\"Plan a zero-downtime migration of provider failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.195,\"id\":\"worker_a\",\"p95_latency_ms\":22302,\"pass_rate\":0.481,\"provider_group\":\"provider_x\",\"quality\":0.7346,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.264,\"id\":\"worker_b\",\"p95_latency_ms\":73592,\"pass_rate\":0.7694,\"provider_group\":\"provider_z\",\"quality\":0.8334,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.697,\"id\":\"worker_c\",\"p95_latency_ms\":72918,\"pass_rate\":0.9705,\"provider_group\":\"provider_y\",\"quality\":0.9167,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.156,\"estimated_latency_ms\":168812,\"estimated_pass_rate\":0.6992,\"estimated_quality\":0.877,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.418,\"max_latency_ms\":198123,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6865, "category": "coding", "complexity": 0.7435, "consequence": 0.7285, "evidence_need": 0.7371, "family": "migration_planning", "input_tokens": 2920, "request_id": "synthetic-migration_planning-000220", "reversibility": 0.6487, "risk": 0.4271, "summary": "Plan a zero-downtime migration of provider failover with rollback checkpoints."}, "workers": [{"credits_per_call": 1.195, "id": "worker_a", "p95_latency_ms": 22302, "pass_rate": 0.481, "provider_group": "provider_x", "quality": 0.7346, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.264, "id": "worker_b", "p95_latency_ms": 73592, "pass_rate": 0.7694, "provider_group": "provider_z", "quality": 0.8334, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.697, "id": "worker_c", "p95_latency_ms": 72918, "pass_rate": 0.9705, "provider_group": "provider_y", "quality": 0.9167, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.45, "max_latency_ms": 38211, "minimum_quality": 0.5151}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000773", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.089, "estimated_latency_ms": 19141, "estimated_pass_rate": 0.2828, "estimated_quality": 0.5427, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.45, "max_latency_ms": 38211, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2828, "quality": 0.5427, "quality_target_met": true, "utility": 32.640378}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.45,\"max_latency_ms\":38211,\"minimum_quality\":0.5151},\"task\":{\"ambiguity\":0.6386,\"category\":\"analysis\",\"complexity\":0.7029,\"consequence\":0.3381,\"evidence_need\":0.7239,\"family\":\"performance_tuning\",\"input_tokens\":4880,\"reversibility\":0.5628,\"risk\":0.3282,\"summary\":\"Review a proposed optimization for prompt caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.888,\"id\":\"worker_a\",\"p95_latency_ms\":47447,\"pass_rate\":0.9497,\"provider_group\":\"provider_z\",\"quality\":0.8982,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.845,\"id\":\"worker_b\",\"p95_latency_ms\":75503,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.089,\"id\":\"worker_c\",\"p95_latency_ms\":19141,\"pass_rate\":0.4374,\"provider_group\":\"provider_y\",\"quality\":0.6563,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.089,\"estimated_latency_ms\":19141,\"estimated_pass_rate\":0.2828,\"estimated_quality\":0.5427,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.45,\"max_latency_ms\":38211,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6386, "category": "analysis", "complexity": 0.7029, "consequence": 0.3381, "evidence_need": 0.7239, "family": "performance_tuning", "input_tokens": 4880, "request_id": "synthetic-performance_tuning-000773", "reversibility": 0.5628, "risk": 0.3282, "summary": "Review a proposed optimization for prompt caching for hidden regressions."}, "workers": [{"credits_per_call": 3.888, "id": "worker_a", "p95_latency_ms": 47447, "pass_rate": 0.9497, "provider_group": "provider_z", "quality": 0.8982, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.845, "id": "worker_b", "p95_latency_ms": 75503, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.089, "id": "worker_c", "p95_latency_ms": 19141, "pass_rate": 0.4374, "provider_group": "provider_y", "quality": 0.6563, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.417, "max_latency_ms": 105661, "minimum_quality": 0.8151}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001269", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.117, "estimated_latency_ms": 78647, "estimated_pass_rate": 0.7083, "estimated_quality": 0.9132, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.417, "max_latency_ms": 105661, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7083, "quality": 0.9132, "quality_target_met": true, "utility": 65.219734}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.417,\"max_latency_ms\":105661,\"minimum_quality\":0.8151},\"task\":{\"ambiguity\":0.6294,\"category\":\"review\",\"complexity\":0.6443,\"consequence\":0.6056,\"evidence_need\":0.836,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.5473,\"risk\":0.5978,\"summary\":\"Review whether observability satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":14.148,\"id\":\"worker_a\",\"p95_latency_ms\":56231,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9802,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.638,\"id\":\"worker_b\",\"p95_latency_ms\":73758,\"pass_rate\":0.7463,\"provider_group\":\"provider_y\",\"quality\":0.8726,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.969,\"id\":\"worker_c\",\"p95_latency_ms\":22416,\"pass_rate\":0.5217,\"provider_group\":\"provider_x\",\"quality\":0.7957,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.117,\"estimated_latency_ms\":78647,\"estimated_pass_rate\":0.7083,\"estimated_quality\":0.9132,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.417,\"max_latency_ms\":105661,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6294, "category": "review", "complexity": 0.6443, "consequence": 0.6056, "evidence_need": 0.836, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-001269", "reversibility": 0.5473, "risk": 0.5978, "summary": "Review whether observability satisfies a documented retention policy."}, "workers": [{"credits_per_call": 14.148, "id": "worker_a", "p95_latency_ms": 56231, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9802, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.638, "id": "worker_b", "p95_latency_ms": 73758, "pass_rate": 0.7463, "provider_group": "provider_y", "quality": 0.8726, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.969, "id": "worker_c", "p95_latency_ms": 22416, "pass_rate": 0.5217, "provider_group": "provider_x", "quality": 0.7957, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.684, "max_latency_ms": 221534, "minimum_quality": 0.8111}, "family": "migration_planning", "id": "synthetic-migration_planning-000562", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.383, "estimated_latency_ms": 75149, "estimated_pass_rate": 0.7217, "estimated_quality": 0.858, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.684, "max_latency_ms": 221534, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7217, "quality": 0.858, "quality_target_met": true, "utility": 64.404416}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.684,\"max_latency_ms\":221534,\"minimum_quality\":0.8111},\"task\":{\"ambiguity\":0.3281,\"category\":\"coding\",\"complexity\":0.5636,\"consequence\":0.686,\"evidence_need\":0.8496,\"family\":\"migration_planning\",\"input_tokens\":7160,\"reversibility\":0.551,\"risk\":0.4305,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":3.324,\"id\":\"worker_a\",\"p95_latency_ms\":53251,\"pass_rate\":0.7645,\"provider_group\":\"provider_y\",\"quality\":0.8736,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.059,\"id\":\"worker_b\",\"p95_latency_ms\":21898,\"pass_rate\":0.6112,\"provider_group\":\"provider_z\",\"quality\":0.756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.079,\"id\":\"worker_c\",\"p95_latency_ms\":78251,\"pass_rate\":0.8965,\"provider_group\":\"provider_x\",\"quality\":0.9735,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.383,\"estimated_latency_ms\":75149,\"estimated_pass_rate\":0.7217,\"estimated_quality\":0.858,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.684,\"max_latency_ms\":221534,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3281, "category": "coding", "complexity": 0.5636, "consequence": 0.686, "evidence_need": 0.8496, "family": "migration_planning", "input_tokens": 7160, "request_id": "synthetic-migration_planning-000562", "reversibility": 0.551, "risk": 0.4305, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 3.324, "id": "worker_a", "p95_latency_ms": 53251, "pass_rate": 0.7645, "provider_group": "provider_y", "quality": 0.8736, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.059, "id": "worker_b", "p95_latency_ms": 21898, "pass_rate": 0.6112, "provider_group": "provider_z", "quality": 0.756, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.079, "id": "worker_c", "p95_latency_ms": 78251, "pass_rate": 0.8965, "provider_group": "provider_x", "quality": 0.9735, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.287, "max_latency_ms": 144186, "minimum_quality": 0.7778}, "family": "migration_planning", "id": "synthetic-migration_planning-000583", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.215, "estimated_latency_ms": 91460, "estimated_pass_rate": 0.753, "estimated_quality": 0.822, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.287, "max_latency_ms": 144186, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.753, "quality": 0.822, "quality_target_met": true, "utility": 62.373446}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.287,\"max_latency_ms\":144186,\"minimum_quality\":0.7778},\"task\":{\"ambiguity\":0.7181,\"category\":\"coding\",\"complexity\":0.8018,\"consequence\":0.589,\"evidence_need\":0.7434,\"family\":\"migration_planning\",\"input_tokens\":6640,\"reversibility\":0.4724,\"risk\":0.4751,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.816,\"id\":\"worker_a\",\"p95_latency_ms\":22409,\"pass_rate\":0.5716,\"provider_group\":\"provider_x\",\"quality\":0.7133,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.399,\"id\":\"worker_b\",\"p95_latency_ms\":69051,\"pass_rate\":0.8774,\"provider_group\":\"provider_y\",\"quality\":0.9175,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.427,\"id\":\"worker_c\",\"p95_latency_ms\":91378,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9258,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.215,\"estimated_latency_ms\":91460,\"estimated_pass_rate\":0.753,\"estimated_quality\":0.822,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.287,\"max_latency_ms\":144186,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7181, "category": "coding", "complexity": 0.8018, "consequence": 0.589, "evidence_need": 0.7434, "family": "migration_planning", "input_tokens": 6640, "request_id": "synthetic-migration_planning-000583", "reversibility": 0.4724, "risk": 0.4751, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 0.816, "id": "worker_a", "p95_latency_ms": 22409, "pass_rate": 0.5716, "provider_group": "provider_x", "quality": 0.7133, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.399, "id": "worker_b", "p95_latency_ms": 69051, "pass_rate": 0.8774, "provider_group": "provider_y", "quality": 0.9175, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.427, "id": "worker_c", "p95_latency_ms": 91378, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9258, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.182, "max_latency_ms": 188909, "minimum_quality": 0.7711}, "family": "migration_planning", "id": "synthetic-migration_planning-000931", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.778, "estimated_latency_ms": 81960, "estimated_pass_rate": 0.8353, "estimated_quality": 0.8873, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.182, "max_latency_ms": 188909, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8353, "quality": 0.8873, "quality_target_met": true, "utility": 64.717727}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.182,\"max_latency_ms\":188909,\"minimum_quality\":0.7711},\"task\":{\"ambiguity\":0.4973,\"category\":\"coding\",\"complexity\":0.6999,\"consequence\":0.8347,\"evidence_need\":0.4844,\"family\":\"migration_planning\",\"input_tokens\":520,\"reversibility\":0.6496,\"risk\":0.5668,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":1.02,\"id\":\"worker_a\",\"p95_latency_ms\":18750,\"pass_rate\":0.5302,\"provider_group\":\"provider_x\",\"quality\":0.7473,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.758,\"id\":\"worker_b\",\"p95_latency_ms\":63210,\"pass_rate\":0.9593,\"provider_group\":\"provider_z\",\"quality\":0.9883,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.234,\"id\":\"worker_c\",\"p95_latency_ms\":52954,\"pass_rate\":0.861,\"provider_group\":\"provider_y\",\"quality\":0.8658,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.778,\"estimated_latency_ms\":81960,\"estimated_pass_rate\":0.8353,\"estimated_quality\":0.8873,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.182,\"max_latency_ms\":188909,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4973, "category": "coding", "complexity": 0.6999, "consequence": 0.8347, "evidence_need": 0.4844, "family": "migration_planning", "input_tokens": 520, "request_id": "synthetic-migration_planning-000931", "reversibility": 0.6496, "risk": 0.5668, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 1.02, "id": "worker_a", "p95_latency_ms": 18750, "pass_rate": 0.5302, "provider_group": "provider_x", "quality": 0.7473, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.758, "id": "worker_b", "p95_latency_ms": 63210, "pass_rate": 0.9593, "provider_group": "provider_z", "quality": 0.9883, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.234, "id": "worker_c", "p95_latency_ms": 52954, "pass_rate": 0.861, "provider_group": "provider_y", "quality": 0.8658, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 5.85, "max_latency_ms": 142306, "minimum_quality": 0.8009}, "family": "migration_planning", "id": "synthetic-migration_planning-001243", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.73, "estimated_latency_ms": 86732, "estimated_pass_rate": 0.7063, "estimated_quality": 0.8118, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.85, "max_latency_ms": 142306, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7063, "quality": 0.8118, "quality_target_met": true, "utility": 55.438245}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":5.85,\"max_latency_ms\":142306,\"minimum_quality\":0.8009},\"task\":{\"ambiguity\":0.3383,\"category\":\"coding\",\"complexity\":0.8031,\"consequence\":0.4033,\"evidence_need\":0.4842,\"family\":\"migration_planning\",\"input_tokens\":1200,\"reversibility\":0.6041,\"risk\":0.5223,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.578,\"id\":\"worker_a\",\"p95_latency_ms\":61591,\"pass_rate\":0.789,\"provider_group\":\"provider_x\",\"quality\":0.889,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.152,\"id\":\"worker_b\",\"p95_latency_ms\":25141,\"pass_rate\":0.5236,\"provider_group\":\"provider_y\",\"quality\":0.7679,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.643,\"id\":\"worker_c\",\"p95_latency_ms\":80927,\"pass_rate\":0.9455,\"provider_group\":\"provider_z\",\"quality\":0.9771,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.73,\"estimated_latency_ms\":86732,\"estimated_pass_rate\":0.7063,\"estimated_quality\":0.8118,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.85,\"max_latency_ms\":142306,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3383, "category": "coding", "complexity": 0.8031, "consequence": 0.4033, "evidence_need": 0.4842, "family": "migration_planning", "input_tokens": 1200, "request_id": "synthetic-migration_planning-001243", "reversibility": 0.6041, "risk": 0.5223, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 3.578, "id": "worker_a", "p95_latency_ms": 61591, "pass_rate": 0.789, "provider_group": "provider_x", "quality": 0.889, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.152, "id": "worker_b", "p95_latency_ms": 25141, "pass_rate": 0.5236, "provider_group": "provider_y", "quality": 0.7679, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.643, "id": "worker_c", "p95_latency_ms": 80927, "pass_rate": 0.9455, "provider_group": "provider_z", "quality": 0.9771, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.967, "max_latency_ms": 172954, "minimum_quality": 0.9162}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000912", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.107, "estimated_latency_ms": 145348, "estimated_pass_rate": 0.7478, "estimated_quality": 0.9118, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.967, "max_latency_ms": 172954, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7478, "quality": 0.9118, "quality_target_met": false, "utility": 75.718347}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.967,\"max_latency_ms\":172954,\"minimum_quality\":0.9162},\"task\":{\"ambiguity\":0.7259,\"category\":\"review\",\"complexity\":0.7714,\"consequence\":0.6764,\"evidence_need\":0.7141,\"family\":\"policy_compliance\",\"input_tokens\":240,\"reversibility\":0.3995,\"risk\":0.6276,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.01,\"id\":\"worker_a\",\"p95_latency_ms\":57628,\"pass_rate\":0.8098,\"provider_group\":\"provider_y\",\"quality\":0.8207,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.02,\"id\":\"worker_b\",\"p95_latency_ms\":25572,\"pass_rate\":0.5735,\"provider_group\":\"provider_x\",\"quality\":0.7272,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.077,\"id\":\"worker_c\",\"p95_latency_ms\":62148,\"pass_rate\":0.9636,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.107,\"estimated_latency_ms\":145348,\"estimated_pass_rate\":0.7478,\"estimated_quality\":0.9118,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.967,\"max_latency_ms\":172954,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7259, "category": "review", "complexity": 0.7714, "consequence": 0.6764, "evidence_need": 0.7141, "family": "policy_compliance", "input_tokens": 240, "request_id": "synthetic-policy_compliance-000912", "reversibility": 0.3995, "risk": 0.6276, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 4.01, "id": "worker_a", "p95_latency_ms": 57628, "pass_rate": 0.8098, "provider_group": "provider_y", "quality": 0.8207, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.02, "id": "worker_b", "p95_latency_ms": 25572, "pass_rate": 0.5735, "provider_group": "provider_x", "quality": 0.7272, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.077, "id": "worker_c", "p95_latency_ms": 62148, "pass_rate": 0.9636, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.79, "max_latency_ms": 26581, "minimum_quality": 0.6741}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000744", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.213, "estimated_latency_ms": 23342, "estimated_pass_rate": 0.3507, "estimated_quality": 0.6111, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.79, "max_latency_ms": 26581, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3507, "quality": 0.6111, "quality_target_met": false, "utility": 34.613067}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.79,\"max_latency_ms\":26581,\"minimum_quality\":0.6741},\"task\":{\"ambiguity\":0.7273,\"category\":\"review\",\"complexity\":0.7477,\"consequence\":0.9344,\"evidence_need\":0.5706,\"family\":\"policy_compliance\",\"input_tokens\":320,\"reversibility\":0.3521,\"risk\":0.4525,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.135,\"id\":\"worker_a\",\"p95_latency_ms\":71969,\"pass_rate\":0.7709,\"provider_group\":\"provider_y\",\"quality\":0.8394,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.839,\"id\":\"worker_b\",\"p95_latency_ms\":89614,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9776,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.213,\"id\":\"worker_c\",\"p95_latency_ms\":23342,\"pass_rate\":0.5271,\"provider_group\":\"provider_x\",\"quality\":0.7477,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.213,\"estimated_latency_ms\":23342,\"estimated_pass_rate\":0.3507,\"estimated_quality\":0.6111,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.79,\"max_latency_ms\":26581,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7273, "category": "review", "complexity": 0.7477, "consequence": 0.9344, "evidence_need": 0.5706, "family": "policy_compliance", "input_tokens": 320, "request_id": "synthetic-policy_compliance-000744", "reversibility": 0.3521, "risk": 0.4525, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.135, "id": "worker_a", "p95_latency_ms": 71969, "pass_rate": 0.7709, "provider_group": "provider_y", "quality": 0.8394, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.839, "id": "worker_b", "p95_latency_ms": 89614, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9776, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.213, "id": "worker_c", "p95_latency_ms": 23342, "pass_rate": 0.5271, "provider_group": "provider_x", "quality": 0.7477, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.47, "max_latency_ms": 227326, "minimum_quality": 0.9416}, "family": "migration_planning", "id": "synthetic-migration_planning-000886", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.166, "estimated_latency_ms": 184972, "estimated_pass_rate": 0.7816, "estimated_quality": 0.8973, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.47, "max_latency_ms": 227326, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7816, "quality": 0.8973, "quality_target_met": false, "utility": 70.189793}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.47,\"max_latency_ms\":227326,\"minimum_quality\":0.9416},\"task\":{\"ambiguity\":0.425,\"category\":\"coding\",\"complexity\":0.725,\"consequence\":0.7826,\"evidence_need\":0.4806,\"family\":\"migration_planning\",\"input_tokens\":280,\"reversibility\":0.2858,\"risk\":0.6085,\"summary\":\"Plan a zero-downtime migration of schema evolution with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.845,\"id\":\"worker_a\",\"p95_latency_ms\":21514,\"pass_rate\":0.5857,\"provider_group\":\"provider_y\",\"quality\":0.7847,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.883,\"id\":\"worker_b\",\"p95_latency_ms\":80405,\"pass_rate\":0.8216,\"provider_group\":\"provider_x\",\"quality\":0.8521,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.438,\"id\":\"worker_c\",\"p95_latency_ms\":83053,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.941,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.166,\"estimated_latency_ms\":184972,\"estimated_pass_rate\":0.7816,\"estimated_quality\":0.8973,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.47,\"max_latency_ms\":227326,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.425, "category": "coding", "complexity": 0.725, "consequence": 0.7826, "evidence_need": 0.4806, "family": "migration_planning", "input_tokens": 280, "request_id": "synthetic-migration_planning-000886", "reversibility": 0.2858, "risk": 0.6085, "summary": "Plan a zero-downtime migration of schema evolution with rollback checkpoints."}, "workers": [{"credits_per_call": 0.845, "id": "worker_a", "p95_latency_ms": 21514, "pass_rate": 0.5857, "provider_group": "provider_y", "quality": 0.7847, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.883, "id": "worker_b", "p95_latency_ms": 80405, "pass_rate": 0.8216, "provider_group": "provider_x", "quality": 0.8521, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.438, "id": "worker_c", "p95_latency_ms": 83053, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.941, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.276, "max_latency_ms": 212480, "minimum_quality": 0.8953}, "family": "migration_planning", "id": "synthetic-migration_planning-000778", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.265, "estimated_latency_ms": 152793, "estimated_pass_rate": 0.7102, "estimated_quality": 0.8957, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.276, "max_latency_ms": 212480, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7102, "quality": 0.8957, "quality_target_met": true, "utility": 69.805588}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.276,\"max_latency_ms\":212480,\"minimum_quality\":0.8953},\"task\":{\"ambiguity\":0.4293,\"category\":\"coding\",\"complexity\":0.7567,\"consequence\":0.7937,\"evidence_need\":0.7142,\"family\":\"migration_planning\",\"input_tokens\":6600,\"reversibility\":0.4029,\"risk\":0.4021,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.413,\"id\":\"worker_a\",\"p95_latency_ms\":65324,\"pass_rate\":0.9843,\"provider_group\":\"provider_x\",\"quality\":0.9491,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.87,\"id\":\"worker_b\",\"p95_latency_ms\":19669,\"pass_rate\":0.4719,\"provider_group\":\"provider_y\",\"quality\":0.732,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.982,\"id\":\"worker_c\",\"p95_latency_ms\":67800,\"pass_rate\":0.7528,\"provider_group\":\"provider_z\",\"quality\":0.8735,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.265,\"estimated_latency_ms\":152793,\"estimated_pass_rate\":0.7102,\"estimated_quality\":0.8957,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.276,\"max_latency_ms\":212480,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4293, "category": "coding", "complexity": 0.7567, "consequence": 0.7937, "evidence_need": 0.7142, "family": "migration_planning", "input_tokens": 6600, "request_id": "synthetic-migration_planning-000778", "reversibility": 0.4029, "risk": 0.4021, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 11.413, "id": "worker_a", "p95_latency_ms": 65324, "pass_rate": 0.9843, "provider_group": "provider_x", "quality": 0.9491, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.87, "id": "worker_b", "p95_latency_ms": 19669, "pass_rate": 0.4719, "provider_group": "provider_y", "quality": 0.732, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.982, "id": "worker_c", "p95_latency_ms": 67800, "pass_rate": 0.7528, "provider_group": "provider_z", "quality": 0.8735, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.849, "max_latency_ms": 179501, "minimum_quality": 0.8687}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000420", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.984, "estimated_latency_ms": 128229, "estimated_pass_rate": 0.7132, "estimated_quality": 0.8906, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.849, "max_latency_ms": 179501, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7132, "quality": 0.8906, "quality_target_met": true, "utility": 75.120482}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.849,\"max_latency_ms\":179501,\"minimum_quality\":0.8687},\"task\":{\"ambiguity\":0.6592,\"category\":\"review\",\"complexity\":0.5126,\"consequence\":0.708,\"evidence_need\":0.7835,\"family\":\"policy_compliance\",\"input_tokens\":6840,\"reversibility\":0.4757,\"risk\":0.7899,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.967,\"id\":\"worker_a\",\"p95_latency_ms\":46162,\"pass_rate\":0.7779,\"provider_group\":\"provider_x\",\"quality\":0.8965,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.804,\"id\":\"worker_b\",\"p95_latency_ms\":18897,\"pass_rate\":0.4702,\"provider_group\":\"provider_z\",\"quality\":0.6928,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.213,\"id\":\"worker_c\",\"p95_latency_ms\":63170,\"pass_rate\":0.8891,\"provider_group\":\"provider_y\",\"quality\":0.9447,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.984,\"estimated_latency_ms\":128229,\"estimated_pass_rate\":0.7132,\"estimated_quality\":0.8906,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.849,\"max_latency_ms\":179501,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6592, "category": "review", "complexity": 0.5126, "consequence": 0.708, "evidence_need": 0.7835, "family": "policy_compliance", "input_tokens": 6840, "request_id": "synthetic-policy_compliance-000420", "reversibility": 0.4757, "risk": 0.7899, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 2.967, "id": "worker_a", "p95_latency_ms": 46162, "pass_rate": 0.7779, "provider_group": "provider_x", "quality": 0.8965, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.804, "id": "worker_b", "p95_latency_ms": 18897, "pass_rate": 0.4702, "provider_group": "provider_z", "quality": 0.6928, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.213, "id": "worker_c", "p95_latency_ms": 63170, "pass_rate": 0.8891, "provider_group": "provider_y", "quality": 0.9447, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.303, "max_latency_ms": 25929, "minimum_quality": 0.5555}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000372", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.815, "estimated_latency_ms": 20726, "estimated_pass_rate": 0.4354, "estimated_quality": 0.7021, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.303, "max_latency_ms": 25929, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4354, "quality": 0.7021, "quality_target_met": true, "utility": 32.616084}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.303,\"max_latency_ms\":25929,\"minimum_quality\":0.5555},\"task\":{\"ambiguity\":0.5901,\"category\":\"review\",\"complexity\":0.4203,\"consequence\":0.8,\"evidence_need\":0.8805,\"family\":\"policy_compliance\",\"input_tokens\":5240,\"reversibility\":0.3987,\"risk\":0.7362,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.61,\"id\":\"worker_a\",\"p95_latency_ms\":51367,\"pass_rate\":0.9882,\"provider_group\":\"provider_z\",\"quality\":0.9649,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.815,\"id\":\"worker_b\",\"p95_latency_ms\":20726,\"pass_rate\":0.5873,\"provider_group\":\"provider_y\",\"quality\":0.7908,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.309,\"id\":\"worker_c\",\"p95_latency_ms\":51704,\"pass_rate\":0.7924,\"provider_group\":\"provider_x\",\"quality\":0.8997,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.815,\"estimated_latency_ms\":20726,\"estimated_pass_rate\":0.4354,\"estimated_quality\":0.7021,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.303,\"max_latency_ms\":25929,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5901, "category": "review", "complexity": 0.4203, "consequence": 0.8, "evidence_need": 0.8805, "family": "policy_compliance", "input_tokens": 5240, "request_id": "synthetic-policy_compliance-000372", "reversibility": 0.3987, "risk": 0.7362, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 11.61, "id": "worker_a", "p95_latency_ms": 51367, "pass_rate": 0.9882, "provider_group": "provider_z", "quality": 0.9649, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.815, "id": "worker_b", "p95_latency_ms": 20726, "pass_rate": 0.5873, "provider_group": "provider_y", "quality": 0.7908, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.309, "id": "worker_c", "p95_latency_ms": 51704, "pass_rate": 0.7924, "provider_group": "provider_x", "quality": 0.8997, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.731, "max_latency_ms": 211250, "minimum_quality": 0.9397}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000656", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.453, "estimated_latency_ms": 168305, "estimated_pass_rate": 0.7075, "estimated_quality": 0.9278, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.731, "max_latency_ms": 211250, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7075, "quality": 0.9278, "quality_target_met": false, "utility": 70.389674}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.731,\"max_latency_ms\":211250,\"minimum_quality\":0.9397},\"task\":{\"ambiguity\":0.7164,\"category\":\"analysis\",\"complexity\":0.7547,\"consequence\":0.5563,\"evidence_need\":0.615,\"family\":\"performance_tuning\",\"input_tokens\":5440,\"reversibility\":0.7004,\"risk\":0.2755,\"summary\":\"Choose between throughput and tail-latency changes for access control.\"},\"workers\":[{\"credits_per_call\":3.801,\"id\":\"worker_a\",\"p95_latency_ms\":75944,\"pass_rate\":0.8743,\"provider_group\":\"provider_x\",\"quality\":0.9288,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.719,\"id\":\"worker_b\",\"p95_latency_ms\":65557,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.933,\"id\":\"worker_c\",\"p95_latency_ms\":26804,\"pass_rate\":0.4202,\"provider_group\":\"provider_y\",\"quality\":0.7399,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.453,\"estimated_latency_ms\":168305,\"estimated_pass_rate\":0.7075,\"estimated_quality\":0.9278,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.731,\"max_latency_ms\":211250,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7164, "category": "analysis", "complexity": 0.7547, "consequence": 0.5563, "evidence_need": 0.615, "family": "performance_tuning", "input_tokens": 5440, "request_id": "synthetic-performance_tuning-000656", "reversibility": 0.7004, "risk": 0.2755, "summary": "Choose between throughput and tail-latency changes for access control."}, "workers": [{"credits_per_call": 3.801, "id": "worker_a", "p95_latency_ms": 75944, "pass_rate": 0.8743, "provider_group": "provider_x", "quality": 0.9288, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.719, "id": "worker_b", "p95_latency_ms": 65557, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.933, "id": "worker_c", "p95_latency_ms": 26804, "pass_rate": 0.4202, "provider_group": "provider_y", "quality": 0.7399, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.785, "max_latency_ms": 97874, "minimum_quality": 0.8678}, "family": "migration_planning", "id": "synthetic-migration_planning-000421", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.553, "estimated_latency_ms": 77339, "estimated_pass_rate": 0.7189, "estimated_quality": 0.8972, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.785, "max_latency_ms": 97874, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7189, "quality": 0.8972, "quality_target_met": true, "utility": 65.785557}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.785,\"max_latency_ms\":97874,\"minimum_quality\":0.8678},\"task\":{\"ambiguity\":0.372,\"category\":\"coding\",\"complexity\":0.5568,\"consequence\":0.5833,\"evidence_need\":0.7706,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.4715,\"risk\":0.6664,\"summary\":\"Design compatibility testing for a artifact storage schema migration.\"},\"workers\":[{\"credits_per_call\":0.822,\"id\":\"worker_a\",\"p95_latency_ms\":21652,\"pass_rate\":0.5439,\"provider_group\":\"provider_y\",\"quality\":0.7792,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.731,\"id\":\"worker_b\",\"p95_latency_ms\":55687,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9125,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.969,\"id\":\"worker_c\",\"p95_latency_ms\":48961,\"pass_rate\":0.7655,\"provider_group\":\"provider_x\",\"quality\":0.8448,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.553,\"estimated_latency_ms\":77339,\"estimated_pass_rate\":0.7189,\"estimated_quality\":0.8972,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.785,\"max_latency_ms\":97874,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.372, "category": "coding", "complexity": 0.5568, "consequence": 0.5833, "evidence_need": 0.7706, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-000421", "reversibility": 0.4715, "risk": 0.6664, "summary": "Design compatibility testing for a artifact storage schema migration."}, "workers": [{"credits_per_call": 0.822, "id": "worker_a", "p95_latency_ms": 21652, "pass_rate": 0.5439, "provider_group": "provider_y", "quality": 0.7792, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.731, "id": "worker_b", "p95_latency_ms": 55687, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9125, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.969, "id": "worker_c", "p95_latency_ms": 48961, "pass_rate": 0.7655, "provider_group": "provider_x", "quality": 0.8448, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.623, "max_latency_ms": 184245, "minimum_quality": 0.7384}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000879", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.745, "estimated_latency_ms": 100546, "estimated_pass_rate": 0.6144, "estimated_quality": 0.7615, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.623, "max_latency_ms": 184245, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6144, "quality": 0.7615, "quality_target_met": true, "utility": 57.386107}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.623,\"max_latency_ms\":184245,\"minimum_quality\":0.7384},\"task\":{\"ambiguity\":0.7551,\"category\":\"review\",\"complexity\":0.7664,\"consequence\":0.6182,\"evidence_need\":0.5686,\"family\":\"policy_compliance\",\"input_tokens\":1600,\"reversibility\":0.6729,\"risk\":0.6617,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.766,\"id\":\"worker_a\",\"p95_latency_ms\":81987,\"pass_rate\":0.7839,\"provider_group\":\"provider_x\",\"quality\":0.8322,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.776,\"id\":\"worker_b\",\"p95_latency_ms\":75365,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9822,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.979,\"id\":\"worker_c\",\"p95_latency_ms\":18559,\"pass_rate\":0.5756,\"provider_group\":\"provider_z\",\"quality\":0.7416,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.745,\"estimated_latency_ms\":100546,\"estimated_pass_rate\":0.6144,\"estimated_quality\":0.7615,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.623,\"max_latency_ms\":184245,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7551, "category": "review", "complexity": 0.7664, "consequence": 0.6182, "evidence_need": 0.5686, "family": "policy_compliance", "input_tokens": 1600, "request_id": "synthetic-policy_compliance-000879", "reversibility": 0.6729, "risk": 0.6617, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 2.766, "id": "worker_a", "p95_latency_ms": 81987, "pass_rate": 0.7839, "provider_group": "provider_x", "quality": 0.8322, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.776, "id": "worker_b", "p95_latency_ms": 75365, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9822, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.979, "id": "worker_c", "p95_latency_ms": 18559, "pass_rate": 0.5756, "provider_group": "provider_z", "quality": 0.7416, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.743, "max_latency_ms": 195970, "minimum_quality": 0.8539}, "family": "migration_planning", "id": "synthetic-migration_planning-001048", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.883, "estimated_latency_ms": 152192, "estimated_pass_rate": 0.7582, "estimated_quality": 0.9038, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.743, "max_latency_ms": 195970, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7582, "quality": 0.9038, "quality_target_met": true, "utility": 73.580161}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.743,\"max_latency_ms\":195970,\"minimum_quality\":0.8539},\"task\":{\"ambiguity\":0.6593,\"category\":\"coding\",\"complexity\":0.694,\"consequence\":0.8402,\"evidence_need\":0.4604,\"family\":\"migration_planning\",\"input_tokens\":3640,\"reversibility\":0.6023,\"risk\":0.5983,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.867,\"id\":\"worker_a\",\"p95_latency_ms\":72087,\"pass_rate\":0.8084,\"provider_group\":\"provider_z\",\"quality\":0.861,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.009,\"id\":\"worker_b\",\"p95_latency_ms\":30313,\"pass_rate\":0.5767,\"provider_group\":\"provider_y\",\"quality\":0.7092,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.007,\"id\":\"worker_c\",\"p95_latency_ms\":49792,\"pass_rate\":0.9601,\"provider_group\":\"provider_x\",\"quality\":0.9931,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.883,\"estimated_latency_ms\":152192,\"estimated_pass_rate\":0.7582,\"estimated_quality\":0.9038,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.743,\"max_latency_ms\":195970,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6593, "category": "coding", "complexity": 0.694, "consequence": 0.8402, "evidence_need": 0.4604, "family": "migration_planning", "input_tokens": 3640, "request_id": "synthetic-migration_planning-001048", "reversibility": 0.6023, "risk": 0.5983, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 3.867, "id": "worker_a", "p95_latency_ms": 72087, "pass_rate": 0.8084, "provider_group": "provider_z", "quality": 0.861, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.009, "id": "worker_b", "p95_latency_ms": 30313, "pass_rate": 0.5767, "provider_group": "provider_y", "quality": 0.7092, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.007, "id": "worker_c", "p95_latency_ms": 49792, "pass_rate": 0.9601, "provider_group": "provider_x", "quality": 0.9931, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.979, "max_latency_ms": 184185, "minimum_quality": 0.8764}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001005", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.314, "estimated_latency_ms": 146247, "estimated_pass_rate": 0.6901, "estimated_quality": 0.8951, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.979, "max_latency_ms": 184185, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6901, "quality": 0.8951, "quality_target_met": true, "utility": 72.125451}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.979,\"max_latency_ms\":184185,\"minimum_quality\":0.8764},\"task\":{\"ambiguity\":0.4255,\"category\":\"review\",\"complexity\":0.6355,\"consequence\":0.9036,\"evidence_need\":0.8521,\"family\":\"policy_compliance\",\"input_tokens\":2920,\"reversibility\":0.5952,\"risk\":0.5348,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.166,\"id\":\"worker_a\",\"p95_latency_ms\":22085,\"pass_rate\":0.497,\"provider_group\":\"provider_x\",\"quality\":0.7198,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.469,\"id\":\"worker_b\",\"p95_latency_ms\":61642,\"pass_rate\":0.915,\"provider_group\":\"provider_y\",\"quality\":0.9643,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.679,\"id\":\"worker_c\",\"p95_latency_ms\":62520,\"pass_rate\":0.7273,\"provider_group\":\"provider_z\",\"quality\":0.7955,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.314,\"estimated_latency_ms\":146247,\"estimated_pass_rate\":0.6901,\"estimated_quality\":0.8951,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.979,\"max_latency_ms\":184185,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4255, "category": "review", "complexity": 0.6355, "consequence": 0.9036, "evidence_need": 0.8521, "family": "policy_compliance", "input_tokens": 2920, "request_id": "synthetic-policy_compliance-001005", "reversibility": 0.5952, "risk": 0.5348, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.166, "id": "worker_a", "p95_latency_ms": 22085, "pass_rate": 0.497, "provider_group": "provider_x", "quality": 0.7198, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.469, "id": "worker_b", "p95_latency_ms": 61642, "pass_rate": 0.915, "provider_group": "provider_y", "quality": 0.9643, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.679, "id": "worker_c", "p95_latency_ms": 62520, "pass_rate": 0.7273, "provider_group": "provider_z", "quality": 0.7955, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.644, "max_latency_ms": 26659, "minimum_quality": 0.5282}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000494", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.97, "estimated_latency_ms": 19618, "estimated_pass_rate": 0.2937, "estimated_quality": 0.5125, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.644, "max_latency_ms": 26659, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2937, "quality": 0.5125, "quality_target_met": false, "utility": 30.42904}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.644,\"max_latency_ms\":26659,\"minimum_quality\":0.5282},\"task\":{\"ambiguity\":0.6762,\"category\":\"analysis\",\"complexity\":0.7445,\"consequence\":0.2988,\"evidence_need\":0.5643,\"family\":\"performance_tuning\",\"input_tokens\":440,\"reversibility\":0.7483,\"risk\":0.5378,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":10.312,\"id\":\"worker_a\",\"p95_latency_ms\":86904,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.97,\"id\":\"worker_b\",\"p95_latency_ms\":19618,\"pass_rate\":0.4728,\"provider_group\":\"provider_y\",\"quality\":0.6513,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.113,\"id\":\"worker_c\",\"p95_latency_ms\":66607,\"pass_rate\":0.9508,\"provider_group\":\"provider_x\",\"quality\":0.8712,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.97,\"estimated_latency_ms\":19618,\"estimated_pass_rate\":0.2937,\"estimated_quality\":0.5125,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.644,\"max_latency_ms\":26659,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6762, "category": "analysis", "complexity": 0.7445, "consequence": 0.2988, "evidence_need": 0.5643, "family": "performance_tuning", "input_tokens": 440, "request_id": "synthetic-performance_tuning-000494", "reversibility": 0.7483, "risk": 0.5378, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 10.312, "id": "worker_a", "p95_latency_ms": 86904, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.97, "id": "worker_b", "p95_latency_ms": 19618, "pass_rate": 0.4728, "provider_group": "provider_y", "quality": 0.6513, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.113, "id": "worker_c", "p95_latency_ms": 66607, "pass_rate": 0.9508, "provider_group": "provider_x", "quality": 0.8712, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.312, "max_latency_ms": 181878, "minimum_quality": 0.8906}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000354", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.4, "estimated_latency_ms": 95783, "estimated_pass_rate": 0.6655, "estimated_quality": 0.8124, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.312, "max_latency_ms": 181878, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6655, "quality": 0.8124, "quality_target_met": false, "utility": 63.269118}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.312,\"max_latency_ms\":181878,\"minimum_quality\":0.8906},\"task\":{\"ambiguity\":0.7386,\"category\":\"review\",\"complexity\":0.7468,\"consequence\":0.7952,\"evidence_need\":0.8386,\"family\":\"policy_compliance\",\"input_tokens\":4760,\"reversibility\":0.5454,\"risk\":0.798,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.611,\"id\":\"worker_a\",\"p95_latency_ms\":61303,\"pass_rate\":0.8978,\"provider_group\":\"provider_y\",\"quality\":0.9616,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.968,\"id\":\"worker_b\",\"p95_latency_ms\":25809,\"pass_rate\":0.5596,\"provider_group\":\"provider_z\",\"quality\":0.6881,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.432,\"id\":\"worker_c\",\"p95_latency_ms\":69974,\"pass_rate\":0.8615,\"provider_group\":\"provider_x\",\"quality\":0.887,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.4,\"estimated_latency_ms\":95783,\"estimated_pass_rate\":0.6655,\"estimated_quality\":0.8124,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.312,\"max_latency_ms\":181878,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7386, "category": "review", "complexity": 0.7468, "consequence": 0.7952, "evidence_need": 0.8386, "family": "policy_compliance", "input_tokens": 4760, "request_id": "synthetic-policy_compliance-000354", "reversibility": 0.5454, "risk": 0.798, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.611, "id": "worker_a", "p95_latency_ms": 61303, "pass_rate": 0.8978, "provider_group": "provider_y", "quality": 0.9616, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.968, "id": "worker_b", "p95_latency_ms": 25809, "pass_rate": 0.5596, "provider_group": "provider_z", "quality": 0.6881, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.432, "id": "worker_c", "p95_latency_ms": 69974, "pass_rate": 0.8615, "provider_group": "provider_x", "quality": 0.887, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.062, "max_latency_ms": 207213, "minimum_quality": 0.9143}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000198", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.842, "estimated_latency_ms": 160100, "estimated_pass_rate": 0.7797, "estimated_quality": 0.9177, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.062, "max_latency_ms": 207213, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7797, "quality": 0.9177, "quality_target_met": true, "utility": 78.551097}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.062,\"max_latency_ms\":207213,\"minimum_quality\":0.9143},\"task\":{\"ambiguity\":0.3651,\"category\":\"review\",\"complexity\":0.6595,\"consequence\":0.7159,\"evidence_need\":0.9264,\"family\":\"policy_compliance\",\"input_tokens\":5080,\"reversibility\":0.5793,\"risk\":0.8174,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.356,\"id\":\"worker_a\",\"p95_latency_ms\":59107,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9831,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.608,\"id\":\"worker_b\",\"p95_latency_ms\":75142,\"pass_rate\":0.7901,\"provider_group\":\"provider_y\",\"quality\":0.8167,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.878,\"id\":\"worker_c\",\"p95_latency_ms\":25851,\"pass_rate\":0.5712,\"provider_group\":\"provider_z\",\"quality\":0.7362,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.842,\"estimated_latency_ms\":160100,\"estimated_pass_rate\":0.7797,\"estimated_quality\":0.9177,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.062,\"max_latency_ms\":207213,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3651, "category": "review", "complexity": 0.6595, "consequence": 0.7159, "evidence_need": 0.9264, "family": "policy_compliance", "input_tokens": 5080, "request_id": "synthetic-policy_compliance-000198", "reversibility": 0.5793, "risk": 0.8174, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 11.356, "id": "worker_a", "p95_latency_ms": 59107, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9831, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.608, "id": "worker_b", "p95_latency_ms": 75142, "pass_rate": 0.7901, "provider_group": "provider_y", "quality": 0.8167, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.878, "id": "worker_c", "p95_latency_ms": 25851, "pass_rate": 0.5712, "provider_group": "provider_z", "quality": 0.7362, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.703, "max_latency_ms": 43517, "minimum_quality": 0.765}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000584", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.823, "estimated_latency_ms": 23938, "estimated_pass_rate": 0.3752, "estimated_quality": 0.5517, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.703, "max_latency_ms": 43517, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3752, "quality": 0.5517, "quality_target_met": false, "utility": 29.269687}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.703,\"max_latency_ms\":43517,\"minimum_quality\":0.765},\"task\":{\"ambiguity\":0.539,\"category\":\"analysis\",\"complexity\":0.5764,\"consequence\":0.4315,\"evidence_need\":0.3553,\"family\":\"performance_tuning\",\"input_tokens\":1240,\"reversibility\":0.6163,\"risk\":0.399,\"summary\":\"Choose between throughput and tail-latency changes for API-key rotation.\"},\"workers\":[{\"credits_per_call\":13.599,\"id\":\"worker_a\",\"p95_latency_ms\":89817,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.823,\"id\":\"worker_b\",\"p95_latency_ms\":23938,\"pass_rate\":0.5136,\"provider_group\":\"provider_x\",\"quality\":0.6409,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.555,\"id\":\"worker_c\",\"p95_latency_ms\":44019,\"pass_rate\":0.9196,\"provider_group\":\"provider_z\",\"quality\":0.8654,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.823,\"estimated_latency_ms\":23938,\"estimated_pass_rate\":0.3752,\"estimated_quality\":0.5517,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.703,\"max_latency_ms\":43517,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.539, "category": "analysis", "complexity": 0.5764, "consequence": 0.4315, "evidence_need": 0.3553, "family": "performance_tuning", "input_tokens": 1240, "request_id": "synthetic-performance_tuning-000584", "reversibility": 0.6163, "risk": 0.399, "summary": "Choose between throughput and tail-latency changes for API-key rotation."}, "workers": [{"credits_per_call": 13.599, "id": "worker_a", "p95_latency_ms": 89817, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.823, "id": "worker_b", "p95_latency_ms": 23938, "pass_rate": 0.5136, "provider_group": "provider_x", "quality": 0.6409, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.555, "id": "worker_c", "p95_latency_ms": 44019, "pass_rate": 0.9196, "provider_group": "provider_z", "quality": 0.8654, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.795, "max_latency_ms": 25098, "minimum_quality": 0.5692}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000509", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.05, "estimated_latency_ms": 21494, "estimated_pass_rate": 0.2845, "estimated_quality": 0.6141, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.795, "max_latency_ms": 25098, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2845, "quality": 0.6141, "quality_target_met": true, "utility": 38.827128}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.795,\"max_latency_ms\":25098,\"minimum_quality\":0.5692},\"task\":{\"ambiguity\":0.5978,\"category\":\"analysis\",\"complexity\":0.7976,\"consequence\":0.5777,\"evidence_need\":0.7088,\"family\":\"performance_tuning\",\"input_tokens\":7000,\"reversibility\":0.7445,\"risk\":0.2433,\"summary\":\"Choose between throughput and tail-latency changes for usage metering.\"},\"workers\":[{\"credits_per_call\":3.402,\"id\":\"worker_a\",\"p95_latency_ms\":44383,\"pass_rate\":0.9434,\"provider_group\":\"provider_x\",\"quality\":0.9148,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.806,\"id\":\"worker_b\",\"p95_latency_ms\":77538,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.05,\"id\":\"worker_c\",\"p95_latency_ms\":21494,\"pass_rate\":0.4398,\"provider_group\":\"provider_z\",\"quality\":0.736,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.05,\"estimated_latency_ms\":21494,\"estimated_pass_rate\":0.2845,\"estimated_quality\":0.6141,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.795,\"max_latency_ms\":25098,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5978, "category": "analysis", "complexity": 0.7976, "consequence": 0.5777, "evidence_need": 0.7088, "family": "performance_tuning", "input_tokens": 7000, "request_id": "synthetic-performance_tuning-000509", "reversibility": 0.7445, "risk": 0.2433, "summary": "Choose between throughput and tail-latency changes for usage metering."}, "workers": [{"credits_per_call": 3.402, "id": "worker_a", "p95_latency_ms": 44383, "pass_rate": 0.9434, "provider_group": "provider_x", "quality": 0.9148, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.806, "id": "worker_b", "p95_latency_ms": 77538, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.05, "id": "worker_c", "p95_latency_ms": 21494, "pass_rate": 0.4398, "provider_group": "provider_z", "quality": 0.736, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.584, "max_latency_ms": 37948, "minimum_quality": 0.7382}, "family": "migration_planning", "id": "synthetic-migration_planning-000586", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.16, "estimated_latency_ms": 29681, "estimated_pass_rate": 0.2875, "estimated_quality": 0.585, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.584, "max_latency_ms": 37948, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2875, "quality": 0.585, "quality_target_met": false, "utility": 30.662803}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.584,\"max_latency_ms\":37948,\"minimum_quality\":0.7382},\"task\":{\"ambiguity\":0.7264,\"category\":\"coding\",\"complexity\":0.8218,\"consequence\":0.5778,\"evidence_need\":0.4923,\"family\":\"migration_planning\",\"input_tokens\":1360,\"reversibility\":0.4341,\"risk\":0.6941,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":2.743,\"id\":\"worker_a\",\"p95_latency_ms\":71108,\"pass_rate\":0.8066,\"provider_group\":\"provider_z\",\"quality\":0.8895,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.16,\"id\":\"worker_b\",\"p95_latency_ms\":29681,\"pass_rate\":0.4918,\"provider_group\":\"provider_x\",\"quality\":0.7524,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.759,\"id\":\"worker_c\",\"p95_latency_ms\":68169,\"pass_rate\":0.9361,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.16,\"estimated_latency_ms\":29681,\"estimated_pass_rate\":0.2875,\"estimated_quality\":0.585,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.584,\"max_latency_ms\":37948,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7264, "category": "coding", "complexity": 0.8218, "consequence": 0.5778, "evidence_need": 0.4923, "family": "migration_planning", "input_tokens": 1360, "request_id": "synthetic-migration_planning-000586", "reversibility": 0.4341, "risk": 0.6941, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 2.743, "id": "worker_a", "p95_latency_ms": 71108, "pass_rate": 0.8066, "provider_group": "provider_z", "quality": 0.8895, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.16, "id": "worker_b", "p95_latency_ms": 29681, "pass_rate": 0.4918, "provider_group": "provider_x", "quality": 0.7524, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.759, "id": "worker_c", "p95_latency_ms": 68169, "pass_rate": 0.9361, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.793, "max_latency_ms": 34152, "minimum_quality": 0.5204}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000783", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.0, "estimated_latency_ms": 20755, "estimated_pass_rate": 0.4313, "estimated_quality": 0.565, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.793, "max_latency_ms": 34152, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4313, "quality": 0.565, "quality_target_met": true, "utility": 25.310515}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.793,\"max_latency_ms\":34152,\"minimum_quality\":0.5204},\"task\":{\"ambiguity\":0.3825,\"category\":\"review\",\"complexity\":0.6507,\"consequence\":0.6423,\"evidence_need\":0.876,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.3332,\"risk\":0.7656,\"summary\":\"Identify evidence required before approving provider failover for production.\"},\"workers\":[{\"credits_per_call\":15.315,\"id\":\"worker_a\",\"p95_latency_ms\":52653,\"pass_rate\":0.9399,\"provider_group\":\"provider_y\",\"quality\":0.9901,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.0,\"id\":\"worker_b\",\"p95_latency_ms\":20755,\"pass_rate\":0.596,\"provider_group\":\"provider_x\",\"quality\":0.6835,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.17,\"id\":\"worker_c\",\"p95_latency_ms\":78562,\"pass_rate\":0.7498,\"provider_group\":\"provider_z\",\"quality\":0.828,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.0,\"estimated_latency_ms\":20755,\"estimated_pass_rate\":0.4313,\"estimated_quality\":0.565,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.793,\"max_latency_ms\":34152,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3825, "category": "review", "complexity": 0.6507, "consequence": 0.6423, "evidence_need": 0.876, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-000783", "reversibility": 0.3332, "risk": 0.7656, "summary": "Identify evidence required before approving provider failover for production."}, "workers": [{"credits_per_call": 15.315, "id": "worker_a", "p95_latency_ms": 52653, "pass_rate": 0.9399, "provider_group": "provider_y", "quality": 0.9901, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.0, "id": "worker_b", "p95_latency_ms": 20755, "pass_rate": 0.596, "provider_group": "provider_x", "quality": 0.6835, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.17, "id": "worker_c", "p95_latency_ms": 78562, "pass_rate": 0.7498, "provider_group": "provider_z", "quality": 0.828, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.332, "max_latency_ms": 178012, "minimum_quality": 0.8131}, "family": "migration_planning", "id": "synthetic-migration_planning-000985", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.905, "estimated_latency_ms": 110512, "estimated_pass_rate": 0.7557, "estimated_quality": 0.8262, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.332, "max_latency_ms": 178012, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7557, "quality": 0.8262, "quality_target_met": true, "utility": 59.325715}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.332,\"max_latency_ms\":178012,\"minimum_quality\":0.8131},\"task\":{\"ambiguity\":0.3928,\"category\":\"coding\",\"complexity\":0.9334,\"consequence\":0.8455,\"evidence_need\":0.6082,\"family\":\"migration_planning\",\"input_tokens\":360,\"reversibility\":0.6492,\"risk\":0.7382,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":15.045,\"id\":\"worker_a\",\"p95_latency_ms\":84498,\"pass_rate\":0.8883,\"provider_group\":\"provider_x\",\"quality\":0.9528,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.314,\"id\":\"worker_b\",\"p95_latency_ms\":66468,\"pass_rate\":0.8085,\"provider_group\":\"provider_z\",\"quality\":0.8879,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.86,\"id\":\"worker_c\",\"p95_latency_ms\":26014,\"pass_rate\":0.5266,\"provider_group\":\"provider_y\",\"quality\":0.6918,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.905,\"estimated_latency_ms\":110512,\"estimated_pass_rate\":0.7557,\"estimated_quality\":0.8262,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.332,\"max_latency_ms\":178012,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3928, "category": "coding", "complexity": 0.9334, "consequence": 0.8455, "evidence_need": 0.6082, "family": "migration_planning", "input_tokens": 360, "request_id": "synthetic-migration_planning-000985", "reversibility": 0.6492, "risk": 0.7382, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 15.045, "id": "worker_a", "p95_latency_ms": 84498, "pass_rate": 0.8883, "provider_group": "provider_x", "quality": 0.9528, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.314, "id": "worker_b", "p95_latency_ms": 66468, "pass_rate": 0.8085, "provider_group": "provider_z", "quality": 0.8879, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.86, "id": "worker_c", "p95_latency_ms": 26014, "pass_rate": 0.5266, "provider_group": "provider_y", "quality": 0.6918, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.136, "max_latency_ms": 25437, "minimum_quality": 0.4999}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000525", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.897, "estimated_latency_ms": 18376, "estimated_pass_rate": 0.3409, "estimated_quality": 0.6302, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.136, "max_latency_ms": 25437, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3409, "quality": 0.6302, "quality_target_met": true, "utility": 41.660013}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.136,\"max_latency_ms\":25437,\"minimum_quality\":0.4999},\"task\":{\"ambiguity\":0.6805,\"category\":\"review\",\"complexity\":0.5815,\"consequence\":0.5678,\"evidence_need\":0.8309,\"family\":\"policy_compliance\",\"input_tokens\":4240,\"reversibility\":0.4541,\"risk\":0.5132,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.897,\"id\":\"worker_a\",\"p95_latency_ms\":18376,\"pass_rate\":0.4999,\"provider_group\":\"provider_z\",\"quality\":0.738,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.839,\"id\":\"worker_b\",\"p95_latency_ms\":84055,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.773,\"id\":\"worker_c\",\"p95_latency_ms\":74674,\"pass_rate\":0.7781,\"provider_group\":\"provider_y\",\"quality\":0.8726,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.897,\"estimated_latency_ms\":18376,\"estimated_pass_rate\":0.3409,\"estimated_quality\":0.6302,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.136,\"max_latency_ms\":25437,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6805, "category": "review", "complexity": 0.5815, "consequence": 0.5678, "evidence_need": 0.8309, "family": "policy_compliance", "input_tokens": 4240, "request_id": "synthetic-policy_compliance-000525", "reversibility": 0.4541, "risk": 0.5132, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 0.897, "id": "worker_a", "p95_latency_ms": 18376, "pass_rate": 0.4999, "provider_group": "provider_z", "quality": 0.738, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.839, "id": "worker_b", "p95_latency_ms": 84055, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.773, "id": "worker_c", "p95_latency_ms": 74674, "pass_rate": 0.7781, "provider_group": "provider_y", "quality": 0.8726, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.626, "max_latency_ms": 36516, "minimum_quality": 0.5556}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000053", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.148, "estimated_latency_ms": 23737, "estimated_pass_rate": 0.3904, "estimated_quality": 0.6137, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.626, "max_latency_ms": 36516, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3904, "quality": 0.6137, "quality_target_met": true, "utility": 41.887187}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.626,\"max_latency_ms\":36516,\"minimum_quality\":0.5556},\"task\":{\"ambiguity\":0.7028,\"category\":\"analysis\",\"complexity\":0.4549,\"consequence\":0.2834,\"evidence_need\":0.4303,\"family\":\"performance_tuning\",\"input_tokens\":5560,\"reversibility\":0.7906,\"risk\":0.5524,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":1.148,\"id\":\"worker_a\",\"p95_latency_ms\":23737,\"pass_rate\":0.5399,\"provider_group\":\"provider_z\",\"quality\":0.7029,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.796,\"id\":\"worker_b\",\"p95_latency_ms\":61294,\"pass_rate\":0.8329,\"provider_group\":\"provider_y\",\"quality\":0.8943,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.302,\"id\":\"worker_c\",\"p95_latency_ms\":66775,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.148,\"estimated_latency_ms\":23737,\"estimated_pass_rate\":0.3904,\"estimated_quality\":0.6137,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.626,\"max_latency_ms\":36516,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7028, "category": "analysis", "complexity": 0.4549, "consequence": 0.2834, "evidence_need": 0.4303, "family": "performance_tuning", "input_tokens": 5560, "request_id": "synthetic-performance_tuning-000053", "reversibility": 0.7906, "risk": 0.5524, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 1.148, "id": "worker_a", "p95_latency_ms": 23737, "pass_rate": 0.5399, "provider_group": "provider_z", "quality": 0.7029, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.796, "id": "worker_b", "p95_latency_ms": 61294, "pass_rate": 0.8329, "provider_group": "provider_y", "quality": 0.8943, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.302, "id": "worker_c", "p95_latency_ms": 66775, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.963, "max_latency_ms": 203366, "minimum_quality": 0.9143}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000303", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.649, "estimated_latency_ms": 164640, "estimated_pass_rate": 0.7804, "estimated_quality": 0.9205, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.963, "max_latency_ms": 203366, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7804, "quality": 0.9205, "quality_target_met": true, "utility": 79.077323}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.963,\"max_latency_ms\":203366,\"minimum_quality\":0.9143},\"task\":{\"ambiguity\":0.735,\"category\":\"review\",\"complexity\":0.6458,\"consequence\":0.6746,\"evidence_need\":0.914,\"family\":\"policy_compliance\",\"input_tokens\":880,\"reversibility\":0.6673,\"risk\":0.6458,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.467,\"id\":\"worker_a\",\"p95_latency_ms\":80001,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9729,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.991,\"id\":\"worker_b\",\"p95_latency_ms\":51904,\"pass_rate\":0.8669,\"provider_group\":\"provider_z\",\"quality\":0.8506,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.191,\"id\":\"worker_c\",\"p95_latency_ms\":32735,\"pass_rate\":0.569,\"provider_group\":\"provider_y\",\"quality\":0.7305,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.649,\"estimated_latency_ms\":164640,\"estimated_pass_rate\":0.7804,\"estimated_quality\":0.9205,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.963,\"max_latency_ms\":203366,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.735, "category": "review", "complexity": 0.6458, "consequence": 0.6746, "evidence_need": 0.914, "family": "policy_compliance", "input_tokens": 880, "request_id": "synthetic-policy_compliance-000303", "reversibility": 0.6673, "risk": 0.6458, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.467, "id": "worker_a", "p95_latency_ms": 80001, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9729, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.991, "id": "worker_b", "p95_latency_ms": 51904, "pass_rate": 0.8669, "provider_group": "provider_z", "quality": 0.8506, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.191, "id": "worker_c", "p95_latency_ms": 32735, "pass_rate": 0.569, "provider_group": "provider_y", "quality": 0.7305, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.449, "max_latency_ms": 256414, "minimum_quality": 0.9583}, "family": "migration_planning", "id": "synthetic-migration_planning-000973", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.73, "estimated_latency_ms": 182670, "estimated_pass_rate": 0.7201, "estimated_quality": 0.9109, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.449, "max_latency_ms": 256414, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7201, "quality": 0.9109, "quality_target_met": false, "utility": 69.059295}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.449,\"max_latency_ms\":256414,\"minimum_quality\":0.9583},\"task\":{\"ambiguity\":0.716,\"category\":\"coding\",\"complexity\":0.6166,\"consequence\":0.5233,\"evidence_need\":0.6471,\"family\":\"migration_planning\",\"input_tokens\":2920,\"reversibility\":0.6096,\"risk\":0.4815,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.714,\"id\":\"worker_a\",\"p95_latency_ms\":82558,\"pass_rate\":0.9838,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.976,\"id\":\"worker_b\",\"p95_latency_ms\":76233,\"pass_rate\":0.8628,\"provider_group\":\"provider_z\",\"quality\":0.797,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.04,\"id\":\"worker_c\",\"p95_latency_ms\":23879,\"pass_rate\":0.474,\"provider_group\":\"provider_y\",\"quality\":0.6886,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.73,\"estimated_latency_ms\":182670,\"estimated_pass_rate\":0.7201,\"estimated_quality\":0.9109,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.449,\"max_latency_ms\":256414,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.716, "category": "coding", "complexity": 0.6166, "consequence": 0.5233, "evidence_need": 0.6471, "family": "migration_planning", "input_tokens": 2920, "request_id": "synthetic-migration_planning-000973", "reversibility": 0.6096, "risk": 0.4815, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 14.714, "id": "worker_a", "p95_latency_ms": 82558, "pass_rate": 0.9838, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.976, "id": "worker_b", "p95_latency_ms": 76233, "pass_rate": 0.8628, "provider_group": "provider_z", "quality": 0.797, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.04, "id": "worker_c", "p95_latency_ms": 23879, "pass_rate": 0.474, "provider_group": "provider_y", "quality": 0.6886, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.564, "max_latency_ms": 33524, "minimum_quality": 0.6203}, "family": "migration_planning", "id": "synthetic-migration_planning-000700", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.086, "estimated_latency_ms": 30113, "estimated_pass_rate": 0.3614, "estimated_quality": 0.6148, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.564, "max_latency_ms": 33524, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3614, "quality": 0.6148, "quality_target_met": false, "utility": 39.191767}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.564,\"max_latency_ms\":33524,\"minimum_quality\":0.6203},\"task\":{\"ambiguity\":0.7368,\"category\":\"coding\",\"complexity\":0.5838,\"consequence\":0.6887,\"evidence_need\":0.4905,\"family\":\"migration_planning\",\"input_tokens\":200,\"reversibility\":0.5469,\"risk\":0.5983,\"summary\":\"Review a staged migration for provider failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.086,\"id\":\"worker_a\",\"p95_latency_ms\":30113,\"pass_rate\":0.5318,\"provider_group\":\"provider_y\",\"quality\":0.7329,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.773,\"id\":\"worker_b\",\"p95_latency_ms\":88917,\"pass_rate\":0.9788,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.153,\"id\":\"worker_c\",\"p95_latency_ms\":72973,\"pass_rate\":0.8015,\"provider_group\":\"provider_z\",\"quality\":0.8131,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.086,\"estimated_latency_ms\":30113,\"estimated_pass_rate\":0.3614,\"estimated_quality\":0.6148,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.564,\"max_latency_ms\":33524,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7368, "category": "coding", "complexity": 0.5838, "consequence": 0.6887, "evidence_need": 0.4905, "family": "migration_planning", "input_tokens": 200, "request_id": "synthetic-migration_planning-000700", "reversibility": 0.5469, "risk": 0.5983, "summary": "Review a staged migration for provider failover and identify irreversible steps."}, "workers": [{"credits_per_call": 1.086, "id": "worker_a", "p95_latency_ms": 30113, "pass_rate": 0.5318, "provider_group": "provider_y", "quality": 0.7329, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.773, "id": "worker_b", "p95_latency_ms": 88917, "pass_rate": 0.9788, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.153, "id": "worker_c", "p95_latency_ms": 72973, "pass_rate": 0.8015, "provider_group": "provider_z", "quality": 0.8131, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 12.727, "max_latency_ms": 151234, "minimum_quality": 0.8858}, "family": "migration_planning", "id": "synthetic-migration_planning-000235", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.436, "estimated_latency_ms": 104713, "estimated_pass_rate": 0.6491, "estimated_quality": 0.823, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.727, "max_latency_ms": 151234, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6491, "quality": 0.823, "quality_target_met": false, "utility": 59.706371}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":12.727,\"max_latency_ms\":151234,\"minimum_quality\":0.8858},\"task\":{\"ambiguity\":0.7487,\"category\":\"coding\",\"complexity\":0.5743,\"consequence\":0.6388,\"evidence_need\":0.5397,\"family\":\"migration_planning\",\"input_tokens\":8000,\"reversibility\":0.3274,\"risk\":0.5667,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.962,\"id\":\"worker_a\",\"p95_latency_ms\":22645,\"pass_rate\":0.5732,\"provider_group\":\"provider_z\",\"quality\":0.784,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.474,\"id\":\"worker_b\",\"p95_latency_ms\":82068,\"pass_rate\":0.8427,\"provider_group\":\"provider_y\",\"quality\":0.8815,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.481,\"id\":\"worker_c\",\"p95_latency_ms\":64388,\"pass_rate\":0.9447,\"provider_group\":\"provider_x\",\"quality\":0.994,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.436,\"estimated_latency_ms\":104713,\"estimated_pass_rate\":0.6491,\"estimated_quality\":0.823,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.727,\"max_latency_ms\":151234,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7487, "category": "coding", "complexity": 0.5743, "consequence": 0.6388, "evidence_need": 0.5397, "family": "migration_planning", "input_tokens": 8000, "request_id": "synthetic-migration_planning-000235", "reversibility": 0.3274, "risk": 0.5667, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 0.962, "id": "worker_a", "p95_latency_ms": 22645, "pass_rate": 0.5732, "provider_group": "provider_z", "quality": 0.784, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.474, "id": "worker_b", "p95_latency_ms": 82068, "pass_rate": 0.8427, "provider_group": "provider_y", "quality": 0.8815, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.481, "id": "worker_c", "p95_latency_ms": 64388, "pass_rate": 0.9447, "provider_group": "provider_x", "quality": 0.994, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.389, "max_latency_ms": 289864, "minimum_quality": 0.7536}, "family": "migration_planning", "id": "synthetic-migration_planning-001267", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.803, "estimated_latency_ms": 156939, "estimated_pass_rate": 0.8892, "estimated_quality": 0.888, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.389, "max_latency_ms": 289864, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8892, "quality": 0.888, "quality_target_met": true, "utility": 70.911159}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.389,\"max_latency_ms\":289864,\"minimum_quality\":0.7536},\"task\":{\"ambiguity\":0.6854,\"category\":\"coding\",\"complexity\":0.5588,\"consequence\":0.7667,\"evidence_need\":0.7335,\"family\":\"migration_planning\",\"input_tokens\":1240,\"reversibility\":0.618,\"risk\":0.6939,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.208,\"id\":\"worker_a\",\"p95_latency_ms\":79715,\"pass_rate\":0.7809,\"provider_group\":\"provider_y\",\"quality\":0.8246,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.595,\"id\":\"worker_b\",\"p95_latency_ms\":77224,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9127,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.108,\"id\":\"worker_c\",\"p95_latency_ms\":31789,\"pass_rate\":0.5143,\"provider_group\":\"provider_z\",\"quality\":0.7072,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.803,\"estimated_latency_ms\":156939,\"estimated_pass_rate\":0.8892,\"estimated_quality\":0.888,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.389,\"max_latency_ms\":289864,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6854, "category": "coding", "complexity": 0.5588, "consequence": 0.7667, "evidence_need": 0.7335, "family": "migration_planning", "input_tokens": 1240, "request_id": "synthetic-migration_planning-001267", "reversibility": 0.618, "risk": 0.6939, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 3.208, "id": "worker_a", "p95_latency_ms": 79715, "pass_rate": 0.7809, "provider_group": "provider_y", "quality": 0.8246, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.595, "id": "worker_b", "p95_latency_ms": 77224, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9127, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.108, "id": "worker_c", "p95_latency_ms": 31789, "pass_rate": 0.5143, "provider_group": "provider_z", "quality": 0.7072, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.176, "max_latency_ms": 50392, "minimum_quality": 0.5355}, "family": "migration_planning", "id": "synthetic-migration_planning-000385", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.06, "estimated_latency_ms": 30484, "estimated_pass_rate": 0.3249, "estimated_quality": 0.685, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.176, "max_latency_ms": 50392, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3249, "quality": 0.685, "quality_target_met": true, "utility": 43.810556}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.176,\"max_latency_ms\":50392,\"minimum_quality\":0.5355},\"task\":{\"ambiguity\":0.459,\"category\":\"coding\",\"complexity\":0.7145,\"consequence\":0.4455,\"evidence_need\":0.8092,\"family\":\"migration_planning\",\"input_tokens\":120,\"reversibility\":0.4386,\"risk\":0.4443,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":1.06,\"id\":\"worker_a\",\"p95_latency_ms\":30484,\"pass_rate\":0.4766,\"provider_group\":\"provider_y\",\"quality\":0.797,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.585,\"id\":\"worker_b\",\"p95_latency_ms\":72421,\"pass_rate\":0.8414,\"provider_group\":\"provider_z\",\"quality\":0.8406,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.778,\"id\":\"worker_c\",\"p95_latency_ms\":55741,\"pass_rate\":0.9045,\"provider_group\":\"provider_x\",\"quality\":0.9728,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.06,\"estimated_latency_ms\":30484,\"estimated_pass_rate\":0.3249,\"estimated_quality\":0.685,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.176,\"max_latency_ms\":50392,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.459, "category": "coding", "complexity": 0.7145, "consequence": 0.4455, "evidence_need": 0.8092, "family": "migration_planning", "input_tokens": 120, "request_id": "synthetic-migration_planning-000385", "reversibility": 0.4386, "risk": 0.4443, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 1.06, "id": "worker_a", "p95_latency_ms": 30484, "pass_rate": 0.4766, "provider_group": "provider_y", "quality": 0.797, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.585, "id": "worker_b", "p95_latency_ms": 72421, "pass_rate": 0.8414, "provider_group": "provider_z", "quality": 0.8406, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.778, "id": "worker_c", "p95_latency_ms": 55741, "pass_rate": 0.9045, "provider_group": "provider_x", "quality": 0.9728, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.551, "max_latency_ms": 164421, "minimum_quality": 0.8957}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000210", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.318, "estimated_latency_ms": 158899, "estimated_pass_rate": 0.8472, "estimated_quality": 0.8984, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.551, "max_latency_ms": 164421, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8472, "quality": 0.8984, "quality_target_met": true, "utility": 68.572284}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.551,\"max_latency_ms\":164421,\"minimum_quality\":0.8957},\"task\":{\"ambiguity\":0.7422,\"category\":\"review\",\"complexity\":0.8124,\"consequence\":0.7686,\"evidence_need\":0.6006,\"family\":\"policy_compliance\",\"input_tokens\":3120,\"reversibility\":0.4016,\"risk\":0.6711,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.07,\"id\":\"worker_a\",\"p95_latency_ms\":84284,\"pass_rate\":0.9553,\"provider_group\":\"provider_y\",\"quality\":0.9213,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.974,\"id\":\"worker_b\",\"p95_latency_ms\":23818,\"pass_rate\":0.5121,\"provider_group\":\"provider_x\",\"quality\":0.7808,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.248,\"id\":\"worker_c\",\"p95_latency_ms\":74615,\"pass_rate\":0.8861,\"provider_group\":\"provider_z\",\"quality\":0.9143,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.318,\"estimated_latency_ms\":158899,\"estimated_pass_rate\":0.8472,\"estimated_quality\":0.8984,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.551,\"max_latency_ms\":164421,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7422, "category": "review", "complexity": 0.8124, "consequence": 0.7686, "evidence_need": 0.6006, "family": "policy_compliance", "input_tokens": 3120, "request_id": "synthetic-policy_compliance-000210", "reversibility": 0.4016, "risk": 0.6711, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 12.07, "id": "worker_a", "p95_latency_ms": 84284, "pass_rate": 0.9553, "provider_group": "provider_y", "quality": 0.9213, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.974, "id": "worker_b", "p95_latency_ms": 23818, "pass_rate": 0.5121, "provider_group": "provider_x", "quality": 0.7808, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.248, "id": "worker_c", "p95_latency_ms": 74615, "pass_rate": 0.8861, "provider_group": "provider_z", "quality": 0.9143, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.556, "max_latency_ms": 37399, "minimum_quality": 0.6413}, "family": "migration_planning", "id": "synthetic-migration_planning-000424", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.075, "estimated_latency_ms": 27962, "estimated_pass_rate": 0.4239, "estimated_quality": 0.5908, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.556, "max_latency_ms": 37399, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4239, "quality": 0.5908, "quality_target_met": false, "utility": 26.666382}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.556,\"max_latency_ms\":37399,\"minimum_quality\":0.6413},\"task\":{\"ambiguity\":0.3346,\"category\":\"coding\",\"complexity\":0.7177,\"consequence\":0.7552,\"evidence_need\":0.4899,\"family\":\"migration_planning\",\"input_tokens\":1520,\"reversibility\":0.4372,\"risk\":0.7434,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":2.969,\"id\":\"worker_a\",\"p95_latency_ms\":59909,\"pass_rate\":0.792,\"provider_group\":\"provider_y\",\"quality\":0.8894,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.34,\"id\":\"worker_b\",\"p95_latency_ms\":57888,\"pass_rate\":0.9182,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.075,\"id\":\"worker_c\",\"p95_latency_ms\":27962,\"pass_rate\":0.5908,\"provider_group\":\"provider_z\",\"quality\":0.7166,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.075,\"estimated_latency_ms\":27962,\"estimated_pass_rate\":0.4239,\"estimated_quality\":0.5908,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.556,\"max_latency_ms\":37399,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3346, "category": "coding", "complexity": 0.7177, "consequence": 0.7552, "evidence_need": 0.4899, "family": "migration_planning", "input_tokens": 1520, "request_id": "synthetic-migration_planning-000424", "reversibility": 0.4372, "risk": 0.7434, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 2.969, "id": "worker_a", "p95_latency_ms": 59909, "pass_rate": 0.792, "provider_group": "provider_y", "quality": 0.8894, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.34, "id": "worker_b", "p95_latency_ms": 57888, "pass_rate": 0.9182, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.075, "id": "worker_c", "p95_latency_ms": 27962, "pass_rate": 0.5908, "provider_group": "provider_z", "quality": 0.7166, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.581, "max_latency_ms": 89614, "minimum_quality": 0.6807}, "family": "migration_planning", "id": "synthetic-migration_planning-000184", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.043, "estimated_latency_ms": 86021, "estimated_pass_rate": 0.6265, "estimated_quality": 0.8481, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.581, "max_latency_ms": 89614, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6265, "quality": 0.8481, "quality_target_met": true, "utility": 61.720911}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.581,\"max_latency_ms\":89614,\"minimum_quality\":0.6807},\"task\":{\"ambiguity\":0.4921,\"category\":\"coding\",\"complexity\":0.7157,\"consequence\":0.5623,\"evidence_need\":0.6889,\"family\":\"migration_planning\",\"input_tokens\":4320,\"reversibility\":0.252,\"risk\":0.7295,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":11.03,\"id\":\"worker_a\",\"p95_latency_ms\":59742,\"pass_rate\":0.9601,\"provider_group\":\"provider_z\",\"quality\":0.9158,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.849,\"id\":\"worker_b\",\"p95_latency_ms\":22670,\"pass_rate\":0.4722,\"provider_group\":\"provider_x\",\"quality\":0.7922,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.194,\"id\":\"worker_c\",\"p95_latency_ms\":63351,\"pass_rate\":0.8433,\"provider_group\":\"provider_y\",\"quality\":0.9181,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.043,\"estimated_latency_ms\":86021,\"estimated_pass_rate\":0.6265,\"estimated_quality\":0.8481,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.581,\"max_latency_ms\":89614,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4921, "category": "coding", "complexity": 0.7157, "consequence": 0.5623, "evidence_need": 0.6889, "family": "migration_planning", "input_tokens": 4320, "request_id": "synthetic-migration_planning-000184", "reversibility": 0.252, "risk": 0.7295, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 11.03, "id": "worker_a", "p95_latency_ms": 59742, "pass_rate": 0.9601, "provider_group": "provider_z", "quality": 0.9158, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.849, "id": "worker_b", "p95_latency_ms": 22670, "pass_rate": 0.4722, "provider_group": "provider_x", "quality": 0.7922, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.194, "id": "worker_c", "p95_latency_ms": 63351, "pass_rate": 0.8433, "provider_group": "provider_y", "quality": 0.9181, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.183, "max_latency_ms": 131462, "minimum_quality": 0.7922}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000621", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.101, "estimated_latency_ms": 75587, "estimated_pass_rate": 0.6349, "estimated_quality": 0.8145, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.183, "max_latency_ms": 131462, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6349, "quality": 0.8145, "quality_target_met": true, "utility": 62.870739}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.183,\"max_latency_ms\":131462,\"minimum_quality\":0.7922},\"task\":{\"ambiguity\":0.3821,\"category\":\"review\",\"complexity\":0.5158,\"consequence\":0.7812,\"evidence_need\":0.8584,\"family\":\"policy_compliance\",\"input_tokens\":160,\"reversibility\":0.4322,\"risk\":0.6112,\"summary\":\"Review whether observability satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.001,\"id\":\"worker_a\",\"p95_latency_ms\":45421,\"pass_rate\":0.768,\"provider_group\":\"provider_z\",\"quality\":0.8251,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.065,\"id\":\"worker_b\",\"p95_latency_ms\":86567,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.1,\"id\":\"worker_c\",\"p95_latency_ms\":30166,\"pass_rate\":0.5681,\"provider_group\":\"provider_x\",\"quality\":0.7291,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.101,\"estimated_latency_ms\":75587,\"estimated_pass_rate\":0.6349,\"estimated_quality\":0.8145,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.183,\"max_latency_ms\":131462,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3821, "category": "review", "complexity": 0.5158, "consequence": 0.7812, "evidence_need": 0.8584, "family": "policy_compliance", "input_tokens": 160, "request_id": "synthetic-policy_compliance-000621", "reversibility": 0.4322, "risk": 0.6112, "summary": "Review whether observability satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.001, "id": "worker_a", "p95_latency_ms": 45421, "pass_rate": 0.768, "provider_group": "provider_z", "quality": 0.8251, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.065, "id": "worker_b", "p95_latency_ms": 86567, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.1, "id": "worker_c", "p95_latency_ms": 30166, "pass_rate": 0.5681, "provider_group": "provider_x", "quality": 0.7291, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.764, "max_latency_ms": 213132, "minimum_quality": 0.89}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000746", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.325, "estimated_latency_ms": 164833, "estimated_pass_rate": 0.7559, "estimated_quality": 0.9298, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 24.764, "max_latency_ms": 213132, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7559, "quality": 0.9298, "quality_target_met": true, "utility": 73.236666}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.764,\"max_latency_ms\":213132,\"minimum_quality\":0.89},\"task\":{\"ambiguity\":0.6242,\"category\":\"analysis\",\"complexity\":0.6833,\"consequence\":0.2908,\"evidence_need\":0.6476,\"family\":\"performance_tuning\",\"input_tokens\":440,\"reversibility\":0.5865,\"risk\":0.5112,\"summary\":\"Choose between throughput and tail-latency changes for private credentials.\"},\"workers\":[{\"credits_per_call\":13.1,\"id\":\"worker_a\",\"p95_latency_ms\":52144,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.212,\"id\":\"worker_b\",\"p95_latency_ms\":82416,\"pass_rate\":0.8689,\"provider_group\":\"provider_y\",\"quality\":0.9396,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.013,\"id\":\"worker_c\",\"p95_latency_ms\":30273,\"pass_rate\":0.4741,\"provider_group\":\"provider_z\",\"quality\":0.7363,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.325,\"estimated_latency_ms\":164833,\"estimated_pass_rate\":0.7559,\"estimated_quality\":0.9298,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.764,\"max_latency_ms\":213132,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6242, "category": "analysis", "complexity": 0.6833, "consequence": 0.2908, "evidence_need": 0.6476, "family": "performance_tuning", "input_tokens": 440, "request_id": "synthetic-performance_tuning-000746", "reversibility": 0.5865, "risk": 0.5112, "summary": "Choose between throughput and tail-latency changes for private credentials."}, "workers": [{"credits_per_call": 13.1, "id": "worker_a", "p95_latency_ms": 52144, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.212, "id": "worker_b", "p95_latency_ms": 82416, "pass_rate": 0.8689, "provider_group": "provider_y", "quality": 0.9396, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.013, "id": "worker_c", "p95_latency_ms": 30273, "pass_rate": 0.4741, "provider_group": "provider_z", "quality": 0.7363, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.964, "max_latency_ms": 47485, "minimum_quality": 0.5609}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000534", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.178, "estimated_latency_ms": 27970, "estimated_pass_rate": 0.3296, "estimated_quality": 0.6694, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.964, "max_latency_ms": 47485, "policy": "lean", "reason_codes": ["high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3296, "quality": 0.6694, "quality_target_met": true, "utility": 27.763527}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.964,\"max_latency_ms\":47485,\"minimum_quality\":0.5609},\"task\":{\"ambiguity\":0.6696,\"category\":\"review\",\"complexity\":0.4007,\"consequence\":0.85,\"evidence_need\":0.6211,\"family\":\"policy_compliance\",\"input_tokens\":880,\"reversibility\":0.5791,\"risk\":0.733,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.027,\"id\":\"worker_a\",\"p95_latency_ms\":79908,\"pass_rate\":0.8209,\"provider_group\":\"provider_x\",\"quality\":0.8704,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.962,\"id\":\"worker_b\",\"p95_latency_ms\":86720,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9164,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.178,\"id\":\"worker_c\",\"p95_latency_ms\":27970,\"pass_rate\":0.485,\"provider_group\":\"provider_z\",\"quality\":0.7596,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.178,\"estimated_latency_ms\":27970,\"estimated_pass_rate\":0.3296,\"estimated_quality\":0.6694,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.964,\"max_latency_ms\":47485,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6696, "category": "review", "complexity": 0.4007, "consequence": 0.85, "evidence_need": 0.6211, "family": "policy_compliance", "input_tokens": 880, "request_id": "synthetic-policy_compliance-000534", "reversibility": 0.5791, "risk": 0.733, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.027, "id": "worker_a", "p95_latency_ms": 79908, "pass_rate": 0.8209, "provider_group": "provider_x", "quality": 0.8704, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.962, "id": "worker_b", "p95_latency_ms": 86720, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9164, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.178, "id": "worker_c", "p95_latency_ms": 27970, "pass_rate": 0.485, "provider_group": "provider_z", "quality": 0.7596, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.9, "max_latency_ms": 205425, "minimum_quality": 0.918}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000435", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.162, "estimated_latency_ms": 159949, "estimated_pass_rate": 0.7203, "estimated_quality": 0.9176, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.9, "max_latency_ms": 205425, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7203, "quality": 0.9176, "quality_target_met": false, "utility": 75.30153}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.9,\"max_latency_ms\":205425,\"minimum_quality\":0.918},\"task\":{\"ambiguity\":0.7387,\"category\":\"review\",\"complexity\":0.8236,\"consequence\":0.6434,\"evidence_need\":0.8429,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.6331,\"risk\":0.4557,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.054,\"id\":\"worker_a\",\"p95_latency_ms\":74149,\"pass_rate\":0.753,\"provider_group\":\"provider_y\",\"quality\":0.8278,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.018,\"id\":\"worker_b\",\"p95_latency_ms\":24814,\"pass_rate\":0.5277,\"provider_group\":\"provider_x\",\"quality\":0.7653,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.09,\"id\":\"worker_c\",\"p95_latency_ms\":60986,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9891,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.162,\"estimated_latency_ms\":159949,\"estimated_pass_rate\":0.7203,\"estimated_quality\":0.9176,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.9,\"max_latency_ms\":205425,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7387, "category": "review", "complexity": 0.8236, "consequence": 0.6434, "evidence_need": 0.8429, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-000435", "reversibility": 0.6331, "risk": 0.4557, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.054, "id": "worker_a", "p95_latency_ms": 74149, "pass_rate": 0.753, "provider_group": "provider_y", "quality": 0.8278, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.018, "id": "worker_b", "p95_latency_ms": 24814, "pass_rate": 0.5277, "provider_group": "provider_x", "quality": 0.7653, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.09, "id": "worker_c", "p95_latency_ms": 60986, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9891, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.691, "max_latency_ms": 222170, "minimum_quality": 0.8438}, "family": "migration_planning", "id": "synthetic-migration_planning-000001", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.831, "estimated_latency_ms": 83699, "estimated_pass_rate": 0.5829, "estimated_quality": 0.7621, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.691, "max_latency_ms": 222170, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5829, "quality": 0.7621, "quality_target_met": false, "utility": 52.991603}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.691,\"max_latency_ms\":222170,\"minimum_quality\":0.8438},\"task\":{\"ambiguity\":0.5621,\"category\":\"coding\",\"complexity\":0.7942,\"consequence\":0.5131,\"evidence_need\":0.4701,\"family\":\"migration_planning\",\"input_tokens\":4520,\"reversibility\":0.4312,\"risk\":0.5227,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.972,\"id\":\"worker_a\",\"p95_latency_ms\":32100,\"pass_rate\":0.4847,\"provider_group\":\"provider_z\",\"quality\":0.7163,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.859,\"id\":\"worker_b\",\"p95_latency_ms\":51599,\"pass_rate\":0.8641,\"provider_group\":\"provider_x\",\"quality\":0.8527,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.588,\"id\":\"worker_c\",\"p95_latency_ms\":79539,\"pass_rate\":0.9051,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.831,\"estimated_latency_ms\":83699,\"estimated_pass_rate\":0.5829,\"estimated_quality\":0.7621,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.691,\"max_latency_ms\":222170,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5621, "category": "coding", "complexity": 0.7942, "consequence": 0.5131, "evidence_need": 0.4701, "family": "migration_planning", "input_tokens": 4520, "request_id": "synthetic-migration_planning-000001", "reversibility": 0.4312, "risk": 0.5227, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 0.972, "id": "worker_a", "p95_latency_ms": 32100, "pass_rate": 0.4847, "provider_group": "provider_z", "quality": 0.7163, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.859, "id": "worker_b", "p95_latency_ms": 51599, "pass_rate": 0.8641, "provider_group": "provider_x", "quality": 0.8527, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.588, "id": "worker_c", "p95_latency_ms": 79539, "pass_rate": 0.9051, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.28, "max_latency_ms": 207090, "minimum_quality": 0.8474}, "family": "migration_planning", "id": "synthetic-migration_planning-000559", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.725, "estimated_latency_ms": 188626, "estimated_pass_rate": 0.7612, "estimated_quality": 0.874, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.28, "max_latency_ms": 207090, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7612, "quality": 0.874, "quality_target_met": true, "utility": 70.529993}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.28,\"max_latency_ms\":207090,\"minimum_quality\":0.8474},\"task\":{\"ambiguity\":0.4564,\"category\":\"coding\",\"complexity\":0.9135,\"consequence\":0.504,\"evidence_need\":0.4354,\"family\":\"migration_planning\",\"input_tokens\":4280,\"reversibility\":0.2516,\"risk\":0.5905,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.424,\"id\":\"worker_a\",\"p95_latency_ms\":78445,\"pass_rate\":0.7745,\"provider_group\":\"provider_y\",\"quality\":0.9195,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.426,\"id\":\"worker_b\",\"p95_latency_ms\":83931,\"pass_rate\":0.9609,\"provider_group\":\"provider_z\",\"quality\":0.9286,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.875,\"id\":\"worker_c\",\"p95_latency_ms\":26250,\"pass_rate\":0.5853,\"provider_group\":\"provider_x\",\"quality\":0.7342,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.725,\"estimated_latency_ms\":188626,\"estimated_pass_rate\":0.7612,\"estimated_quality\":0.874,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.28,\"max_latency_ms\":207090,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4564, "category": "coding", "complexity": 0.9135, "consequence": 0.504, "evidence_need": 0.4354, "family": "migration_planning", "input_tokens": 4280, "request_id": "synthetic-migration_planning-000559", "reversibility": 0.2516, "risk": 0.5905, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.424, "id": "worker_a", "p95_latency_ms": 78445, "pass_rate": 0.7745, "provider_group": "provider_y", "quality": 0.9195, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.426, "id": "worker_b", "p95_latency_ms": 83931, "pass_rate": 0.9609, "provider_group": "provider_z", "quality": 0.9286, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.875, "id": "worker_c", "p95_latency_ms": 26250, "pass_rate": 0.5853, "provider_group": "provider_x", "quality": 0.7342, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.501, "max_latency_ms": 205905, "minimum_quality": 0.7863}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000296", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.914, "estimated_latency_ms": 177565, "estimated_pass_rate": 0.7761, "estimated_quality": 0.9192, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.501, "max_latency_ms": 205905, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7761, "quality": 0.9192, "quality_target_met": true, "utility": 72.378189}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.501,\"max_latency_ms\":205905,\"minimum_quality\":0.7863},\"task\":{\"ambiguity\":0.6109,\"category\":\"analysis\",\"complexity\":0.7359,\"consequence\":0.3387,\"evidence_need\":0.4504,\"family\":\"performance_tuning\",\"input_tokens\":6040,\"reversibility\":0.5906,\"risk\":0.5233,\"summary\":\"Choose between throughput and tail-latency changes for private credentials.\"},\"workers\":[{\"credits_per_call\":11.969,\"id\":\"worker_a\",\"p95_latency_ms\":79201,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.944,\"id\":\"worker_b\",\"p95_latency_ms\":76757,\"pass_rate\":0.8467,\"provider_group\":\"provider_z\",\"quality\":0.9611,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.001,\"id\":\"worker_c\",\"p95_latency_ms\":21607,\"pass_rate\":0.5354,\"provider_group\":\"provider_y\",\"quality\":0.7261,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.914,\"estimated_latency_ms\":177565,\"estimated_pass_rate\":0.7761,\"estimated_quality\":0.9192,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.501,\"max_latency_ms\":205905,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6109, "category": "analysis", "complexity": 0.7359, "consequence": 0.3387, "evidence_need": 0.4504, "family": "performance_tuning", "input_tokens": 6040, "request_id": "synthetic-performance_tuning-000296", "reversibility": 0.5906, "risk": 0.5233, "summary": "Choose between throughput and tail-latency changes for private credentials."}, "workers": [{"credits_per_call": 11.969, "id": "worker_a", "p95_latency_ms": 79201, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.944, "id": "worker_b", "p95_latency_ms": 76757, "pass_rate": 0.8467, "provider_group": "provider_z", "quality": 0.9611, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.001, "id": "worker_c", "p95_latency_ms": 21607, "pass_rate": 0.5354, "provider_group": "provider_y", "quality": 0.7261, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.855, "max_latency_ms": 35665, "minimum_quality": 0.5907}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000101", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.126, "estimated_latency_ms": 24207, "estimated_pass_rate": 0.3673, "estimated_quality": 0.5327, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.855, "max_latency_ms": 35665, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3673, "quality": 0.5327, "quality_target_met": false, "utility": 33.224232}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.855,\"max_latency_ms\":35665,\"minimum_quality\":0.5907},\"task\":{\"ambiguity\":0.4409,\"category\":\"analysis\",\"complexity\":0.6225,\"consequence\":0.5134,\"evidence_need\":0.5235,\"family\":\"performance_tuning\",\"input_tokens\":3440,\"reversibility\":0.6322,\"risk\":0.5208,\"summary\":\"Review a proposed optimization for artifact storage for hidden regressions.\"},\"workers\":[{\"credits_per_call\":4.031,\"id\":\"worker_a\",\"p95_latency_ms\":65390,\"pass_rate\":0.9153,\"provider_group\":\"provider_z\",\"quality\":0.9381,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.981,\"id\":\"worker_b\",\"p95_latency_ms\":60117,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.126,\"id\":\"worker_c\",\"p95_latency_ms\":24207,\"pass_rate\":0.5134,\"provider_group\":\"provider_x\",\"quality\":0.6324,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.126,\"estimated_latency_ms\":24207,\"estimated_pass_rate\":0.3673,\"estimated_quality\":0.5327,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.855,\"max_latency_ms\":35665,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4409, "category": "analysis", "complexity": 0.6225, "consequence": 0.5134, "evidence_need": 0.5235, "family": "performance_tuning", "input_tokens": 3440, "request_id": "synthetic-performance_tuning-000101", "reversibility": 0.6322, "risk": 0.5208, "summary": "Review a proposed optimization for artifact storage for hidden regressions."}, "workers": [{"credits_per_call": 4.031, "id": "worker_a", "p95_latency_ms": 65390, "pass_rate": 0.9153, "provider_group": "provider_z", "quality": 0.9381, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.981, "id": "worker_b", "p95_latency_ms": 60117, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.126, "id": "worker_c", "p95_latency_ms": 24207, "pass_rate": 0.5134, "provider_group": "provider_x", "quality": 0.6324, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.12, "max_latency_ms": 147771, "minimum_quality": 0.8567}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001155", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.864, "estimated_latency_ms": 142862, "estimated_pass_rate": 0.7462, "estimated_quality": 0.9193, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.12, "max_latency_ms": 147771, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7462, "quality": 0.9193, "quality_target_met": true, "utility": 79.324691}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.12,\"max_latency_ms\":147771,\"minimum_quality\":0.8567},\"task\":{\"ambiguity\":0.4951,\"category\":\"review\",\"complexity\":0.8162,\"consequence\":0.7071,\"evidence_need\":0.8245,\"family\":\"policy_compliance\",\"input_tokens\":7360,\"reversibility\":0.3616,\"risk\":0.8401,\"summary\":\"Review whether observability satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.309,\"id\":\"worker_a\",\"p95_latency_ms\":45641,\"pass_rate\":0.7843,\"provider_group\":\"provider_x\",\"quality\":0.852,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.796,\"id\":\"worker_b\",\"p95_latency_ms\":30818,\"pass_rate\":0.4844,\"provider_group\":\"provider_y\",\"quality\":0.7452,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.759,\"id\":\"worker_c\",\"p95_latency_ms\":66403,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.864,\"estimated_latency_ms\":142862,\"estimated_pass_rate\":0.7462,\"estimated_quality\":0.9193,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.12,\"max_latency_ms\":147771,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4951, "category": "review", "complexity": 0.8162, "consequence": 0.7071, "evidence_need": 0.8245, "family": "policy_compliance", "input_tokens": 7360, "request_id": "synthetic-policy_compliance-001155", "reversibility": 0.3616, "risk": 0.8401, "summary": "Review whether observability satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.309, "id": "worker_a", "p95_latency_ms": 45641, "pass_rate": 0.7843, "provider_group": "provider_x", "quality": 0.852, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.796, "id": "worker_b", "p95_latency_ms": 30818, "pass_rate": 0.4844, "provider_group": "provider_y", "quality": 0.7452, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.759, "id": "worker_c", "p95_latency_ms": 66403, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.327, "max_latency_ms": 212611, "minimum_quality": 0.8473}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000293", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.074, "estimated_latency_ms": 82049, "estimated_pass_rate": 0.7201, "estimated_quality": 0.8479, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.327, "max_latency_ms": 212611, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7201, "quality": 0.8479, "quality_target_met": true, "utility": 59.740707}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.327,\"max_latency_ms\":212611,\"minimum_quality\":0.8473},\"task\":{\"ambiguity\":0.523,\"category\":\"analysis\",\"complexity\":0.8147,\"consequence\":0.5262,\"evidence_need\":0.6579,\"family\":\"performance_tuning\",\"input_tokens\":4560,\"reversibility\":0.6351,\"risk\":0.4084,\"summary\":\"Diagnose latency variance in prompt caching and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.972,\"id\":\"worker_a\",\"p95_latency_ms\":22409,\"pass_rate\":0.4143,\"provider_group\":\"provider_y\",\"quality\":0.6451,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.102,\"id\":\"worker_b\",\"p95_latency_ms\":59640,\"pass_rate\":0.8521,\"provider_group\":\"provider_z\",\"quality\":0.9738,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.267,\"id\":\"worker_c\",\"p95_latency_ms\":83375,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.074,\"estimated_latency_ms\":82049,\"estimated_pass_rate\":0.7201,\"estimated_quality\":0.8479,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.327,\"max_latency_ms\":212611,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.523, "category": "analysis", "complexity": 0.8147, "consequence": 0.5262, "evidence_need": 0.6579, "family": "performance_tuning", "input_tokens": 4560, "request_id": "synthetic-performance_tuning-000293", "reversibility": 0.6351, "risk": 0.4084, "summary": "Diagnose latency variance in prompt caching and rank the next experiments."}, "workers": [{"credits_per_call": 0.972, "id": "worker_a", "p95_latency_ms": 22409, "pass_rate": 0.4143, "provider_group": "provider_y", "quality": 0.6451, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.102, "id": "worker_b", "p95_latency_ms": 59640, "pass_rate": 0.8521, "provider_group": "provider_z", "quality": 0.9738, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.267, "id": "worker_c", "p95_latency_ms": 83375, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.759, "max_latency_ms": 25841, "minimum_quality": 0.6512}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000432", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.953, "estimated_latency_ms": 23157, "estimated_pass_rate": 0.3639, "estimated_quality": 0.586, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.759, "max_latency_ms": 25841, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3639, "quality": 0.586, "quality_target_met": false, "utility": 16.194087}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.759,\"max_latency_ms\":25841,\"minimum_quality\":0.6512},\"task\":{\"ambiguity\":0.6919,\"category\":\"review\",\"complexity\":0.6118,\"consequence\":0.9316,\"evidence_need\":0.9102,\"family\":\"policy_compliance\",\"input_tokens\":120,\"reversibility\":0.5413,\"risk\":0.7718,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.719,\"id\":\"worker_a\",\"p95_latency_ms\":71099,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9582,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.075,\"id\":\"worker_b\",\"p95_latency_ms\":70461,\"pass_rate\":0.8659,\"provider_group\":\"provider_x\",\"quality\":0.9184,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.953,\"id\":\"worker_c\",\"p95_latency_ms\":23157,\"pass_rate\":0.5481,\"provider_group\":\"provider_y\",\"quality\":0.7187,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.953,\"estimated_latency_ms\":23157,\"estimated_pass_rate\":0.3639,\"estimated_quality\":0.586,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.759,\"max_latency_ms\":25841,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6919, "category": "review", "complexity": 0.6118, "consequence": 0.9316, "evidence_need": 0.9102, "family": "policy_compliance", "input_tokens": 120, "request_id": "synthetic-policy_compliance-000432", "reversibility": 0.5413, "risk": 0.7718, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 12.719, "id": "worker_a", "p95_latency_ms": 71099, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9582, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.075, "id": "worker_b", "p95_latency_ms": 70461, "pass_rate": 0.8659, "provider_group": "provider_x", "quality": 0.9184, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.953, "id": "worker_c", "p95_latency_ms": 23157, "pass_rate": 0.5481, "provider_group": "provider_y", "quality": 0.7187, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.75, "max_latency_ms": 193340, "minimum_quality": 0.9326}, "family": "migration_planning", "id": "synthetic-migration_planning-000118", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.701, "estimated_latency_ms": 164881, "estimated_pass_rate": 0.7728, "estimated_quality": 0.9051, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.75, "max_latency_ms": 193340, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7728, "quality": 0.9051, "quality_target_met": false, "utility": 73.223362}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.75,\"max_latency_ms\":193340,\"minimum_quality\":0.9326},\"task\":{\"ambiguity\":0.3922,\"category\":\"coding\",\"complexity\":0.8877,\"consequence\":0.8425,\"evidence_need\":0.4028,\"family\":\"migration_planning\",\"input_tokens\":4600,\"reversibility\":0.5649,\"risk\":0.5904,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":10.701,\"id\":\"worker_a\",\"p95_latency_ms\":67385,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.975,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.806,\"id\":\"worker_b\",\"p95_latency_ms\":24717,\"pass_rate\":0.6034,\"provider_group\":\"provider_y\",\"quality\":0.7904,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.194,\"id\":\"worker_c\",\"p95_latency_ms\":72779,\"pass_rate\":0.7501,\"provider_group\":\"provider_x\",\"quality\":0.8775,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.701,\"estimated_latency_ms\":164881,\"estimated_pass_rate\":0.7728,\"estimated_quality\":0.9051,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.75,\"max_latency_ms\":193340,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3922, "category": "coding", "complexity": 0.8877, "consequence": 0.8425, "evidence_need": 0.4028, "family": "migration_planning", "input_tokens": 4600, "request_id": "synthetic-migration_planning-000118", "reversibility": 0.5649, "risk": 0.5904, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 10.701, "id": "worker_a", "p95_latency_ms": 67385, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.975, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.806, "id": "worker_b", "p95_latency_ms": 24717, "pass_rate": 0.6034, "provider_group": "provider_y", "quality": 0.7904, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.194, "id": "worker_c", "p95_latency_ms": 72779, "pass_rate": 0.7501, "provider_group": "provider_x", "quality": 0.8775, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.629, "max_latency_ms": 172681, "minimum_quality": 0.8866}, "family": "migration_planning", "id": "synthetic-migration_planning-000748", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.869, "estimated_latency_ms": 146478, "estimated_pass_rate": 0.7263, "estimated_quality": 0.9007, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.629, "max_latency_ms": 172681, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7263, "quality": 0.9007, "quality_target_met": true, "utility": 72.557943}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.629,\"max_latency_ms\":172681,\"minimum_quality\":0.8866},\"task\":{\"ambiguity\":0.66,\"category\":\"coding\",\"complexity\":0.8699,\"consequence\":0.4612,\"evidence_need\":0.5712,\"family\":\"migration_planning\",\"input_tokens\":1000,\"reversibility\":0.4352,\"risk\":0.5189,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":12.744,\"id\":\"worker_a\",\"p95_latency_ms\":75407,\"pass_rate\":0.9623,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.026,\"id\":\"worker_b\",\"p95_latency_ms\":23827,\"pass_rate\":0.5258,\"provider_group\":\"provider_y\",\"quality\":0.68,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.099,\"id\":\"worker_c\",\"p95_latency_ms\":47244,\"pass_rate\":0.8536,\"provider_group\":\"provider_x\",\"quality\":0.8114,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.869,\"estimated_latency_ms\":146478,\"estimated_pass_rate\":0.7263,\"estimated_quality\":0.9007,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.629,\"max_latency_ms\":172681,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.66, "category": "coding", "complexity": 0.8699, "consequence": 0.4612, "evidence_need": 0.5712, "family": "migration_planning", "input_tokens": 1000, "request_id": "synthetic-migration_planning-000748", "reversibility": 0.4352, "risk": 0.5189, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 12.744, "id": "worker_a", "p95_latency_ms": 75407, "pass_rate": 0.9623, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.026, "id": "worker_b", "p95_latency_ms": 23827, "pass_rate": 0.5258, "provider_group": "provider_y", "quality": 0.68, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.099, "id": "worker_c", "p95_latency_ms": 47244, "pass_rate": 0.8536, "provider_group": "provider_x", "quality": 0.8114, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.647, "max_latency_ms": 158974, "minimum_quality": 0.6882}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000720", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.649, "estimated_latency_ms": 107544, "estimated_pass_rate": 0.6147, "estimated_quality": 0.8125, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.647, "max_latency_ms": 158974, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6147, "quality": 0.8125, "quality_target_met": true, "utility": 64.138531}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.647,\"max_latency_ms\":158974,\"minimum_quality\":0.6882},\"task\":{\"ambiguity\":0.6098,\"category\":\"review\",\"complexity\":0.5219,\"consequence\":0.7475,\"evidence_need\":0.7668,\"family\":\"policy_compliance\",\"input_tokens\":1880,\"reversibility\":0.4864,\"risk\":0.7709,\"summary\":\"Review whether provider failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.772,\"id\":\"worker_a\",\"p95_latency_ms\":64071,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9401,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.798,\"id\":\"worker_b\",\"p95_latency_ms\":28335,\"pass_rate\":0.4678,\"provider_group\":\"provider_x\",\"quality\":0.697,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.851,\"id\":\"worker_c\",\"p95_latency_ms\":79209,\"pass_rate\":0.7965,\"provider_group\":\"provider_z\",\"quality\":0.9081,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.649,\"estimated_latency_ms\":107544,\"estimated_pass_rate\":0.6147,\"estimated_quality\":0.8125,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.647,\"max_latency_ms\":158974,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6098, "category": "review", "complexity": 0.5219, "consequence": 0.7475, "evidence_need": 0.7668, "family": "policy_compliance", "input_tokens": 1880, "request_id": "synthetic-policy_compliance-000720", "reversibility": 0.4864, "risk": 0.7709, "summary": "Review whether provider failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.772, "id": "worker_a", "p95_latency_ms": 64071, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9401, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.798, "id": "worker_b", "p95_latency_ms": 28335, "pass_rate": 0.4678, "provider_group": "provider_x", "quality": 0.697, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.851, "id": "worker_c", "p95_latency_ms": 79209, "pass_rate": 0.7965, "provider_group": "provider_z", "quality": 0.9081, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.001, "max_latency_ms": 38536, "minimum_quality": 0.6228}, "family": "migration_planning", "id": "synthetic-migration_planning-000238", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.064, "estimated_latency_ms": 25804, "estimated_pass_rate": 0.3347, "estimated_quality": 0.5943, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.001, "max_latency_ms": 38536, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3347, "quality": 0.5943, "quality_target_met": false, "utility": 37.852626}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.001,\"max_latency_ms\":38536,\"minimum_quality\":0.6228},\"task\":{\"ambiguity\":0.6603,\"category\":\"coding\",\"complexity\":0.6291,\"consequence\":0.4993,\"evidence_need\":0.8437,\"family\":\"migration_planning\",\"input_tokens\":4880,\"reversibility\":0.4409,\"risk\":0.556,\"summary\":\"Plan a zero-downtime migration of usage metering with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.186,\"id\":\"worker_a\",\"p95_latency_ms\":50087,\"pass_rate\":0.8279,\"provider_group\":\"provider_x\",\"quality\":0.8332,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.669,\"id\":\"worker_b\",\"p95_latency_ms\":85249,\"pass_rate\":0.985,\"provider_group\":\"provider_z\",\"quality\":0.9278,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.064,\"id\":\"worker_c\",\"p95_latency_ms\":25804,\"pass_rate\":0.501,\"provider_group\":\"provider_y\",\"quality\":0.7124,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.064,\"estimated_latency_ms\":25804,\"estimated_pass_rate\":0.3347,\"estimated_quality\":0.5943,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.001,\"max_latency_ms\":38536,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6603, "category": "coding", "complexity": 0.6291, "consequence": 0.4993, "evidence_need": 0.8437, "family": "migration_planning", "input_tokens": 4880, "request_id": "synthetic-migration_planning-000238", "reversibility": 0.4409, "risk": 0.556, "summary": "Plan a zero-downtime migration of usage metering with rollback checkpoints."}, "workers": [{"credits_per_call": 4.186, "id": "worker_a", "p95_latency_ms": 50087, "pass_rate": 0.8279, "provider_group": "provider_x", "quality": 0.8332, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.669, "id": "worker_b", "p95_latency_ms": 85249, "pass_rate": 0.985, "provider_group": "provider_z", "quality": 0.9278, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.064, "id": "worker_c", "p95_latency_ms": 25804, "pass_rate": 0.501, "provider_group": "provider_y", "quality": 0.7124, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.966, "max_latency_ms": 88411, "minimum_quality": 0.7345}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000102", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.416, "estimated_latency_ms": 81805, "estimated_pass_rate": 0.8145, "estimated_quality": 0.8821, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.966, "max_latency_ms": 88411, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8145, "quality": 0.8821, "quality_target_met": true, "utility": 62.963033}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.966,\"max_latency_ms\":88411,\"minimum_quality\":0.7345},\"task\":{\"ambiguity\":0.5413,\"category\":\"review\",\"complexity\":0.8269,\"consequence\":0.7593,\"evidence_need\":0.5632,\"family\":\"policy_compliance\",\"input_tokens\":4040,\"reversibility\":0.4319,\"risk\":0.5949,\"summary\":\"Audit a workflow for webhook delivery against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.063,\"id\":\"worker_a\",\"p95_latency_ms\":22788,\"pass_rate\":0.6154,\"provider_group\":\"provider_y\",\"quality\":0.7491,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.353,\"id\":\"worker_b\",\"p95_latency_ms\":59017,\"pass_rate\":0.9386,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.528,\"id\":\"worker_c\",\"p95_latency_ms\":65335,\"pass_rate\":0.8361,\"provider_group\":\"provider_x\",\"quality\":0.8469,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.416,\"estimated_latency_ms\":81805,\"estimated_pass_rate\":0.8145,\"estimated_quality\":0.8821,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.966,\"max_latency_ms\":88411,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5413, "category": "review", "complexity": 0.8269, "consequence": 0.7593, "evidence_need": 0.5632, "family": "policy_compliance", "input_tokens": 4040, "request_id": "synthetic-policy_compliance-000102", "reversibility": 0.4319, "risk": 0.5949, "summary": "Audit a workflow for webhook delivery against explicit access requirements."}, "workers": [{"credits_per_call": 1.063, "id": "worker_a", "p95_latency_ms": 22788, "pass_rate": 0.6154, "provider_group": "provider_y", "quality": 0.7491, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.353, "id": "worker_b", "p95_latency_ms": 59017, "pass_rate": 0.9386, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.528, "id": "worker_c", "p95_latency_ms": 65335, "pass_rate": 0.8361, "provider_group": "provider_x", "quality": 0.8469, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.146, "max_latency_ms": 241777, "minimum_quality": 0.9598}, "family": "migration_planning", "id": "synthetic-migration_planning-000889", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.483, "estimated_latency_ms": 187298, "estimated_pass_rate": 0.771, "estimated_quality": 0.914, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.146, "max_latency_ms": 241777, "policy": "strict", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.771, "quality": 0.914, "quality_target_met": false, "utility": 75.052906}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.146,\"max_latency_ms\":241777,\"minimum_quality\":0.9598},\"task\":{\"ambiguity\":0.663,\"category\":\"coding\",\"complexity\":0.6915,\"consequence\":0.5531,\"evidence_need\":0.7216,\"family\":\"migration_planning\",\"input_tokens\":360,\"reversibility\":0.3378,\"risk\":0.7117,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.89,\"id\":\"worker_a\",\"p95_latency_ms\":22585,\"pass_rate\":0.5146,\"provider_group\":\"provider_x\",\"quality\":0.6827,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.228,\"id\":\"worker_b\",\"p95_latency_ms\":81908,\"pass_rate\":0.85,\"provider_group\":\"provider_y\",\"quality\":0.8946,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.365,\"id\":\"worker_c\",\"p95_latency_ms\":82805,\"pass_rate\":0.9921,\"provider_group\":\"provider_z\",\"quality\":0.9857,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.483,\"estimated_latency_ms\":187298,\"estimated_pass_rate\":0.771,\"estimated_quality\":0.914,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.146,\"max_latency_ms\":241777,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.663, "category": "coding", "complexity": 0.6915, "consequence": 0.5531, "evidence_need": 0.7216, "family": "migration_planning", "input_tokens": 360, "request_id": "synthetic-migration_planning-000889", "reversibility": 0.3378, "risk": 0.7117, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 0.89, "id": "worker_a", "p95_latency_ms": 22585, "pass_rate": 0.5146, "provider_group": "provider_x", "quality": 0.6827, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.228, "id": "worker_b", "p95_latency_ms": 81908, "pass_rate": 0.85, "provider_group": "provider_y", "quality": 0.8946, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.365, "id": "worker_c", "p95_latency_ms": 82805, "pass_rate": 0.9921, "provider_group": "provider_z", "quality": 0.9857, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.84, "max_latency_ms": 40581, "minimum_quality": 0.709}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000794", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.07, "estimated_latency_ms": 23092, "estimated_pass_rate": 0.3172, "estimated_quality": 0.6141, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.84, "max_latency_ms": 40581, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3172, "quality": 0.6141, "quality_target_met": false, "utility": 34.752691}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.84,\"max_latency_ms\":40581,\"minimum_quality\":0.709},\"task\":{\"ambiguity\":0.7495,\"category\":\"analysis\",\"complexity\":0.6072,\"consequence\":0.4956,\"evidence_need\":0.629,\"family\":\"performance_tuning\",\"input_tokens\":2080,\"reversibility\":0.729,\"risk\":0.2877,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.07,\"id\":\"worker_a\",\"p95_latency_ms\":23092,\"pass_rate\":0.466,\"provider_group\":\"provider_x\",\"quality\":0.715,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.535,\"id\":\"worker_b\",\"p95_latency_ms\":47256,\"pass_rate\":0.802,\"provider_group\":\"provider_y\",\"quality\":0.9373,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.417,\"id\":\"worker_c\",\"p95_latency_ms\":72220,\"pass_rate\":0.9678,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.07,\"estimated_latency_ms\":23092,\"estimated_pass_rate\":0.3172,\"estimated_quality\":0.6141,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.84,\"max_latency_ms\":40581,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7495, "category": "analysis", "complexity": 0.6072, "consequence": 0.4956, "evidence_need": 0.629, "family": "performance_tuning", "input_tokens": 2080, "request_id": "synthetic-performance_tuning-000794", "reversibility": 0.729, "risk": 0.2877, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 1.07, "id": "worker_a", "p95_latency_ms": 23092, "pass_rate": 0.466, "provider_group": "provider_x", "quality": 0.715, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.535, "id": "worker_b", "p95_latency_ms": 47256, "pass_rate": 0.802, "provider_group": "provider_y", "quality": 0.9373, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.417, "id": "worker_c", "p95_latency_ms": 72220, "pass_rate": 0.9678, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.131, "max_latency_ms": 120825, "minimum_quality": 0.7216}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000096", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.305, "estimated_latency_ms": 63091, "estimated_pass_rate": 0.7395, "estimated_quality": 0.8682, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.131, "max_latency_ms": 120825, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7395, "quality": 0.8682, "quality_target_met": true, "utility": 66.628638}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.131,\"max_latency_ms\":120825,\"minimum_quality\":0.7216},\"task\":{\"ambiguity\":0.7065,\"category\":\"review\",\"complexity\":0.4375,\"consequence\":0.8925,\"evidence_need\":0.6814,\"family\":\"policy_compliance\",\"input_tokens\":1800,\"reversibility\":0.3359,\"risk\":0.5633,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":11.483,\"id\":\"worker_a\",\"p95_latency_ms\":90278,\"pass_rate\":0.983,\"provider_group\":\"provider_y\",\"quality\":0.9497,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.961,\"id\":\"worker_b\",\"p95_latency_ms\":18210,\"pass_rate\":0.5534,\"provider_group\":\"provider_x\",\"quality\":0.774,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.344,\"id\":\"worker_c\",\"p95_latency_ms\":44881,\"pass_rate\":0.8136,\"provider_group\":\"provider_z\",\"quality\":0.9166,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.305,\"estimated_latency_ms\":63091,\"estimated_pass_rate\":0.7395,\"estimated_quality\":0.8682,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.131,\"max_latency_ms\":120825,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7065, "category": "review", "complexity": 0.4375, "consequence": 0.8925, "evidence_need": 0.6814, "family": "policy_compliance", "input_tokens": 1800, "request_id": "synthetic-policy_compliance-000096", "reversibility": 0.3359, "risk": 0.5633, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 11.483, "id": "worker_a", "p95_latency_ms": 90278, "pass_rate": 0.983, "provider_group": "provider_y", "quality": 0.9497, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.961, "id": "worker_b", "p95_latency_ms": 18210, "pass_rate": 0.5534, "provider_group": "provider_x", "quality": 0.774, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.344, "id": "worker_c", "p95_latency_ms": 44881, "pass_rate": 0.8136, "provider_group": "provider_z", "quality": 0.9166, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.208, "max_latency_ms": 106136, "minimum_quality": 0.7402}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000981", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.904, "estimated_latency_ms": 75524, "estimated_pass_rate": 0.6626, "estimated_quality": 0.799, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.208, "max_latency_ms": 106136, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6626, "quality": 0.799, "quality_target_met": true, "utility": 61.375786}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.208,\"max_latency_ms\":106136,\"minimum_quality\":0.7402},\"task\":{\"ambiguity\":0.7519,\"category\":\"review\",\"complexity\":0.5858,\"consequence\":0.6306,\"evidence_need\":0.7718,\"family\":\"policy_compliance\",\"input_tokens\":7600,\"reversibility\":0.5042,\"risk\":0.8074,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.774,\"id\":\"worker_a\",\"p95_latency_ms\":78840,\"pass_rate\":0.9057,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.776,\"id\":\"worker_b\",\"p95_latency_ms\":49596,\"pass_rate\":0.8153,\"provider_group\":\"provider_x\",\"quality\":0.8708,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.128,\"id\":\"worker_c\",\"p95_latency_ms\":25928,\"pass_rate\":0.5758,\"provider_group\":\"provider_y\",\"quality\":0.6952,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.904,\"estimated_latency_ms\":75524,\"estimated_pass_rate\":0.6626,\"estimated_quality\":0.799,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.208,\"max_latency_ms\":106136,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7519, "category": "review", "complexity": 0.5858, "consequence": 0.6306, "evidence_need": 0.7718, "family": "policy_compliance", "input_tokens": 7600, "request_id": "synthetic-policy_compliance-000981", "reversibility": 0.5042, "risk": 0.8074, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 12.774, "id": "worker_a", "p95_latency_ms": 78840, "pass_rate": 0.9057, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.776, "id": "worker_b", "p95_latency_ms": 49596, "pass_rate": 0.8153, "provider_group": "provider_x", "quality": 0.8708, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.128, "id": "worker_c", "p95_latency_ms": 25928, "pass_rate": 0.5758, "provider_group": "provider_y", "quality": 0.6952, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 22.158, "max_latency_ms": 226192, "minimum_quality": 0.819}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000011", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.835, "estimated_latency_ms": 135090, "estimated_pass_rate": 0.8905, "estimated_quality": 0.9527, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 22.158, "max_latency_ms": 226192, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8905, "quality": 0.9527, "quality_target_met": true, "utility": 72.577948}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":22.158,\"max_latency_ms\":226192,\"minimum_quality\":0.819},\"task\":{\"ambiguity\":0.4208,\"category\":\"analysis\",\"complexity\":0.6344,\"consequence\":0.6544,\"evidence_need\":0.7005,\"family\":\"performance_tuning\",\"input_tokens\":1960,\"reversibility\":0.8001,\"risk\":0.3731,\"summary\":\"Choose between throughput and tail-latency changes for database failover.\"},\"workers\":[{\"credits_per_call\":12.663,\"id\":\"worker_a\",\"p95_latency_ms\":53766,\"pass_rate\":0.9575,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.172,\"id\":\"worker_b\",\"p95_latency_ms\":81324,\"pass_rate\":0.8978,\"provider_group\":\"provider_z\",\"quality\":0.8756,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.985,\"id\":\"worker_c\",\"p95_latency_ms\":22703,\"pass_rate\":0.5439,\"provider_group\":\"provider_y\",\"quality\":0.6318,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.835,\"estimated_latency_ms\":135090,\"estimated_pass_rate\":0.8905,\"estimated_quality\":0.9527,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":22.158,\"max_latency_ms\":226192,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4208, "category": "analysis", "complexity": 0.6344, "consequence": 0.6544, "evidence_need": 0.7005, "family": "performance_tuning", "input_tokens": 1960, "request_id": "synthetic-performance_tuning-000011", "reversibility": 0.8001, "risk": 0.3731, "summary": "Choose between throughput and tail-latency changes for database failover."}, "workers": [{"credits_per_call": 12.663, "id": "worker_a", "p95_latency_ms": 53766, "pass_rate": 0.9575, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.172, "id": "worker_b", "p95_latency_ms": 81324, "pass_rate": 0.8978, "provider_group": "provider_z", "quality": 0.8756, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.985, "id": "worker_c", "p95_latency_ms": 22703, "pass_rate": 0.5439, "provider_group": "provider_y", "quality": 0.6318, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.527, "max_latency_ms": 127391, "minimum_quality": 0.8985}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001230", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.891, "estimated_latency_ms": 94701, "estimated_pass_rate": 0.6176, "estimated_quality": 0.7576, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.527, "max_latency_ms": 127391, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6176, "quality": 0.7576, "quality_target_met": false, "utility": 52.757653}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.527,\"max_latency_ms\":127391,\"minimum_quality\":0.8985},\"task\":{\"ambiguity\":0.7795,\"category\":\"review\",\"complexity\":0.7901,\"consequence\":0.7627,\"evidence_need\":0.5905,\"family\":\"policy_compliance\",\"input_tokens\":7440,\"reversibility\":0.5035,\"risk\":0.7419,\"summary\":\"Review whether OAuth callbacks satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.921,\"id\":\"worker_a\",\"p95_latency_ms\":73243,\"pass_rate\":0.8049,\"provider_group\":\"provider_x\",\"quality\":0.823,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.97,\"id\":\"worker_b\",\"p95_latency_ms\":21458,\"pass_rate\":0.5668,\"provider_group\":\"provider_z\",\"quality\":0.7283,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.137,\"id\":\"worker_c\",\"p95_latency_ms\":58819,\"pass_rate\":0.9608,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.891,\"estimated_latency_ms\":94701,\"estimated_pass_rate\":0.6176,\"estimated_quality\":0.7576,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.527,\"max_latency_ms\":127391,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7795, "category": "review", "complexity": 0.7901, "consequence": 0.7627, "evidence_need": 0.5905, "family": "policy_compliance", "input_tokens": 7440, "request_id": "synthetic-policy_compliance-001230", "reversibility": 0.5035, "risk": 0.7419, "summary": "Review whether OAuth callbacks satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.921, "id": "worker_a", "p95_latency_ms": 73243, "pass_rate": 0.8049, "provider_group": "provider_x", "quality": 0.823, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.97, "id": "worker_b", "p95_latency_ms": 21458, "pass_rate": 0.5668, "provider_group": "provider_z", "quality": 0.7283, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.137, "id": "worker_c", "p95_latency_ms": 58819, "pass_rate": 0.9608, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.34, "max_latency_ms": 291888, "minimum_quality": 0.76}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000336", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.432, "estimated_latency_ms": 162053, "estimated_pass_rate": 0.854, "estimated_quality": 0.8993, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.34, "max_latency_ms": 291888, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.854, "quality": 0.8993, "quality_target_met": true, "utility": 71.311873}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.34,\"max_latency_ms\":291888,\"minimum_quality\":0.76},\"task\":{\"ambiguity\":0.7422,\"category\":\"review\",\"complexity\":0.5607,\"consequence\":0.7243,\"evidence_need\":0.8271,\"family\":\"policy_compliance\",\"input_tokens\":2320,\"reversibility\":0.3757,\"risk\":0.807,\"summary\":\"Audit a workflow for provider failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.116,\"id\":\"worker_a\",\"p95_latency_ms\":28848,\"pass_rate\":0.5779,\"provider_group\":\"provider_x\",\"quality\":0.7756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.631,\"id\":\"worker_b\",\"p95_latency_ms\":86288,\"pass_rate\":0.9352,\"provider_group\":\"provider_z\",\"quality\":0.9241,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.801,\"id\":\"worker_c\",\"p95_latency_ms\":75765,\"pass_rate\":0.7982,\"provider_group\":\"provider_y\",\"quality\":0.8347,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.432,\"estimated_latency_ms\":162053,\"estimated_pass_rate\":0.854,\"estimated_quality\":0.8993,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.34,\"max_latency_ms\":291888,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7422, "category": "review", "complexity": 0.5607, "consequence": 0.7243, "evidence_need": 0.8271, "family": "policy_compliance", "input_tokens": 2320, "request_id": "synthetic-policy_compliance-000336", "reversibility": 0.3757, "risk": 0.807, "summary": "Audit a workflow for provider failover against explicit access requirements."}, "workers": [{"credits_per_call": 1.116, "id": "worker_a", "p95_latency_ms": 28848, "pass_rate": 0.5779, "provider_group": "provider_x", "quality": 0.7756, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.631, "id": "worker_b", "p95_latency_ms": 86288, "pass_rate": 0.9352, "provider_group": "provider_z", "quality": 0.9241, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.801, "id": "worker_c", "p95_latency_ms": 75765, "pass_rate": 0.7982, "provider_group": "provider_y", "quality": 0.8347, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 13.797, "max_latency_ms": 154991, "minimum_quality": 0.6916}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000111", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.899, "estimated_latency_ms": 84505, "estimated_pass_rate": 0.8052, "estimated_quality": 0.8364, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.797, "max_latency_ms": 154991, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.8052, "quality": 0.8364, "quality_target_met": true, "utility": 65.789299}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":13.797,\"max_latency_ms\":154991,\"minimum_quality\":0.6916},\"task\":{\"ambiguity\":0.5128,\"category\":\"review\",\"complexity\":0.4513,\"consequence\":0.7922,\"evidence_need\":0.5553,\"family\":\"policy_compliance\",\"input_tokens\":4200,\"reversibility\":0.454,\"risk\":0.5056,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.88,\"id\":\"worker_a\",\"p95_latency_ms\":20262,\"pass_rate\":0.606,\"provider_group\":\"provider_x\",\"quality\":0.7375,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.019,\"id\":\"worker_b\",\"p95_latency_ms\":64243,\"pass_rate\":0.8815,\"provider_group\":\"provider_z\",\"quality\":0.8588,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.977,\"id\":\"worker_c\",\"p95_latency_ms\":62212,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9654,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.899,\"estimated_latency_ms\":84505,\"estimated_pass_rate\":0.8052,\"estimated_quality\":0.8364,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.797,\"max_latency_ms\":154991,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5128, "category": "review", "complexity": 0.4513, "consequence": 0.7922, "evidence_need": 0.5553, "family": "policy_compliance", "input_tokens": 4200, "request_id": "synthetic-policy_compliance-000111", "reversibility": 0.454, "risk": 0.5056, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 0.88, "id": "worker_a", "p95_latency_ms": 20262, "pass_rate": 0.606, "provider_group": "provider_x", "quality": 0.7375, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.019, "id": "worker_b", "p95_latency_ms": 64243, "pass_rate": 0.8815, "provider_group": "provider_z", "quality": 0.8588, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.977, "id": "worker_c", "p95_latency_ms": 62212, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9654, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.965, "max_latency_ms": 24271, "minimum_quality": 0.5156}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000876", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.153, "estimated_latency_ms": 19402, "estimated_pass_rate": 0.4029, "estimated_quality": 0.5733, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.965, "max_latency_ms": 24271, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4029, "quality": 0.5733, "quality_target_met": true, "utility": 20.847634}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.965,\"max_latency_ms\":24271,\"minimum_quality\":0.5156},\"task\":{\"ambiguity\":0.5367,\"category\":\"review\",\"complexity\":0.5619,\"consequence\":0.8756,\"evidence_need\":0.8565,\"family\":\"policy_compliance\",\"input_tokens\":6960,\"reversibility\":0.3829,\"risk\":0.7993,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.236,\"id\":\"worker_a\",\"p95_latency_ms\":66315,\"pass_rate\":0.7744,\"provider_group\":\"provider_z\",\"quality\":0.9229,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.153,\"id\":\"worker_b\",\"p95_latency_ms\":19402,\"pass_rate\":0.5719,\"provider_group\":\"provider_x\",\"quality\":0.6885,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.133,\"id\":\"worker_c\",\"p95_latency_ms\":81711,\"pass_rate\":0.9397,\"provider_group\":\"provider_y\",\"quality\":0.9719,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.153,\"estimated_latency_ms\":19402,\"estimated_pass_rate\":0.4029,\"estimated_quality\":0.5733,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.965,\"max_latency_ms\":24271,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5367, "category": "review", "complexity": 0.5619, "consequence": 0.8756, "evidence_need": 0.8565, "family": "policy_compliance", "input_tokens": 6960, "request_id": "synthetic-policy_compliance-000876", "reversibility": 0.3829, "risk": 0.7993, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 4.236, "id": "worker_a", "p95_latency_ms": 66315, "pass_rate": 0.7744, "provider_group": "provider_z", "quality": 0.9229, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.153, "id": "worker_b", "p95_latency_ms": 19402, "pass_rate": 0.5719, "provider_group": "provider_x", "quality": 0.6885, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.133, "id": "worker_c", "p95_latency_ms": 81711, "pass_rate": 0.9397, "provider_group": "provider_y", "quality": 0.9719, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.409, "max_latency_ms": 180258, "minimum_quality": 0.7959}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001056", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.77, "estimated_latency_ms": 168934, "estimated_pass_rate": 0.7964, "estimated_quality": 0.9445, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.409, "max_latency_ms": 180258, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7964, "quality": 0.9445, "quality_target_met": true, "utility": 81.274678}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.409,\"max_latency_ms\":180258,\"minimum_quality\":0.7959},\"task\":{\"ambiguity\":0.4316,\"category\":\"review\",\"complexity\":0.6947,\"consequence\":0.8388,\"evidence_need\":0.9102,\"family\":\"policy_compliance\",\"input_tokens\":3840,\"reversibility\":0.4665,\"risk\":0.7352,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.247,\"id\":\"worker_a\",\"p95_latency_ms\":83671,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.114,\"id\":\"worker_b\",\"p95_latency_ms\":17608,\"pass_rate\":0.5382,\"provider_group\":\"provider_y\",\"quality\":0.7839,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.409,\"id\":\"worker_c\",\"p95_latency_ms\":67655,\"pass_rate\":0.8839,\"provider_group\":\"provider_x\",\"quality\":0.9017,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.77,\"estimated_latency_ms\":168934,\"estimated_pass_rate\":0.7964,\"estimated_quality\":0.9445,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.409,\"max_latency_ms\":180258,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4316, "category": "review", "complexity": 0.6947, "consequence": 0.8388, "evidence_need": 0.9102, "family": "policy_compliance", "input_tokens": 3840, "request_id": "synthetic-policy_compliance-001056", "reversibility": 0.4665, "risk": 0.7352, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.247, "id": "worker_a", "p95_latency_ms": 83671, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.114, "id": "worker_b", "p95_latency_ms": 17608, "pass_rate": 0.5382, "provider_group": "provider_y", "quality": 0.7839, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.409, "id": "worker_c", "p95_latency_ms": 67655, "pass_rate": 0.8839, "provider_group": "provider_x", "quality": 0.9017, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.602, "max_latency_ms": 162015, "minimum_quality": 0.8647}, "family": "migration_planning", "id": "synthetic-migration_planning-001090", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.12, "estimated_latency_ms": 76278, "estimated_pass_rate": 0.6096, "estimated_quality": 0.7955, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.602, "max_latency_ms": 162015, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6096, "quality": 0.7955, "quality_target_met": false, "utility": 54.996072}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.602,\"max_latency_ms\":162015,\"minimum_quality\":0.8647},\"task\":{\"ambiguity\":0.6742,\"category\":\"coding\",\"complexity\":0.757,\"consequence\":0.5365,\"evidence_need\":0.573,\"family\":\"migration_planning\",\"input_tokens\":7400,\"reversibility\":0.3867,\"risk\":0.3676,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":1.01,\"id\":\"worker_a\",\"p95_latency_ms\":19294,\"pass_rate\":0.5533,\"provider_group\":\"provider_z\",\"quality\":0.756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.11,\"id\":\"worker_b\",\"p95_latency_ms\":56984,\"pass_rate\":0.8837,\"provider_group\":\"provider_x\",\"quality\":0.8499,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.792,\"id\":\"worker_c\",\"p95_latency_ms\":72215,\"pass_rate\":0.9747,\"provider_group\":\"provider_y\",\"quality\":0.9169,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.12,\"estimated_latency_ms\":76278,\"estimated_pass_rate\":0.6096,\"estimated_quality\":0.7955,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.602,\"max_latency_ms\":162015,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6742, "category": "coding", "complexity": 0.757, "consequence": 0.5365, "evidence_need": 0.573, "family": "migration_planning", "input_tokens": 7400, "request_id": "synthetic-migration_planning-001090", "reversibility": 0.3867, "risk": 0.3676, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 1.01, "id": "worker_a", "p95_latency_ms": 19294, "pass_rate": 0.5533, "provider_group": "provider_z", "quality": 0.756, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.11, "id": "worker_b", "p95_latency_ms": 56984, "pass_rate": 0.8837, "provider_group": "provider_x", "quality": 0.8499, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.792, "id": "worker_c", "p95_latency_ms": 72215, "pass_rate": 0.9747, "provider_group": "provider_y", "quality": 0.9169, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.917, "max_latency_ms": 151559, "minimum_quality": 0.8835}, "family": "migration_planning", "id": "synthetic-migration_planning-000634", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.623, "estimated_latency_ms": 127407, "estimated_pass_rate": 0.7049, "estimated_quality": 0.8902, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.917, "max_latency_ms": 151559, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7049, "quality": 0.8902, "quality_target_met": true, "utility": 72.812803}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.917,\"max_latency_ms\":151559,\"minimum_quality\":0.8835},\"task\":{\"ambiguity\":0.6565,\"category\":\"coding\",\"complexity\":0.902,\"consequence\":0.4832,\"evidence_need\":0.4731,\"family\":\"migration_planning\",\"input_tokens\":6960,\"reversibility\":0.6266,\"risk\":0.6327,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.98,\"id\":\"worker_a\",\"p95_latency_ms\":25377,\"pass_rate\":0.4796,\"provider_group\":\"provider_x\",\"quality\":0.6769,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.426,\"id\":\"worker_b\",\"p95_latency_ms\":46338,\"pass_rate\":0.8149,\"provider_group\":\"provider_z\",\"quality\":0.8296,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.217,\"id\":\"worker_c\",\"p95_latency_ms\":55692,\"pass_rate\":0.9459,\"provider_group\":\"provider_y\",\"quality\":0.9912,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.623,\"estimated_latency_ms\":127407,\"estimated_pass_rate\":0.7049,\"estimated_quality\":0.8902,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.917,\"max_latency_ms\":151559,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6565, "category": "coding", "complexity": 0.902, "consequence": 0.4832, "evidence_need": 0.4731, "family": "migration_planning", "input_tokens": 6960, "request_id": "synthetic-migration_planning-000634", "reversibility": 0.6266, "risk": 0.6327, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 0.98, "id": "worker_a", "p95_latency_ms": 25377, "pass_rate": 0.4796, "provider_group": "provider_x", "quality": 0.6769, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.426, "id": "worker_b", "p95_latency_ms": 46338, "pass_rate": 0.8149, "provider_group": "provider_z", "quality": 0.8296, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.217, "id": "worker_c", "p95_latency_ms": 55692, "pass_rate": 0.9459, "provider_group": "provider_y", "quality": 0.9912, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.071, "max_latency_ms": 194173, "minimum_quality": 0.8189}, "family": "migration_planning", "id": "synthetic-migration_planning-000226", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.47, "estimated_latency_ms": 140713, "estimated_pass_rate": 0.7021, "estimated_quality": 0.8613, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.071, "max_latency_ms": 194173, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7021, "quality": 0.8613, "quality_target_met": true, "utility": 68.850285}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.071,\"max_latency_ms\":194173,\"minimum_quality\":0.8189},\"task\":{\"ambiguity\":0.6159,\"category\":\"coding\",\"complexity\":0.6387,\"consequence\":0.5226,\"evidence_need\":0.62,\"family\":\"migration_planning\",\"input_tokens\":4800,\"reversibility\":0.611,\"risk\":0.6635,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.992,\"id\":\"worker_a\",\"p95_latency_ms\":69263,\"pass_rate\":0.8069,\"provider_group\":\"provider_z\",\"quality\":0.8033,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.926,\"id\":\"worker_b\",\"p95_latency_ms\":20272,\"pass_rate\":0.48,\"provider_group\":\"provider_x\",\"quality\":0.6894,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.552,\"id\":\"worker_c\",\"p95_latency_ms\":51178,\"pass_rate\":0.9185,\"provider_group\":\"provider_y\",\"quality\":0.9131,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.47,\"estimated_latency_ms\":140713,\"estimated_pass_rate\":0.7021,\"estimated_quality\":0.8613,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.071,\"max_latency_ms\":194173,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6159, "category": "coding", "complexity": 0.6387, "consequence": 0.5226, "evidence_need": 0.62, "family": "migration_planning", "input_tokens": 4800, "request_id": "synthetic-migration_planning-000226", "reversibility": 0.611, "risk": 0.6635, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.992, "id": "worker_a", "p95_latency_ms": 69263, "pass_rate": 0.8069, "provider_group": "provider_z", "quality": 0.8033, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.926, "id": "worker_b", "p95_latency_ms": 20272, "pass_rate": 0.48, "provider_group": "provider_x", "quality": 0.6894, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.552, "id": "worker_c", "p95_latency_ms": 51178, "pass_rate": 0.9185, "provider_group": "provider_y", "quality": 0.9131, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.269, "max_latency_ms": 222941, "minimum_quality": 0.9482}, "family": "migration_planning", "id": "synthetic-migration_planning-000814", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.969, "estimated_latency_ms": 155571, "estimated_pass_rate": 0.6955, "estimated_quality": 0.8919, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 26.269, "max_latency_ms": 222941, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6955, "quality": 0.8919, "quality_target_met": false, "utility": 70.549651}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.269,\"max_latency_ms\":222941,\"minimum_quality\":0.9482},\"task\":{\"ambiguity\":0.5396,\"category\":\"coding\",\"complexity\":0.8904,\"consequence\":0.4067,\"evidence_need\":0.4659,\"family\":\"migration_planning\",\"input_tokens\":1120,\"reversibility\":0.6293,\"risk\":0.6682,\"summary\":\"Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.764,\"id\":\"worker_a\",\"p95_latency_ms\":77022,\"pass_rate\":0.7319,\"provider_group\":\"provider_y\",\"quality\":0.8339,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.131,\"id\":\"worker_b\",\"p95_latency_ms\":21630,\"pass_rate\":0.4843,\"provider_group\":\"provider_z\",\"quality\":0.7252,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.074,\"id\":\"worker_c\",\"p95_latency_ms\":56919,\"pass_rate\":0.9449,\"provider_group\":\"provider_x\",\"quality\":0.9893,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.969,\"estimated_latency_ms\":155571,\"estimated_pass_rate\":0.6955,\"estimated_quality\":0.8919,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.269,\"max_latency_ms\":222941,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5396, "category": "coding", "complexity": 0.8904, "consequence": 0.4067, "evidence_need": 0.4659, "family": "migration_planning", "input_tokens": 1120, "request_id": "synthetic-migration_planning-000814", "reversibility": 0.6293, "risk": 0.6682, "summary": "Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints."}, "workers": [{"credits_per_call": 3.764, "id": "worker_a", "p95_latency_ms": 77022, "pass_rate": 0.7319, "provider_group": "provider_y", "quality": 0.8339, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.131, "id": "worker_b", "p95_latency_ms": 21630, "pass_rate": 0.4843, "provider_group": "provider_z", "quality": 0.7252, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.074, "id": "worker_c", "p95_latency_ms": 56919, "pass_rate": 0.9449, "provider_group": "provider_x", "quality": 0.9893, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.039, "max_latency_ms": 181451, "minimum_quality": 0.8943}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001212", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.332, "estimated_latency_ms": 160020, "estimated_pass_rate": 0.7658, "estimated_quality": 0.921, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.039, "max_latency_ms": 181451, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7658, "quality": 0.921, "quality_target_met": true, "utility": 78.642506}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.039,\"max_latency_ms\":181451,\"minimum_quality\":0.8943},\"task\":{\"ambiguity\":0.614,\"category\":\"review\",\"complexity\":0.7824,\"consequence\":0.8466,\"evidence_need\":0.8941,\"family\":\"policy_compliance\",\"input_tokens\":2360,\"reversibility\":0.5036,\"risk\":0.5986,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.807,\"id\":\"worker_a\",\"p95_latency_ms\":28674,\"pass_rate\":0.5956,\"provider_group\":\"provider_x\",\"quality\":0.7433,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.825,\"id\":\"worker_b\",\"p95_latency_ms\":61881,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.7,\"id\":\"worker_c\",\"p95_latency_ms\":69465,\"pass_rate\":0.7261,\"provider_group\":\"provider_y\",\"quality\":0.8898,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.332,\"estimated_latency_ms\":160020,\"estimated_pass_rate\":0.7658,\"estimated_quality\":0.921,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.039,\"max_latency_ms\":181451,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.614, "category": "review", "complexity": 0.7824, "consequence": 0.8466, "evidence_need": 0.8941, "family": "policy_compliance", "input_tokens": 2360, "request_id": "synthetic-policy_compliance-001212", "reversibility": 0.5036, "risk": 0.5986, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 0.807, "id": "worker_a", "p95_latency_ms": 28674, "pass_rate": 0.5956, "provider_group": "provider_x", "quality": 0.7433, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.825, "id": "worker_b", "p95_latency_ms": 61881, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.7, "id": "worker_c", "p95_latency_ms": 69465, "pass_rate": 0.7261, "provider_group": "provider_y", "quality": 0.8898, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.474, "max_latency_ms": 38502, "minimum_quality": 0.6574}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000033", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.946, "estimated_latency_ms": 19686, "estimated_pass_rate": 0.4355, "estimated_quality": 0.6704, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.474, "max_latency_ms": 38502, "policy": "lean", "reason_codes": ["high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4355, "quality": 0.6704, "quality_target_met": true, "utility": 46.603821}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.474,\"max_latency_ms\":38502,\"minimum_quality\":0.6574},\"task\":{\"ambiguity\":0.5333,\"category\":\"review\",\"complexity\":0.6503,\"consequence\":0.8869,\"evidence_need\":0.5599,\"family\":\"policy_compliance\",\"input_tokens\":5920,\"reversibility\":0.3993,\"risk\":0.5761,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":0.946,\"id\":\"worker_a\",\"p95_latency_ms\":19686,\"pass_rate\":0.5962,\"provider_group\":\"provider_x\",\"quality\":0.7853,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.73,\"id\":\"worker_b\",\"p95_latency_ms\":64066,\"pass_rate\":0.7265,\"provider_group\":\"provider_z\",\"quality\":0.9232,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.225,\"id\":\"worker_c\",\"p95_latency_ms\":90175,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9867,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.946,\"estimated_latency_ms\":19686,\"estimated_pass_rate\":0.4355,\"estimated_quality\":0.6704,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.474,\"max_latency_ms\":38502,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5333, "category": "review", "complexity": 0.6503, "consequence": 0.8869, "evidence_need": 0.5599, "family": "policy_compliance", "input_tokens": 5920, "request_id": "synthetic-policy_compliance-000033", "reversibility": 0.3993, "risk": 0.5761, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 0.946, "id": "worker_a", "p95_latency_ms": 19686, "pass_rate": 0.5962, "provider_group": "provider_x", "quality": 0.7853, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.73, "id": "worker_b", "p95_latency_ms": 64066, "pass_rate": 0.7265, "provider_group": "provider_z", "quality": 0.9232, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.225, "id": "worker_c", "p95_latency_ms": 90175, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9867, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.469, "max_latency_ms": 35197, "minimum_quality": 0.6244}, "family": "migration_planning", "id": "synthetic-migration_planning-000652", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.036, "estimated_latency_ms": 28027, "estimated_pass_rate": 0.3964, "estimated_quality": 0.6227, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.469, "max_latency_ms": 35197, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3964, "quality": 0.6227, "quality_target_met": false, "utility": 41.132676}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.469,\"max_latency_ms\":35197,\"minimum_quality\":0.6244},\"task\":{\"ambiguity\":0.3859,\"category\":\"coding\",\"complexity\":0.6337,\"consequence\":0.6955,\"evidence_need\":0.6255,\"family\":\"migration_planning\",\"input_tokens\":160,\"reversibility\":0.4737,\"risk\":0.5194,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":13.134,\"id\":\"worker_a\",\"p95_latency_ms\":50613,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.473,\"id\":\"worker_b\",\"p95_latency_ms\":61385,\"pass_rate\":0.8022,\"provider_group\":\"provider_x\",\"quality\":0.8097,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.036,\"id\":\"worker_c\",\"p95_latency_ms\":28027,\"pass_rate\":0.5395,\"provider_group\":\"provider_z\",\"quality\":0.7206,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.036,\"estimated_latency_ms\":28027,\"estimated_pass_rate\":0.3964,\"estimated_quality\":0.6227,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.469,\"max_latency_ms\":35197,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3859, "category": "coding", "complexity": 0.6337, "consequence": 0.6955, "evidence_need": 0.6255, "family": "migration_planning", "input_tokens": 160, "request_id": "synthetic-migration_planning-000652", "reversibility": 0.4737, "risk": 0.5194, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 13.134, "id": "worker_a", "p95_latency_ms": 50613, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.473, "id": "worker_b", "p95_latency_ms": 61385, "pass_rate": 0.8022, "provider_group": "provider_x", "quality": 0.8097, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.036, "id": "worker_c", "p95_latency_ms": 28027, "pass_rate": 0.5395, "provider_group": "provider_z", "quality": 0.7206, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.35, "max_latency_ms": 48017, "minimum_quality": 0.7266}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000477", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.082, "estimated_latency_ms": 31573, "estimated_pass_rate": 0.3109, "estimated_quality": 0.6905, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.35, "max_latency_ms": 48017, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3109, "quality": 0.6905, "quality_target_met": false, "utility": 43.225684}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.35,\"max_latency_ms\":48017,\"minimum_quality\":0.7266},\"task\":{\"ambiguity\":0.364,\"category\":\"review\",\"complexity\":0.691,\"consequence\":0.528,\"evidence_need\":0.7003,\"family\":\"policy_compliance\",\"input_tokens\":5920,\"reversibility\":0.6237,\"risk\":0.6197,\"summary\":\"Identify evidence required before approving artifact storage for production.\"},\"workers\":[{\"credits_per_call\":9.951,\"id\":\"worker_a\",\"p95_latency_ms\":89207,\"pass_rate\":0.9739,\"provider_group\":\"provider_y\",\"quality\":0.9804,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.183,\"id\":\"worker_b\",\"p95_latency_ms\":69998,\"pass_rate\":0.8143,\"provider_group\":\"provider_x\",\"quality\":0.9218,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.082,\"id\":\"worker_c\",\"p95_latency_ms\":31573,\"pass_rate\":0.467,\"provider_group\":\"provider_z\",\"quality\":0.8045,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.082,\"estimated_latency_ms\":31573,\"estimated_pass_rate\":0.3109,\"estimated_quality\":0.6905,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.35,\"max_latency_ms\":48017,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.364, "category": "review", "complexity": 0.691, "consequence": 0.528, "evidence_need": 0.7003, "family": "policy_compliance", "input_tokens": 5920, "request_id": "synthetic-policy_compliance-000477", "reversibility": 0.6237, "risk": 0.6197, "summary": "Identify evidence required before approving artifact storage for production."}, "workers": [{"credits_per_call": 9.951, "id": "worker_a", "p95_latency_ms": 89207, "pass_rate": 0.9739, "provider_group": "provider_y", "quality": 0.9804, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.183, "id": "worker_b", "p95_latency_ms": 69998, "pass_rate": 0.8143, "provider_group": "provider_x", "quality": 0.9218, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.082, "id": "worker_c", "p95_latency_ms": 31573, "pass_rate": 0.467, "provider_group": "provider_z", "quality": 0.8045, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.743, "max_latency_ms": 118120, "minimum_quality": 0.7006}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000453", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.914, "estimated_latency_ms": 101599, "estimated_pass_rate": 0.6215, "estimated_quality": 0.8576, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.743, "max_latency_ms": 118120, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6215, "quality": 0.8576, "quality_target_met": true, "utility": 63.362894}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.743,\"max_latency_ms\":118120,\"minimum_quality\":0.7006},\"task\":{\"ambiguity\":0.6232,\"category\":\"review\",\"complexity\":0.5781,\"consequence\":0.7355,\"evidence_need\":0.9351,\"family\":\"policy_compliance\",\"input_tokens\":4360,\"reversibility\":0.42,\"risk\":0.5137,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.844,\"id\":\"worker_a\",\"p95_latency_ms\":73488,\"pass_rate\":0.8665,\"provider_group\":\"provider_x\",\"quality\":0.9172,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.07,\"id\":\"worker_b\",\"p95_latency_ms\":28111,\"pass_rate\":0.4966,\"provider_group\":\"provider_z\",\"quality\":0.7364,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.171,\"id\":\"worker_c\",\"p95_latency_ms\":50554,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9532,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.914,\"estimated_latency_ms\":101599,\"estimated_pass_rate\":0.6215,\"estimated_quality\":0.8576,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.743,\"max_latency_ms\":118120,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6232, "category": "review", "complexity": 0.5781, "consequence": 0.7355, "evidence_need": 0.9351, "family": "policy_compliance", "input_tokens": 4360, "request_id": "synthetic-policy_compliance-000453", "reversibility": 0.42, "risk": 0.5137, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.844, "id": "worker_a", "p95_latency_ms": 73488, "pass_rate": 0.8665, "provider_group": "provider_x", "quality": 0.9172, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.07, "id": "worker_b", "p95_latency_ms": 28111, "pass_rate": 0.4966, "provider_group": "provider_z", "quality": 0.7364, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.171, "id": "worker_c", "p95_latency_ms": 50554, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9532, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.777, "max_latency_ms": 236478, "minimum_quality": 0.8412}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000576", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.754, "estimated_latency_ms": 175660, "estimated_pass_rate": 0.6983, "estimated_quality": 0.8799, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.777, "max_latency_ms": 236478, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6983, "quality": 0.8799, "quality_target_met": true, "utility": 74.435202}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.777,\"max_latency_ms\":236478,\"minimum_quality\":0.8412},\"task\":{\"ambiguity\":0.6272,\"category\":\"review\",\"complexity\":0.5703,\"consequence\":0.8612,\"evidence_need\":0.8903,\"family\":\"policy_compliance\",\"input_tokens\":1280,\"reversibility\":0.305,\"risk\":0.6632,\"summary\":\"Review whether usage metering satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.824,\"id\":\"worker_a\",\"p95_latency_ms\":61412,\"pass_rate\":0.7364,\"provider_group\":\"provider_z\",\"quality\":0.8704,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.821,\"id\":\"worker_b\",\"p95_latency_ms\":86782,\"pass_rate\":0.929,\"provider_group\":\"provider_x\",\"quality\":0.922,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.109,\"id\":\"worker_c\",\"p95_latency_ms\":27466,\"pass_rate\":0.4575,\"provider_group\":\"provider_y\",\"quality\":0.7016,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.754,\"estimated_latency_ms\":175660,\"estimated_pass_rate\":0.6983,\"estimated_quality\":0.8799,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.777,\"max_latency_ms\":236478,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6272, "category": "review", "complexity": 0.5703, "consequence": 0.8612, "evidence_need": 0.8903, "family": "policy_compliance", "input_tokens": 1280, "request_id": "synthetic-policy_compliance-000576", "reversibility": 0.305, "risk": 0.6632, "summary": "Review whether usage metering satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.824, "id": "worker_a", "p95_latency_ms": 61412, "pass_rate": 0.7364, "provider_group": "provider_z", "quality": 0.8704, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.821, "id": "worker_b", "p95_latency_ms": 86782, "pass_rate": 0.929, "provider_group": "provider_x", "quality": 0.922, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.109, "id": "worker_c", "p95_latency_ms": 27466, "pass_rate": 0.4575, "provider_group": "provider_y", "quality": 0.7016, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.436, "max_latency_ms": 34644, "minimum_quality": 0.6658}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000362", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.998, "estimated_latency_ms": 21386, "estimated_pass_rate": 0.3627, "estimated_quality": 0.5463, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.436, "max_latency_ms": 34644, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3627, "quality": 0.5463, "quality_target_met": false, "utility": 29.550914}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.436,\"max_latency_ms\":34644,\"minimum_quality\":0.6658},\"task\":{\"ambiguity\":0.5715,\"category\":\"analysis\",\"complexity\":0.6333,\"consequence\":0.5466,\"evidence_need\":0.5324,\"family\":\"performance_tuning\",\"input_tokens\":4880,\"reversibility\":0.897,\"risk\":0.4571,\"summary\":\"Review a proposed optimization for deployment rollback for hidden regressions.\"},\"workers\":[{\"credits_per_call\":13.138,\"id\":\"worker_a\",\"p95_latency_ms\":78834,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.978,\"id\":\"worker_b\",\"p95_latency_ms\":66292,\"pass_rate\":0.93,\"provider_group\":\"provider_y\",\"quality\":0.8887,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.998,\"id\":\"worker_c\",\"p95_latency_ms\":21386,\"pass_rate\":0.5147,\"provider_group\":\"provider_z\",\"quality\":0.6521,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.998,\"estimated_latency_ms\":21386,\"estimated_pass_rate\":0.3627,\"estimated_quality\":0.5463,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.436,\"max_latency_ms\":34644,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5715, "category": "analysis", "complexity": 0.6333, "consequence": 0.5466, "evidence_need": 0.5324, "family": "performance_tuning", "input_tokens": 4880, "request_id": "synthetic-performance_tuning-000362", "reversibility": 0.897, "risk": 0.4571, "summary": "Review a proposed optimization for deployment rollback for hidden regressions."}, "workers": [{"credits_per_call": 13.138, "id": "worker_a", "p95_latency_ms": 78834, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.978, "id": "worker_b", "p95_latency_ms": 66292, "pass_rate": 0.93, "provider_group": "provider_y", "quality": 0.8887, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.998, "id": "worker_c", "p95_latency_ms": 21386, "pass_rate": 0.5147, "provider_group": "provider_z", "quality": 0.6521, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.238, "max_latency_ms": 26692, "minimum_quality": 0.5782}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000846", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.961, "estimated_latency_ms": 17948, "estimated_pass_rate": 0.3066, "estimated_quality": 0.6462, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.238, "max_latency_ms": 26692, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3066, "quality": 0.6462, "quality_target_met": true, "utility": 41.084541}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.238,\"max_latency_ms\":26692,\"minimum_quality\":0.5782},\"task\":{\"ambiguity\":0.3561,\"category\":\"review\",\"complexity\":0.738,\"consequence\":0.62,\"evidence_need\":0.832,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.4814,\"risk\":0.7325,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.945,\"id\":\"worker_a\",\"p95_latency_ms\":53251,\"pass_rate\":0.813,\"provider_group\":\"provider_y\",\"quality\":0.8178,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.579,\"id\":\"worker_b\",\"p95_latency_ms\":63545,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9353,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.961,\"id\":\"worker_c\",\"p95_latency_ms\":17948,\"pass_rate\":0.4765,\"provider_group\":\"provider_x\",\"quality\":0.7763,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.961,\"estimated_latency_ms\":17948,\"estimated_pass_rate\":0.3066,\"estimated_quality\":0.6462,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.238,\"max_latency_ms\":26692,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3561, "category": "review", "complexity": 0.738, "consequence": 0.62, "evidence_need": 0.832, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-000846", "reversibility": 0.4814, "risk": 0.7325, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.945, "id": "worker_a", "p95_latency_ms": 53251, "pass_rate": 0.813, "provider_group": "provider_y", "quality": 0.8178, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.579, "id": "worker_b", "p95_latency_ms": 63545, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9353, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.961, "id": "worker_c", "p95_latency_ms": 17948, "pass_rate": 0.4765, "provider_group": "provider_x", "quality": 0.7763, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.402, "max_latency_ms": 226220, "minimum_quality": 0.8901}, "family": "migration_planning", "id": "synthetic-migration_planning-000841", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.398, "estimated_latency_ms": 169710, "estimated_pass_rate": 0.7731, "estimated_quality": 0.9016, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.402, "max_latency_ms": 226220, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7731, "quality": 0.9016, "quality_target_met": true, "utility": 71.213676}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.402,\"max_latency_ms\":226220,\"minimum_quality\":0.8901},\"task\":{\"ambiguity\":0.3968,\"category\":\"coding\",\"complexity\":0.6715,\"consequence\":0.4029,\"evidence_need\":0.4321,\"family\":\"migration_planning\",\"input_tokens\":1680,\"reversibility\":0.2751,\"risk\":0.7059,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.419,\"id\":\"worker_a\",\"p95_latency_ms\":67083,\"pass_rate\":0.9735,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.148,\"id\":\"worker_b\",\"p95_latency_ms\":31983,\"pass_rate\":0.57,\"provider_group\":\"provider_y\",\"quality\":0.6933,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.831,\"id\":\"worker_c\",\"p95_latency_ms\":70644,\"pass_rate\":0.8269,\"provider_group\":\"provider_z\",\"quality\":0.8297,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.398,\"estimated_latency_ms\":169710,\"estimated_pass_rate\":0.7731,\"estimated_quality\":0.9016,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.402,\"max_latency_ms\":226220,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3968, "category": "coding", "complexity": 0.6715, "consequence": 0.4029, "evidence_need": 0.4321, "family": "migration_planning", "input_tokens": 1680, "request_id": "synthetic-migration_planning-000841", "reversibility": 0.2751, "risk": 0.7059, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 14.419, "id": "worker_a", "p95_latency_ms": 67083, "pass_rate": 0.9735, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.148, "id": "worker_b", "p95_latency_ms": 31983, "pass_rate": 0.57, "provider_group": "provider_y", "quality": 0.6933, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.831, "id": "worker_c", "p95_latency_ms": 70644, "pass_rate": 0.8269, "provider_group": "provider_z", "quality": 0.8297, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.379, "max_latency_ms": 28078, "minimum_quality": 0.5159}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000645", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.934, "estimated_latency_ms": 20055, "estimated_pass_rate": 0.3658, "estimated_quality": 0.5798, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.379, "max_latency_ms": 28078, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3658, "quality": 0.5798, "quality_target_met": true, "utility": 38.398615}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.379,\"max_latency_ms\":28078,\"minimum_quality\":0.5159},\"task\":{\"ambiguity\":0.4846,\"category\":\"review\",\"complexity\":0.5954,\"consequence\":0.7282,\"evidence_need\":0.9479,\"family\":\"policy_compliance\",\"input_tokens\":4360,\"reversibility\":0.3655,\"risk\":0.5855,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":0.934,\"id\":\"worker_a\",\"p95_latency_ms\":20055,\"pass_rate\":0.5174,\"provider_group\":\"provider_z\",\"quality\":0.6822,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.242,\"id\":\"worker_b\",\"p95_latency_ms\":76263,\"pass_rate\":0.8253,\"provider_group\":\"provider_x\",\"quality\":0.8397,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.299,\"id\":\"worker_c\",\"p95_latency_ms\":56039,\"pass_rate\":0.9687,\"provider_group\":\"provider_y\",\"quality\":0.956,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.934,\"estimated_latency_ms\":20055,\"estimated_pass_rate\":0.3658,\"estimated_quality\":0.5798,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.379,\"max_latency_ms\":28078,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4846, "category": "review", "complexity": 0.5954, "consequence": 0.7282, "evidence_need": 0.9479, "family": "policy_compliance", "input_tokens": 4360, "request_id": "synthetic-policy_compliance-000645", "reversibility": 0.3655, "risk": 0.5855, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 0.934, "id": "worker_a", "p95_latency_ms": 20055, "pass_rate": 0.5174, "provider_group": "provider_z", "quality": 0.6822, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.242, "id": "worker_b", "p95_latency_ms": 76263, "pass_rate": 0.8253, "provider_group": "provider_x", "quality": 0.8397, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.299, "id": "worker_c", "p95_latency_ms": 56039, "pass_rate": 0.9687, "provider_group": "provider_y", "quality": 0.956, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.328, "max_latency_ms": 181523, "minimum_quality": 0.915}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000600", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.866, "estimated_latency_ms": 143697, "estimated_pass_rate": 0.7593, "estimated_quality": 0.9193, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.328, "max_latency_ms": 181523, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7593, "quality": 0.9193, "quality_target_met": true, "utility": 72.405539}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.328,\"max_latency_ms\":181523,\"minimum_quality\":0.915},\"task\":{\"ambiguity\":0.7361,\"category\":\"review\",\"complexity\":0.7211,\"consequence\":0.7248,\"evidence_need\":0.8621,\"family\":\"policy_compliance\",\"input_tokens\":3120,\"reversibility\":0.4072,\"risk\":0.6165,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.989,\"id\":\"worker_a\",\"p95_latency_ms\":61674,\"pass_rate\":0.9098,\"provider_group\":\"provider_x\",\"quality\":0.9933,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.837,\"id\":\"worker_b\",\"p95_latency_ms\":24273,\"pass_rate\":0.5549,\"provider_group\":\"provider_z\",\"quality\":0.6943,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.04,\"id\":\"worker_c\",\"p95_latency_ms\":57750,\"pass_rate\":0.7678,\"provider_group\":\"provider_y\",\"quality\":0.8927,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.866,\"estimated_latency_ms\":143697,\"estimated_pass_rate\":0.7593,\"estimated_quality\":0.9193,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.328,\"max_latency_ms\":181523,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7361, "category": "review", "complexity": 0.7211, "consequence": 0.7248, "evidence_need": 0.8621, "family": "policy_compliance", "input_tokens": 3120, "request_id": "synthetic-policy_compliance-000600", "reversibility": 0.4072, "risk": 0.6165, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 12.989, "id": "worker_a", "p95_latency_ms": 61674, "pass_rate": 0.9098, "provider_group": "provider_x", "quality": 0.9933, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.837, "id": "worker_b", "p95_latency_ms": 24273, "pass_rate": 0.5549, "provider_group": "provider_z", "quality": 0.6943, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.04, "id": "worker_c", "p95_latency_ms": 57750, "pass_rate": 0.7678, "provider_group": "provider_y", "quality": 0.8927, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.639, "max_latency_ms": 37594, "minimum_quality": 0.7583}, "family": "migration_planning", "id": "synthetic-migration_planning-000520", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.937, "estimated_latency_ms": 19011, "estimated_pass_rate": 0.3511, "estimated_quality": 0.6296, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.639, "max_latency_ms": 37594, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3511, "quality": 0.6296, "quality_target_met": false, "utility": 36.944073}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.639,\"max_latency_ms\":37594,\"minimum_quality\":0.7583},\"task\":{\"ambiguity\":0.4507,\"category\":\"coding\",\"complexity\":0.9041,\"consequence\":0.8162,\"evidence_need\":0.4896,\"family\":\"migration_planning\",\"input_tokens\":1680,\"reversibility\":0.4152,\"risk\":0.3858,\"summary\":\"Design compatibility testing for a observability schema migration.\"},\"workers\":[{\"credits_per_call\":3.817,\"id\":\"worker_a\",\"p95_latency_ms\":48548,\"pass_rate\":0.8607,\"provider_group\":\"provider_z\",\"quality\":0.86,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.43,\"id\":\"worker_b\",\"p95_latency_ms\":62600,\"pass_rate\":0.9069,\"provider_group\":\"provider_x\",\"quality\":0.9915,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.937,\"id\":\"worker_c\",\"p95_latency_ms\":19011,\"pass_rate\":0.5189,\"provider_group\":\"provider_y\",\"quality\":0.7711,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.937,\"estimated_latency_ms\":19011,\"estimated_pass_rate\":0.3511,\"estimated_quality\":0.6296,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.639,\"max_latency_ms\":37594,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4507, "category": "coding", "complexity": 0.9041, "consequence": 0.8162, "evidence_need": 0.4896, "family": "migration_planning", "input_tokens": 1680, "request_id": "synthetic-migration_planning-000520", "reversibility": 0.4152, "risk": 0.3858, "summary": "Design compatibility testing for a observability schema migration."}, "workers": [{"credits_per_call": 3.817, "id": "worker_a", "p95_latency_ms": 48548, "pass_rate": 0.8607, "provider_group": "provider_z", "quality": 0.86, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.43, "id": "worker_b", "p95_latency_ms": 62600, "pass_rate": 0.9069, "provider_group": "provider_x", "quality": 0.9915, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.937, "id": "worker_c", "p95_latency_ms": 19011, "pass_rate": 0.5189, "provider_group": "provider_y", "quality": 0.7711, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.601, "max_latency_ms": 189811, "minimum_quality": 0.9596}, "family": "migration_planning", "id": "synthetic-migration_planning-001228", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.937, "estimated_latency_ms": 145420, "estimated_pass_rate": 0.7081, "estimated_quality": 0.9085, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 15.601, "max_latency_ms": 189811, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7081, "quality": 0.9085, "quality_target_met": false, "utility": 70.82594}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.601,\"max_latency_ms\":189811,\"minimum_quality\":0.9596},\"task\":{\"ambiguity\":0.5192,\"category\":\"coding\",\"complexity\":0.9243,\"consequence\":0.5473,\"evidence_need\":0.6571,\"family\":\"migration_planning\",\"input_tokens\":3120,\"reversibility\":0.5713,\"risk\":0.525,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.063,\"id\":\"worker_a\",\"p95_latency_ms\":26506,\"pass_rate\":0.5071,\"provider_group\":\"provider_y\",\"quality\":0.7147,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.049,\"id\":\"worker_b\",\"p95_latency_ms\":65021,\"pass_rate\":0.9113,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.825,\"id\":\"worker_c\",\"p95_latency_ms\":53893,\"pass_rate\":0.8486,\"provider_group\":\"provider_z\",\"quality\":0.8485,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.937,\"estimated_latency_ms\":145420,\"estimated_pass_rate\":0.7081,\"estimated_quality\":0.9085,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":15.601,\"max_latency_ms\":189811,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5192, "category": "coding", "complexity": 0.9243, "consequence": 0.5473, "evidence_need": 0.6571, "family": "migration_planning", "input_tokens": 3120, "request_id": "synthetic-migration_planning-001228", "reversibility": 0.5713, "risk": 0.525, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 1.063, "id": "worker_a", "p95_latency_ms": 26506, "pass_rate": 0.5071, "provider_group": "provider_y", "quality": 0.7147, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.049, "id": "worker_b", "p95_latency_ms": 65021, "pass_rate": 0.9113, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.825, "id": "worker_c", "p95_latency_ms": 53893, "pass_rate": 0.8486, "provider_group": "provider_z", "quality": 0.8485, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.16, "max_latency_ms": 39938, "minimum_quality": 0.5159}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000546", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.099, "estimated_latency_ms": 21610, "estimated_pass_rate": 0.4084, "estimated_quality": 0.7305, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.16, "max_latency_ms": 39938, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4084, "quality": 0.7305, "quality_target_met": true, "utility": 48.687727}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.16,\"max_latency_ms\":39938,\"minimum_quality\":0.5159},\"task\":{\"ambiguity\":0.6118,\"category\":\"review\",\"complexity\":0.4028,\"consequence\":0.5929,\"evidence_need\":0.6319,\"family\":\"policy_compliance\",\"input_tokens\":5280,\"reversibility\":0.6891,\"risk\":0.5193,\"summary\":\"Review whether tenant-scoped caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.099,\"id\":\"worker_a\",\"p95_latency_ms\":21610,\"pass_rate\":0.5424,\"provider_group\":\"provider_x\",\"quality\":0.8018,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.777,\"id\":\"worker_b\",\"p95_latency_ms\":74488,\"pass_rate\":0.7471,\"provider_group\":\"provider_z\",\"quality\":0.919,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.341,\"id\":\"worker_c\",\"p95_latency_ms\":69879,\"pass_rate\":0.9428,\"provider_group\":\"provider_y\",\"quality\":0.9919,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.099,\"estimated_latency_ms\":21610,\"estimated_pass_rate\":0.4084,\"estimated_quality\":0.7305,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.16,\"max_latency_ms\":39938,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6118, "category": "review", "complexity": 0.4028, "consequence": 0.5929, "evidence_need": 0.6319, "family": "policy_compliance", "input_tokens": 5280, "request_id": "synthetic-policy_compliance-000546", "reversibility": 0.6891, "risk": 0.5193, "summary": "Review whether tenant-scoped caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.099, "id": "worker_a", "p95_latency_ms": 21610, "pass_rate": 0.5424, "provider_group": "provider_x", "quality": 0.8018, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.777, "id": "worker_b", "p95_latency_ms": 74488, "pass_rate": 0.7471, "provider_group": "provider_z", "quality": 0.919, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.341, "id": "worker_c", "p95_latency_ms": 69879, "pass_rate": 0.9428, "provider_group": "provider_y", "quality": 0.9919, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.131, "max_latency_ms": 194184, "minimum_quality": 0.8653}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001206", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.862, "estimated_latency_ms": 120163, "estimated_pass_rate": 0.8816, "estimated_quality": 0.9473, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.131, "max_latency_ms": 194184, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8816, "quality": 0.9473, "quality_target_met": true, "utility": 74.825492}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.131,\"max_latency_ms\":194184,\"minimum_quality\":0.8653},\"task\":{\"ambiguity\":0.4908,\"category\":\"review\",\"complexity\":0.4416,\"consequence\":0.7349,\"evidence_need\":0.772,\"family\":\"policy_compliance\",\"input_tokens\":3280,\"reversibility\":0.6423,\"risk\":0.7646,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.841,\"id\":\"worker_a\",\"p95_latency_ms\":69500,\"pass_rate\":0.8374,\"provider_group\":\"provider_z\",\"quality\":0.8604,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.021,\"id\":\"worker_b\",\"p95_latency_ms\":50663,\"pass_rate\":0.9371,\"provider_group\":\"provider_x\",\"quality\":0.9765,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.191,\"id\":\"worker_c\",\"p95_latency_ms\":21057,\"pass_rate\":0.543,\"provider_group\":\"provider_y\",\"quality\":0.701,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.862,\"estimated_latency_ms\":120163,\"estimated_pass_rate\":0.8816,\"estimated_quality\":0.9473,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.131,\"max_latency_ms\":194184,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4908, "category": "review", "complexity": 0.4416, "consequence": 0.7349, "evidence_need": 0.772, "family": "policy_compliance", "input_tokens": 3280, "request_id": "synthetic-policy_compliance-001206", "reversibility": 0.6423, "risk": 0.7646, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 2.841, "id": "worker_a", "p95_latency_ms": 69500, "pass_rate": 0.8374, "provider_group": "provider_z", "quality": 0.8604, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.021, "id": "worker_b", "p95_latency_ms": 50663, "pass_rate": 0.9371, "provider_group": "provider_x", "quality": 0.9765, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.191, "id": "worker_c", "p95_latency_ms": 21057, "pass_rate": 0.543, "provider_group": "provider_y", "quality": 0.701, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 22.186, "max_latency_ms": 170759, "minimum_quality": 0.8457}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000462", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.724, "estimated_latency_ms": 112398, "estimated_pass_rate": 0.8515, "estimated_quality": 0.913, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 22.186, "max_latency_ms": 170759, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8515, "quality": 0.913, "quality_target_met": true, "utility": 70.999416}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":22.186,\"max_latency_ms\":170759,\"minimum_quality\":0.8457},\"task\":{\"ambiguity\":0.3606,\"category\":\"review\",\"complexity\":0.8426,\"consequence\":0.5954,\"evidence_need\":0.7049,\"family\":\"policy_compliance\",\"input_tokens\":6160,\"reversibility\":0.5752,\"risk\":0.8358,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":1.08,\"id\":\"worker_a\",\"p95_latency_ms\":21786,\"pass_rate\":0.573,\"provider_group\":\"provider_z\",\"quality\":0.7444,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.154,\"id\":\"worker_b\",\"p95_latency_ms\":61337,\"pass_rate\":0.8165,\"provider_group\":\"provider_y\",\"quality\":0.8611,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.57,\"id\":\"worker_c\",\"p95_latency_ms\":51061,\"pass_rate\":0.921,\"provider_group\":\"provider_x\",\"quality\":0.9774,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.724,\"estimated_latency_ms\":112398,\"estimated_pass_rate\":0.8515,\"estimated_quality\":0.913,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":22.186,\"max_latency_ms\":170759,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3606, "category": "review", "complexity": 0.8426, "consequence": 0.5954, "evidence_need": 0.7049, "family": "policy_compliance", "input_tokens": 6160, "request_id": "synthetic-policy_compliance-000462", "reversibility": 0.5752, "risk": 0.8358, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 1.08, "id": "worker_a", "p95_latency_ms": 21786, "pass_rate": 0.573, "provider_group": "provider_z", "quality": 0.7444, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.154, "id": "worker_b", "p95_latency_ms": 61337, "pass_rate": 0.8165, "provider_group": "provider_y", "quality": 0.8611, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.57, "id": "worker_c", "p95_latency_ms": 51061, "pass_rate": 0.921, "provider_group": "provider_x", "quality": 0.9774, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.583, "max_latency_ms": 28982, "minimum_quality": 0.7365}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000017", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.013, "estimated_latency_ms": 17881, "estimated_pass_rate": 0.3458, "estimated_quality": 0.5335, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.583, "max_latency_ms": 28982, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3458, "quality": 0.5335, "quality_target_met": false, "utility": 25.248569}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.583,\"max_latency_ms\":28982,\"minimum_quality\":0.7365},\"task\":{\"ambiguity\":0.7139,\"category\":\"analysis\",\"complexity\":0.804,\"consequence\":0.338,\"evidence_need\":0.4539,\"family\":\"performance_tuning\",\"input_tokens\":1480,\"reversibility\":0.5821,\"risk\":0.252,\"summary\":\"Review a proposed optimization for schema evolution for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.634,\"id\":\"worker_a\",\"p95_latency_ms\":71535,\"pass_rate\":0.8252,\"provider_group\":\"provider_y\",\"quality\":0.9563,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.873,\"id\":\"worker_b\",\"p95_latency_ms\":74798,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.013,\"id\":\"worker_c\",\"p95_latency_ms\":17881,\"pass_rate\":0.5114,\"provider_group\":\"provider_x\",\"quality\":0.665,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.013,\"estimated_latency_ms\":17881,\"estimated_pass_rate\":0.3458,\"estimated_quality\":0.5335,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.583,\"max_latency_ms\":28982,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7139, "category": "analysis", "complexity": 0.804, "consequence": 0.338, "evidence_need": 0.4539, "family": "performance_tuning", "input_tokens": 1480, "request_id": "synthetic-performance_tuning-000017", "reversibility": 0.5821, "risk": 0.252, "summary": "Review a proposed optimization for schema evolution for hidden regressions."}, "workers": [{"credits_per_call": 3.634, "id": "worker_a", "p95_latency_ms": 71535, "pass_rate": 0.8252, "provider_group": "provider_y", "quality": 0.9563, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.873, "id": "worker_b", "p95_latency_ms": 74798, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.013, "id": "worker_c", "p95_latency_ms": 17881, "pass_rate": 0.5114, "provider_group": "provider_x", "quality": 0.665, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.867, "max_latency_ms": 173430, "minimum_quality": 0.9114}, "family": "migration_planning", "id": "synthetic-migration_planning-001081", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.736, "estimated_latency_ms": 166662, "estimated_pass_rate": 0.734, "estimated_quality": 0.9282, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.867, "max_latency_ms": 173430, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.734, "quality": 0.9282, "quality_target_met": true, "utility": 71.872946}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.867,\"max_latency_ms\":173430,\"minimum_quality\":0.9114},\"task\":{\"ambiguity\":0.5989,\"category\":\"coding\",\"complexity\":0.6676,\"consequence\":0.7378,\"evidence_need\":0.6077,\"family\":\"migration_planning\",\"input_tokens\":6840,\"reversibility\":0.3712,\"risk\":0.4724,\"summary\":\"Plan a zero-downtime migration of webhook delivery with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":13.813,\"id\":\"worker_a\",\"p95_latency_ms\":58909,\"pass_rate\":0.943,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.079,\"id\":\"worker_b\",\"p95_latency_ms\":26937,\"pass_rate\":0.4809,\"provider_group\":\"provider_x\",\"quality\":0.752,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.844,\"id\":\"worker_c\",\"p95_latency_ms\":80816,\"pass_rate\":0.8765,\"provider_group\":\"provider_y\",\"quality\":0.9188,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.736,\"estimated_latency_ms\":166662,\"estimated_pass_rate\":0.734,\"estimated_quality\":0.9282,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.867,\"max_latency_ms\":173430,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5989, "category": "coding", "complexity": 0.6676, "consequence": 0.7378, "evidence_need": 0.6077, "family": "migration_planning", "input_tokens": 6840, "request_id": "synthetic-migration_planning-001081", "reversibility": 0.3712, "risk": 0.4724, "summary": "Plan a zero-downtime migration of webhook delivery with rollback checkpoints."}, "workers": [{"credits_per_call": 13.813, "id": "worker_a", "p95_latency_ms": 58909, "pass_rate": 0.943, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.079, "id": "worker_b", "p95_latency_ms": 26937, "pass_rate": 0.4809, "provider_group": "provider_x", "quality": 0.752, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.844, "id": "worker_c", "p95_latency_ms": 80816, "pass_rate": 0.8765, "provider_group": "provider_y", "quality": 0.9188, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.966, "max_latency_ms": 51483, "minimum_quality": 0.5568}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000048", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.18, "estimated_latency_ms": 25863, "estimated_pass_rate": 0.2767, "estimated_quality": 0.5323, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.966, "max_latency_ms": 51483, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2767, "quality": 0.5323, "quality_target_met": false, "utility": 21.486744}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.966,\"max_latency_ms\":51483,\"minimum_quality\":0.5568},\"task\":{\"ambiguity\":0.4173,\"category\":\"review\",\"complexity\":0.7998,\"consequence\":0.5735,\"evidence_need\":0.8894,\"family\":\"policy_compliance\",\"input_tokens\":3120,\"reversibility\":0.6523,\"risk\":0.8197,\"summary\":\"Identify evidence required before approving prompt caching for production.\"},\"workers\":[{\"credits_per_call\":1.18,\"id\":\"worker_a\",\"p95_latency_ms\":25863,\"pass_rate\":0.4653,\"provider_group\":\"provider_y\",\"quality\":0.684,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.162,\"id\":\"worker_b\",\"p95_latency_ms\":72829,\"pass_rate\":0.755,\"provider_group\":\"provider_x\",\"quality\":0.8053,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.217,\"id\":\"worker_c\",\"p95_latency_ms\":73259,\"pass_rate\":0.9015,\"provider_group\":\"provider_z\",\"quality\":0.9536,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.18,\"estimated_latency_ms\":25863,\"estimated_pass_rate\":0.2767,\"estimated_quality\":0.5323,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.966,\"max_latency_ms\":51483,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4173, "category": "review", "complexity": 0.7998, "consequence": 0.5735, "evidence_need": 0.8894, "family": "policy_compliance", "input_tokens": 3120, "request_id": "synthetic-policy_compliance-000048", "reversibility": 0.6523, "risk": 0.8197, "summary": "Identify evidence required before approving prompt caching for production."}, "workers": [{"credits_per_call": 1.18, "id": "worker_a", "p95_latency_ms": 25863, "pass_rate": 0.4653, "provider_group": "provider_y", "quality": 0.684, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.162, "id": "worker_b", "p95_latency_ms": 72829, "pass_rate": 0.755, "provider_group": "provider_x", "quality": 0.8053, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.217, "id": "worker_c", "p95_latency_ms": 73259, "pass_rate": 0.9015, "provider_group": "provider_z", "quality": 0.9536, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.086, "max_latency_ms": 173583, "minimum_quality": 0.9022}, "family": "migration_planning", "id": "synthetic-migration_planning-000181", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.509, "estimated_latency_ms": 169706, "estimated_pass_rate": 0.7502, "estimated_quality": 0.8914, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.086, "max_latency_ms": 173583, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7502, "quality": 0.8914, "quality_target_met": false, "utility": 68.102628}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.086,\"max_latency_ms\":173583,\"minimum_quality\":0.9022},\"task\":{\"ambiguity\":0.496,\"category\":\"coding\",\"complexity\":0.7483,\"consequence\":0.6118,\"evidence_need\":0.427,\"family\":\"migration_planning\",\"input_tokens\":1280,\"reversibility\":0.4719,\"risk\":0.4801,\"summary\":\"Design compatibility testing for a observability schema migration.\"},\"workers\":[{\"credits_per_call\":14.301,\"id\":\"worker_a\",\"p95_latency_ms\":61119,\"pass_rate\":0.9627,\"provider_group\":\"provider_z\",\"quality\":0.952,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.146,\"id\":\"worker_b\",\"p95_latency_ms\":30529,\"pass_rate\":0.5981,\"provider_group\":\"provider_x\",\"quality\":0.7729,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.062,\"id\":\"worker_c\",\"p95_latency_ms\":78058,\"pass_rate\":0.7729,\"provider_group\":\"provider_y\",\"quality\":0.8092,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.509,\"estimated_latency_ms\":169706,\"estimated_pass_rate\":0.7502,\"estimated_quality\":0.8914,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.086,\"max_latency_ms\":173583,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.496, "category": "coding", "complexity": 0.7483, "consequence": 0.6118, "evidence_need": 0.427, "family": "migration_planning", "input_tokens": 1280, "request_id": "synthetic-migration_planning-000181", "reversibility": 0.4719, "risk": 0.4801, "summary": "Design compatibility testing for a observability schema migration."}, "workers": [{"credits_per_call": 14.301, "id": "worker_a", "p95_latency_ms": 61119, "pass_rate": 0.9627, "provider_group": "provider_z", "quality": 0.952, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.146, "id": "worker_b", "p95_latency_ms": 30529, "pass_rate": 0.5981, "provider_group": "provider_x", "quality": 0.7729, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.062, "id": "worker_c", "p95_latency_ms": 78058, "pass_rate": 0.7729, "provider_group": "provider_y", "quality": 0.8092, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.185, "max_latency_ms": 25353, "minimum_quality": 0.7011}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000737", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.927, "estimated_latency_ms": 22852, "estimated_pass_rate": 0.3559, "estimated_quality": 0.5919, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.185, "max_latency_ms": 25353, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3559, "quality": 0.5919, "quality_target_met": false, "utility": 34.203649}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.185,\"max_latency_ms\":25353,\"minimum_quality\":0.7011},\"task\":{\"ambiguity\":0.6461,\"category\":\"analysis\",\"complexity\":0.4855,\"consequence\":0.5356,\"evidence_need\":0.7465,\"family\":\"performance_tuning\",\"input_tokens\":7440,\"reversibility\":0.8688,\"risk\":0.498,\"summary\":\"Choose between throughput and tail-latency changes for deployment rollback.\"},\"workers\":[{\"credits_per_call\":12.619,\"id\":\"worker_a\",\"p95_latency_ms\":66292,\"pass_rate\":0.9724,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.927,\"id\":\"worker_b\",\"p95_latency_ms\":22852,\"pass_rate\":0.5001,\"provider_group\":\"provider_x\",\"quality\":0.679,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.243,\"id\":\"worker_c\",\"p95_latency_ms\":82573,\"pass_rate\":0.8053,\"provider_group\":\"provider_y\",\"quality\":0.9703,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.927,\"estimated_latency_ms\":22852,\"estimated_pass_rate\":0.3559,\"estimated_quality\":0.5919,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.185,\"max_latency_ms\":25353,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6461, "category": "analysis", "complexity": 0.4855, "consequence": 0.5356, "evidence_need": 0.7465, "family": "performance_tuning", "input_tokens": 7440, "request_id": "synthetic-performance_tuning-000737", "reversibility": 0.8688, "risk": 0.498, "summary": "Choose between throughput and tail-latency changes for deployment rollback."}, "workers": [{"credits_per_call": 12.619, "id": "worker_a", "p95_latency_ms": 66292, "pass_rate": 0.9724, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.927, "id": "worker_b", "p95_latency_ms": 22852, "pass_rate": 0.5001, "provider_group": "provider_x", "quality": 0.679, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.243, "id": "worker_c", "p95_latency_ms": 82573, "pass_rate": 0.8053, "provider_group": "provider_y", "quality": 0.9703, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.968, "max_latency_ms": 29947, "minimum_quality": 0.4852}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000333", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.082, "estimated_latency_ms": 24231, "estimated_pass_rate": 0.3414, "estimated_quality": 0.6429, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.968, "max_latency_ms": 29947, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3414, "quality": 0.6429, "quality_target_met": true, "utility": 26.667164}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.968,\"max_latency_ms\":29947,\"minimum_quality\":0.4852},\"task\":{\"ambiguity\":0.6586,\"category\":\"review\",\"complexity\":0.7171,\"consequence\":0.8821,\"evidence_need\":0.7674,\"family\":\"policy_compliance\",\"input_tokens\":6560,\"reversibility\":0.3839,\"risk\":0.7007,\"summary\":\"Audit a workflow for background jobs against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.802,\"id\":\"worker_a\",\"p95_latency_ms\":92193,\"pass_rate\":0.9403,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.112,\"id\":\"worker_b\",\"p95_latency_ms\":79932,\"pass_rate\":0.7642,\"provider_group\":\"provider_y\",\"quality\":0.8617,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.082,\"id\":\"worker_c\",\"p95_latency_ms\":24231,\"pass_rate\":0.5293,\"provider_group\":\"provider_z\",\"quality\":0.7873,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.082,\"estimated_latency_ms\":24231,\"estimated_pass_rate\":0.3414,\"estimated_quality\":0.6429,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.968,\"max_latency_ms\":29947,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6586, "category": "review", "complexity": 0.7171, "consequence": 0.8821, "evidence_need": 0.7674, "family": "policy_compliance", "input_tokens": 6560, "request_id": "synthetic-policy_compliance-000333", "reversibility": 0.3839, "risk": 0.7007, "summary": "Audit a workflow for background jobs against explicit access requirements."}, "workers": [{"credits_per_call": 13.802, "id": "worker_a", "p95_latency_ms": 92193, "pass_rate": 0.9403, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.112, "id": "worker_b", "p95_latency_ms": 79932, "pass_rate": 0.7642, "provider_group": "provider_y", "quality": 0.8617, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.082, "id": "worker_c", "p95_latency_ms": 24231, "pass_rate": 0.5293, "provider_group": "provider_z", "quality": 0.7873, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.731, "max_latency_ms": 35470, "minimum_quality": 0.6794}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000756", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.916, "estimated_latency_ms": 25256, "estimated_pass_rate": 0.4399, "estimated_quality": 0.6092, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.731, "max_latency_ms": 35470, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4399, "quality": 0.6092, "quality_target_met": false, "utility": 38.03131}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.731,\"max_latency_ms\":35470,\"minimum_quality\":0.6794},\"task\":{\"ambiguity\":0.3622,\"category\":\"review\",\"complexity\":0.8078,\"consequence\":0.5077,\"evidence_need\":0.8119,\"family\":\"policy_compliance\",\"input_tokens\":2440,\"reversibility\":0.6095,\"risk\":0.641,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":3.149,\"id\":\"worker_a\",\"p95_latency_ms\":59852,\"pass_rate\":0.878,\"provider_group\":\"provider_z\",\"quality\":0.9068,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.439,\"id\":\"worker_b\",\"p95_latency_ms\":65644,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9715,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.916,\"id\":\"worker_c\",\"p95_latency_ms\":25256,\"pass_rate\":0.6108,\"provider_group\":\"provider_x\",\"quality\":0.7457,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.916,\"estimated_latency_ms\":25256,\"estimated_pass_rate\":0.4399,\"estimated_quality\":0.6092,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.731,\"max_latency_ms\":35470,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3622, "category": "review", "complexity": 0.8078, "consequence": 0.5077, "evidence_need": 0.8119, "family": "policy_compliance", "input_tokens": 2440, "request_id": "synthetic-policy_compliance-000756", "reversibility": 0.6095, "risk": 0.641, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 3.149, "id": "worker_a", "p95_latency_ms": 59852, "pass_rate": 0.878, "provider_group": "provider_z", "quality": 0.9068, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.439, "id": "worker_b", "p95_latency_ms": 65644, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9715, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.916, "id": "worker_c", "p95_latency_ms": 25256, "pass_rate": 0.6108, "provider_group": "provider_x", "quality": 0.7457, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.123, "max_latency_ms": 33639, "minimum_quality": 0.5388}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000407", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.097, "estimated_latency_ms": 19904, "estimated_pass_rate": 0.2893, "estimated_quality": 0.6328, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.123, "max_latency_ms": 33639, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2893, "quality": 0.6328, "quality_target_met": true, "utility": 40.766023}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.123,\"max_latency_ms\":33639,\"minimum_quality\":0.5388},\"task\":{\"ambiguity\":0.5161,\"category\":\"analysis\",\"complexity\":0.4703,\"consequence\":0.3328,\"evidence_need\":0.5178,\"family\":\"performance_tuning\",\"input_tokens\":5440,\"reversibility\":0.5629,\"risk\":0.3083,\"summary\":\"Review a proposed optimization for private credentials for hidden regressions.\"},\"workers\":[{\"credits_per_call\":13.965,\"id\":\"worker_a\",\"p95_latency_ms\":83523,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.097,\"id\":\"worker_b\",\"p95_latency_ms\":19904,\"pass_rate\":0.4066,\"provider_group\":\"provider_y\",\"quality\":0.6947,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.912,\"id\":\"worker_c\",\"p95_latency_ms\":51882,\"pass_rate\":0.9192,\"provider_group\":\"provider_x\",\"quality\":0.9071,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.097,\"estimated_latency_ms\":19904,\"estimated_pass_rate\":0.2893,\"estimated_quality\":0.6328,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.123,\"max_latency_ms\":33639,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5161, "category": "analysis", "complexity": 0.4703, "consequence": 0.3328, "evidence_need": 0.5178, "family": "performance_tuning", "input_tokens": 5440, "request_id": "synthetic-performance_tuning-000407", "reversibility": 0.5629, "risk": 0.3083, "summary": "Review a proposed optimization for private credentials for hidden regressions."}, "workers": [{"credits_per_call": 13.965, "id": "worker_a", "p95_latency_ms": 83523, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.097, "id": "worker_b", "p95_latency_ms": 19904, "pass_rate": 0.4066, "provider_group": "provider_y", "quality": 0.6947, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.912, "id": "worker_c", "p95_latency_ms": 51882, "pass_rate": 0.9192, "provider_group": "provider_x", "quality": 0.9071, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.039, "max_latency_ms": 193873, "minimum_quality": 0.7086}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001176", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.394, "estimated_latency_ms": 94206, "estimated_pass_rate": 0.7678, "estimated_quality": 0.8957, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.039, "max_latency_ms": 193873, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7678, "quality": 0.8957, "quality_target_met": true, "utility": 68.55519}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.039,\"max_latency_ms\":193873,\"minimum_quality\":0.7086},\"task\":{\"ambiguity\":0.5603,\"category\":\"review\",\"complexity\":0.7636,\"consequence\":0.6373,\"evidence_need\":0.7675,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.5711,\"risk\":0.7847,\"summary\":\"Review whether private credentials satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.076,\"id\":\"worker_a\",\"p95_latency_ms\":24129,\"pass_rate\":0.5806,\"provider_group\":\"provider_x\",\"quality\":0.7239,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.723,\"id\":\"worker_b\",\"p95_latency_ms\":81620,\"pass_rate\":0.8001,\"provider_group\":\"provider_y\",\"quality\":0.9118,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.318,\"id\":\"worker_c\",\"p95_latency_ms\":70077,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.394,\"estimated_latency_ms\":94206,\"estimated_pass_rate\":0.7678,\"estimated_quality\":0.8957,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.039,\"max_latency_ms\":193873,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5603, "category": "review", "complexity": 0.7636, "consequence": 0.6373, "evidence_need": 0.7675, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-001176", "reversibility": 0.5711, "risk": 0.7847, "summary": "Review whether private credentials satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.076, "id": "worker_a", "p95_latency_ms": 24129, "pass_rate": 0.5806, "provider_group": "provider_x", "quality": 0.7239, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.723, "id": "worker_b", "p95_latency_ms": 81620, "pass_rate": 0.8001, "provider_group": "provider_y", "quality": 0.9118, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.318, "id": "worker_c", "p95_latency_ms": 70077, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.782, "max_latency_ms": 182742, "minimum_quality": 0.8146}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000408", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.96, "estimated_latency_ms": 154540, "estimated_pass_rate": 0.7845, "estimated_quality": 0.9218, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.782, "max_latency_ms": 182742, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7845, "quality": 0.9218, "quality_target_met": true, "utility": 82.362071}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.782,\"max_latency_ms\":182742,\"minimum_quality\":0.8146},\"task\":{\"ambiguity\":0.451,\"category\":\"review\",\"complexity\":0.7521,\"consequence\":0.9258,\"evidence_need\":0.7848,\"family\":\"policy_compliance\",\"input_tokens\":1560,\"reversibility\":0.4493,\"risk\":0.8486,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.049,\"id\":\"worker_a\",\"p95_latency_ms\":47167,\"pass_rate\":0.8632,\"provider_group\":\"provider_z\",\"quality\":0.898,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.963,\"id\":\"worker_b\",\"p95_latency_ms\":80239,\"pass_rate\":0.9494,\"provider_group\":\"provider_x\",\"quality\":0.9667,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.948,\"id\":\"worker_c\",\"p95_latency_ms\":27134,\"pass_rate\":0.552,\"provider_group\":\"provider_y\",\"quality\":0.7862,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.96,\"estimated_latency_ms\":154540,\"estimated_pass_rate\":0.7845,\"estimated_quality\":0.9218,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.782,\"max_latency_ms\":182742,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.451, "category": "review", "complexity": 0.7521, "consequence": 0.9258, "evidence_need": 0.7848, "family": "policy_compliance", "input_tokens": 1560, "request_id": "synthetic-policy_compliance-000408", "reversibility": 0.4493, "risk": 0.8486, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.049, "id": "worker_a", "p95_latency_ms": 47167, "pass_rate": 0.8632, "provider_group": "provider_z", "quality": 0.898, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.963, "id": "worker_b", "p95_latency_ms": 80239, "pass_rate": 0.9494, "provider_group": "provider_x", "quality": 0.9667, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.948, "id": "worker_c", "p95_latency_ms": 27134, "pass_rate": 0.552, "provider_group": "provider_y", "quality": 0.7862, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 5.626, "max_latency_ms": 131278, "minimum_quality": 0.868}, "family": "migration_planning", "id": "synthetic-migration_planning-000958", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.079, "estimated_latency_ms": 96772, "estimated_pass_rate": 0.651, "estimated_quality": 0.7964, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.626, "max_latency_ms": 131278, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.651, "quality": 0.7964, "quality_target_met": false, "utility": 53.595011}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":5.626,\"max_latency_ms\":131278,\"minimum_quality\":0.868},\"task\":{\"ambiguity\":0.5769,\"category\":\"coding\",\"complexity\":0.6991,\"consequence\":0.5377,\"evidence_need\":0.7916,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.4896,\"risk\":0.3668,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":12.249,\"id\":\"worker_a\",\"p95_latency_ms\":76734,\"pass_rate\":0.9744,\"provider_group\":\"provider_z\",\"quality\":0.9362,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.155,\"id\":\"worker_b\",\"p95_latency_ms\":26010,\"pass_rate\":0.6073,\"provider_group\":\"provider_x\",\"quality\":0.6962,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.924,\"id\":\"worker_c\",\"p95_latency_ms\":70762,\"pass_rate\":0.8718,\"provider_group\":\"provider_y\",\"quality\":0.8542,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.079,\"estimated_latency_ms\":96772,\"estimated_pass_rate\":0.651,\"estimated_quality\":0.7964,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.626,\"max_latency_ms\":131278,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5769, "category": "coding", "complexity": 0.6991, "consequence": 0.5377, "evidence_need": 0.7916, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-000958", "reversibility": 0.4896, "risk": 0.3668, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 12.249, "id": "worker_a", "p95_latency_ms": 76734, "pass_rate": 0.9744, "provider_group": "provider_z", "quality": 0.9362, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.155, "id": "worker_b", "p95_latency_ms": 26010, "pass_rate": 0.6073, "provider_group": "provider_x", "quality": 0.6962, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.924, "id": "worker_c", "p95_latency_ms": 70762, "pass_rate": 0.8718, "provider_group": "provider_y", "quality": 0.8542, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.152, "max_latency_ms": 149370, "minimum_quality": 0.9195}, "family": "migration_planning", "id": "synthetic-migration_planning-000868", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.317, "estimated_latency_ms": 139363, "estimated_pass_rate": 0.6809, "estimated_quality": 0.8621, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.152, "max_latency_ms": 149370, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6809, "quality": 0.8621, "quality_target_met": false, "utility": 68.548966}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.152,\"max_latency_ms\":149370,\"minimum_quality\":0.9195},\"task\":{\"ambiguity\":0.4218,\"category\":\"coding\",\"complexity\":0.9045,\"consequence\":0.704,\"evidence_need\":0.6504,\"family\":\"migration_planning\",\"input_tokens\":1160,\"reversibility\":0.5412,\"risk\":0.5867,\"summary\":\"Plan a zero-downtime migration of access control with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.44,\"id\":\"worker_a\",\"p95_latency_ms\":57327,\"pass_rate\":0.7248,\"provider_group\":\"provider_x\",\"quality\":0.8161,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.847,\"id\":\"worker_b\",\"p95_latency_ms\":57246,\"pass_rate\":0.892,\"provider_group\":\"provider_y\",\"quality\":0.9309,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.03,\"id\":\"worker_c\",\"p95_latency_ms\":24790,\"pass_rate\":0.5132,\"provider_group\":\"provider_z\",\"quality\":0.7102,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.317,\"estimated_latency_ms\":139363,\"estimated_pass_rate\":0.6809,\"estimated_quality\":0.8621,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.152,\"max_latency_ms\":149370,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4218, "category": "coding", "complexity": 0.9045, "consequence": 0.704, "evidence_need": 0.6504, "family": "migration_planning", "input_tokens": 1160, "request_id": "synthetic-migration_planning-000868", "reversibility": 0.5412, "risk": 0.5867, "summary": "Plan a zero-downtime migration of access control with rollback checkpoints."}, "workers": [{"credits_per_call": 3.44, "id": "worker_a", "p95_latency_ms": 57327, "pass_rate": 0.7248, "provider_group": "provider_x", "quality": 0.8161, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.847, "id": "worker_b", "p95_latency_ms": 57246, "pass_rate": 0.892, "provider_group": "provider_y", "quality": 0.9309, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.03, "id": "worker_c", "p95_latency_ms": 24790, "pass_rate": 0.5132, "provider_group": "provider_z", "quality": 0.7102, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.787, "max_latency_ms": 33350, "minimum_quality": 0.5816}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000050", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.106, "estimated_latency_ms": 27183, "estimated_pass_rate": 0.3695, "estimated_quality": 0.6329, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.787, "max_latency_ms": 33350, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3695, "quality": 0.6329, "quality_target_met": true, "utility": 42.026067}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.787,\"max_latency_ms\":33350,\"minimum_quality\":0.5816},\"task\":{\"ambiguity\":0.3498,\"category\":\"analysis\",\"complexity\":0.6713,\"consequence\":0.3477,\"evidence_need\":0.5472,\"family\":\"performance_tuning\",\"input_tokens\":4440,\"reversibility\":0.5932,\"risk\":0.2542,\"summary\":\"Diagnose latency variance in private credentials and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.106,\"id\":\"worker_a\",\"p95_latency_ms\":27183,\"pass_rate\":0.4927,\"provider_group\":\"provider_z\",\"quality\":0.7161,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.06,\"id\":\"worker_b\",\"p95_latency_ms\":77599,\"pass_rate\":0.9521,\"provider_group\":\"provider_x\",\"quality\":0.8748,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.949,\"id\":\"worker_c\",\"p95_latency_ms\":92063,\"pass_rate\":0.9686,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.106,\"estimated_latency_ms\":27183,\"estimated_pass_rate\":0.3695,\"estimated_quality\":0.6329,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.787,\"max_latency_ms\":33350,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3498, "category": "analysis", "complexity": 0.6713, "consequence": 0.3477, "evidence_need": 0.5472, "family": "performance_tuning", "input_tokens": 4440, "request_id": "synthetic-performance_tuning-000050", "reversibility": 0.5932, "risk": 0.2542, "summary": "Diagnose latency variance in private credentials and rank the next experiments."}, "workers": [{"credits_per_call": 1.106, "id": "worker_a", "p95_latency_ms": 27183, "pass_rate": 0.4927, "provider_group": "provider_z", "quality": 0.7161, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.06, "id": "worker_b", "p95_latency_ms": 77599, "pass_rate": 0.9521, "provider_group": "provider_x", "quality": 0.8748, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.949, "id": "worker_c", "p95_latency_ms": 92063, "pass_rate": 0.9686, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.336, "max_latency_ms": 187950, "minimum_quality": 0.9366}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000936", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.025, "estimated_latency_ms": 173539, "estimated_pass_rate": 0.7419, "estimated_quality": 0.9041, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.336, "max_latency_ms": 187950, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7419, "quality": 0.9041, "quality_target_met": false, "utility": 73.833284}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.336,\"max_latency_ms\":187950,\"minimum_quality\":0.9366},\"task\":{\"ambiguity\":0.512,\"category\":\"review\",\"complexity\":0.7594,\"consequence\":0.5826,\"evidence_need\":0.7046,\"family\":\"policy_compliance\",\"input_tokens\":5280,\"reversibility\":0.4125,\"risk\":0.8063,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.035,\"id\":\"worker_a\",\"p95_latency_ms\":26348,\"pass_rate\":0.49,\"provider_group\":\"provider_z\",\"quality\":0.7758,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.964,\"id\":\"worker_b\",\"p95_latency_ms\":67742,\"pass_rate\":0.8138,\"provider_group\":\"provider_y\",\"quality\":0.796,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.026,\"id\":\"worker_c\",\"p95_latency_ms\":79449,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9622,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.025,\"estimated_latency_ms\":173539,\"estimated_pass_rate\":0.7419,\"estimated_quality\":0.9041,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.336,\"max_latency_ms\":187950,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.512, "category": "review", "complexity": 0.7594, "consequence": 0.5826, "evidence_need": 0.7046, "family": "policy_compliance", "input_tokens": 5280, "request_id": "synthetic-policy_compliance-000936", "reversibility": 0.4125, "risk": 0.8063, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 1.035, "id": "worker_a", "p95_latency_ms": 26348, "pass_rate": 0.49, "provider_group": "provider_z", "quality": 0.7758, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.964, "id": "worker_b", "p95_latency_ms": 67742, "pass_rate": 0.8138, "provider_group": "provider_y", "quality": 0.796, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.026, "id": "worker_c", "p95_latency_ms": 79449, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9622, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.301, "max_latency_ms": 192904, "minimum_quality": 0.8556}, "family": "migration_planning", "id": "synthetic-migration_planning-000967", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.965, "estimated_latency_ms": 161122, "estimated_pass_rate": 0.7284, "estimated_quality": 0.898, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.301, "max_latency_ms": 192904, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7284, "quality": 0.898, "quality_target_met": true, "utility": 74.130234}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.301,\"max_latency_ms\":192904,\"minimum_quality\":0.8556},\"task\":{\"ambiguity\":0.518,\"category\":\"coding\",\"complexity\":0.8558,\"consequence\":0.5833,\"evidence_need\":0.594,\"family\":\"migration_planning\",\"input_tokens\":1560,\"reversibility\":0.5131,\"risk\":0.7029,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.013,\"id\":\"worker_a\",\"p95_latency_ms\":20020,\"pass_rate\":0.4552,\"provider_group\":\"provider_x\",\"quality\":0.7445,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.711,\"id\":\"worker_b\",\"p95_latency_ms\":68156,\"pass_rate\":0.8352,\"provider_group\":\"provider_z\",\"quality\":0.8217,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.241,\"id\":\"worker_c\",\"p95_latency_ms\":72946,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9632,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.965,\"estimated_latency_ms\":161122,\"estimated_pass_rate\":0.7284,\"estimated_quality\":0.898,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.301,\"max_latency_ms\":192904,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.518, "category": "coding", "complexity": 0.8558, "consequence": 0.5833, "evidence_need": 0.594, "family": "migration_planning", "input_tokens": 1560, "request_id": "synthetic-migration_planning-000967", "reversibility": 0.5131, "risk": 0.7029, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 1.013, "id": "worker_a", "p95_latency_ms": 20020, "pass_rate": 0.4552, "provider_group": "provider_x", "quality": 0.7445, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.711, "id": "worker_b", "p95_latency_ms": 68156, "pass_rate": 0.8352, "provider_group": "provider_z", "quality": 0.8217, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.241, "id": "worker_c", "p95_latency_ms": 72946, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9632, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.649, "max_latency_ms": 39398, "minimum_quality": 0.6973}, "family": "migration_planning", "id": "synthetic-migration_planning-000307", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.833, "estimated_latency_ms": 31361, "estimated_pass_rate": 0.3093, "estimated_quality": 0.6529, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.649, "max_latency_ms": 39398, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3093, "quality": 0.6529, "quality_target_met": false, "utility": 25.320291}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.649,\"max_latency_ms\":39398,\"minimum_quality\":0.6973},\"task\":{\"ambiguity\":0.6549,\"category\":\"coding\",\"complexity\":0.6273,\"consequence\":0.8216,\"evidence_need\":0.7131,\"family\":\"migration_planning\",\"input_tokens\":1040,\"reversibility\":0.4563,\"risk\":0.73,\"summary\":\"Plan a zero-downtime migration of provider failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.263,\"id\":\"worker_a\",\"p95_latency_ms\":54885,\"pass_rate\":0.7724,\"provider_group\":\"provider_z\",\"quality\":0.9007,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.833,\"id\":\"worker_b\",\"p95_latency_ms\":31361,\"pass_rate\":0.4891,\"provider_group\":\"provider_y\",\"quality\":0.7829,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.139,\"id\":\"worker_c\",\"p95_latency_ms\":72571,\"pass_rate\":0.976,\"provider_group\":\"provider_x\",\"quality\":0.9517,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.833,\"estimated_latency_ms\":31361,\"estimated_pass_rate\":0.3093,\"estimated_quality\":0.6529,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.649,\"max_latency_ms\":39398,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6549, "category": "coding", "complexity": 0.6273, "consequence": 0.8216, "evidence_need": 0.7131, "family": "migration_planning", "input_tokens": 1040, "request_id": "synthetic-migration_planning-000307", "reversibility": 0.4563, "risk": 0.73, "summary": "Plan a zero-downtime migration of provider failover with rollback checkpoints."}, "workers": [{"credits_per_call": 4.263, "id": "worker_a", "p95_latency_ms": 54885, "pass_rate": 0.7724, "provider_group": "provider_z", "quality": 0.9007, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.833, "id": "worker_b", "p95_latency_ms": 31361, "pass_rate": 0.4891, "provider_group": "provider_y", "quality": 0.7829, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.139, "id": "worker_c", "p95_latency_ms": 72571, "pass_rate": 0.976, "provider_group": "provider_x", "quality": 0.9517, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.424, "max_latency_ms": 36334, "minimum_quality": 0.7178}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000176", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.994, "estimated_latency_ms": 27677, "estimated_pass_rate": 0.3488, "estimated_quality": 0.6007, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.424, "max_latency_ms": 36334, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3488, "quality": 0.6007, "quality_target_met": false, "utility": 32.033542}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.424,\"max_latency_ms\":36334,\"minimum_quality\":0.7178},\"task\":{\"ambiguity\":0.4044,\"category\":\"analysis\",\"complexity\":0.7496,\"consequence\":0.4042,\"evidence_need\":0.339,\"family\":\"performance_tuning\",\"input_tokens\":5280,\"reversibility\":0.8715,\"risk\":0.3411,\"summary\":\"Choose between throughput and tail-latency changes for artifact storage.\"},\"workers\":[{\"credits_per_call\":11.026,\"id\":\"worker_a\",\"p95_latency_ms\":73795,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.994,\"id\":\"worker_b\",\"p95_latency_ms\":27677,\"pass_rate\":0.492,\"provider_group\":\"provider_x\",\"quality\":0.708,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.658,\"id\":\"worker_c\",\"p95_latency_ms\":56244,\"pass_rate\":0.9067,\"provider_group\":\"provider_y\",\"quality\":0.9279,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.994,\"estimated_latency_ms\":27677,\"estimated_pass_rate\":0.3488,\"estimated_quality\":0.6007,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.424,\"max_latency_ms\":36334,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4044, "category": "analysis", "complexity": 0.7496, "consequence": 0.4042, "evidence_need": 0.339, "family": "performance_tuning", "input_tokens": 5280, "request_id": "synthetic-performance_tuning-000176", "reversibility": 0.8715, "risk": 0.3411, "summary": "Choose between throughput and tail-latency changes for artifact storage."}, "workers": [{"credits_per_call": 11.026, "id": "worker_a", "p95_latency_ms": 73795, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.994, "id": "worker_b", "p95_latency_ms": 27677, "pass_rate": 0.492, "provider_group": "provider_x", "quality": 0.708, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.658, "id": "worker_c", "p95_latency_ms": 56244, "pass_rate": 0.9067, "provider_group": "provider_y", "quality": 0.9279, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.27, "max_latency_ms": 260293, "minimum_quality": 0.8592}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000345", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.932, "estimated_latency_ms": 154299, "estimated_pass_rate": 0.892, "estimated_quality": 0.8943, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.27, "max_latency_ms": 260293, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.892, "quality": 0.8943, "quality_target_met": true, "utility": 70.472998}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.27,\"max_latency_ms\":260293,\"minimum_quality\":0.8592},\"task\":{\"ambiguity\":0.7563,\"category\":\"review\",\"complexity\":0.5247,\"consequence\":0.816,\"evidence_need\":0.6044,\"family\":\"policy_compliance\",\"input_tokens\":840,\"reversibility\":0.4757,\"risk\":0.5877,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.22,\"id\":\"worker_a\",\"p95_latency_ms\":18892,\"pass_rate\":0.5473,\"provider_group\":\"provider_z\",\"quality\":0.7403,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.887,\"id\":\"worker_b\",\"p95_latency_ms\":87020,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9255,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.045,\"id\":\"worker_c\",\"p95_latency_ms\":67279,\"pass_rate\":0.8011,\"provider_group\":\"provider_y\",\"quality\":0.8324,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.932,\"estimated_latency_ms\":154299,\"estimated_pass_rate\":0.892,\"estimated_quality\":0.8943,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.27,\"max_latency_ms\":260293,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7563, "category": "review", "complexity": 0.5247, "consequence": 0.816, "evidence_need": 0.6044, "family": "policy_compliance", "input_tokens": 840, "request_id": "synthetic-policy_compliance-000345", "reversibility": 0.4757, "risk": 0.5877, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 1.22, "id": "worker_a", "p95_latency_ms": 18892, "pass_rate": 0.5473, "provider_group": "provider_z", "quality": 0.7403, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.887, "id": "worker_b", "p95_latency_ms": 87020, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9255, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.045, "id": "worker_c", "p95_latency_ms": 67279, "pass_rate": 0.8011, "provider_group": "provider_y", "quality": 0.8324, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.297, "max_latency_ms": 168420, "minimum_quality": 0.9327}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000126", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.409, "estimated_latency_ms": 133597, "estimated_pass_rate": 0.701, "estimated_quality": 0.8671, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.297, "max_latency_ms": 168420, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.701, "quality": 0.8671, "quality_target_met": false, "utility": 67.219321}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.297,\"max_latency_ms\":168420,\"minimum_quality\":0.9327},\"task\":{\"ambiguity\":0.7338,\"category\":\"review\",\"complexity\":0.6223,\"consequence\":0.6727,\"evidence_need\":0.6956,\"family\":\"policy_compliance\",\"input_tokens\":1680,\"reversibility\":0.6383,\"risk\":0.5452,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":0.796,\"id\":\"worker_a\",\"p95_latency_ms\":19698,\"pass_rate\":0.5026,\"provider_group\":\"provider_z\",\"quality\":0.6864,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.922,\"id\":\"worker_b\",\"p95_latency_ms\":60558,\"pass_rate\":0.9021,\"provider_group\":\"provider_y\",\"quality\":0.9148,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.691,\"id\":\"worker_c\",\"p95_latency_ms\":53341,\"pass_rate\":0.8439,\"provider_group\":\"provider_x\",\"quality\":0.8009,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.409,\"estimated_latency_ms\":133597,\"estimated_pass_rate\":0.701,\"estimated_quality\":0.8671,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.297,\"max_latency_ms\":168420,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7338, "category": "review", "complexity": 0.6223, "consequence": 0.6727, "evidence_need": 0.6956, "family": "policy_compliance", "input_tokens": 1680, "request_id": "synthetic-policy_compliance-000126", "reversibility": 0.6383, "risk": 0.5452, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 0.796, "id": "worker_a", "p95_latency_ms": 19698, "pass_rate": 0.5026, "provider_group": "provider_z", "quality": 0.6864, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.922, "id": "worker_b", "p95_latency_ms": 60558, "pass_rate": 0.9021, "provider_group": "provider_y", "quality": 0.9148, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.691, "id": "worker_c", "p95_latency_ms": 53341, "pass_rate": 0.8439, "provider_group": "provider_x", "quality": 0.8009, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.662, "max_latency_ms": 34873, "minimum_quality": 0.7153}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000747", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.983, "estimated_latency_ms": 23131, "estimated_pass_rate": 0.2895, "estimated_quality": 0.5935, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.662, "max_latency_ms": 34873, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2895, "quality": 0.5935, "quality_target_met": false, "utility": 16.588586}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.662,\"max_latency_ms\":34873,\"minimum_quality\":0.7153},\"task\":{\"ambiguity\":0.6544,\"category\":\"review\",\"complexity\":0.7972,\"consequence\":0.8625,\"evidence_need\":0.6316,\"family\":\"policy_compliance\",\"input_tokens\":3560,\"reversibility\":0.3054,\"risk\":0.8032,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.948,\"id\":\"worker_a\",\"p95_latency_ms\":61839,\"pass_rate\":0.7675,\"provider_group\":\"provider_x\",\"quality\":0.8604,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.993,\"id\":\"worker_b\",\"p95_latency_ms\":78232,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9945,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.983,\"id\":\"worker_c\",\"p95_latency_ms\":23131,\"pass_rate\":0.4944,\"provider_group\":\"provider_z\",\"quality\":0.7594,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.983,\"estimated_latency_ms\":23131,\"estimated_pass_rate\":0.2895,\"estimated_quality\":0.5935,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.662,\"max_latency_ms\":34873,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6544, "category": "review", "complexity": 0.7972, "consequence": 0.8625, "evidence_need": 0.6316, "family": "policy_compliance", "input_tokens": 3560, "request_id": "synthetic-policy_compliance-000747", "reversibility": 0.3054, "risk": 0.8032, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.948, "id": "worker_a", "p95_latency_ms": 61839, "pass_rate": 0.7675, "provider_group": "provider_x", "quality": 0.8604, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.993, "id": "worker_b", "p95_latency_ms": 78232, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9945, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.983, "id": "worker_c", "p95_latency_ms": 23131, "pass_rate": 0.4944, "provider_group": "provider_z", "quality": 0.7594, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.223, "max_latency_ms": 236226, "minimum_quality": 0.952}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000368", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.596, "estimated_latency_ms": 189620, "estimated_pass_rate": 0.8599, "estimated_quality": 0.9323, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.223, "max_latency_ms": 236226, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8599, "quality": 0.9323, "quality_target_met": false, "utility": 70.062926}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.223,\"max_latency_ms\":236226,\"minimum_quality\":0.952},\"task\":{\"ambiguity\":0.6186,\"category\":\"analysis\",\"complexity\":0.7768,\"consequence\":0.57,\"evidence_need\":0.4285,\"family\":\"performance_tuning\",\"input_tokens\":7800,\"reversibility\":0.6908,\"risk\":0.3935,\"summary\":\"Review a proposed optimization for webhook delivery for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.806,\"id\":\"worker_a\",\"p95_latency_ms\":89447,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.033,\"id\":\"worker_b\",\"p95_latency_ms\":29739,\"pass_rate\":0.483,\"provider_group\":\"provider_z\",\"quality\":0.7139,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.757,\"id\":\"worker_c\",\"p95_latency_ms\":70434,\"pass_rate\":0.9067,\"provider_group\":\"provider_y\",\"quality\":0.9315,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.596,\"estimated_latency_ms\":189620,\"estimated_pass_rate\":0.8599,\"estimated_quality\":0.9323,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.223,\"max_latency_ms\":236226,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6186, "category": "analysis", "complexity": 0.7768, "consequence": 0.57, "evidence_need": 0.4285, "family": "performance_tuning", "input_tokens": 7800, "request_id": "synthetic-performance_tuning-000368", "reversibility": 0.6908, "risk": 0.3935, "summary": "Review a proposed optimization for webhook delivery for hidden regressions."}, "workers": [{"credits_per_call": 11.806, "id": "worker_a", "p95_latency_ms": 89447, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.033, "id": "worker_b", "p95_latency_ms": 29739, "pass_rate": 0.483, "provider_group": "provider_z", "quality": 0.7139, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.757, "id": "worker_c", "p95_latency_ms": 70434, "pass_rate": 0.9067, "provider_group": "provider_y", "quality": 0.9315, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.45, "max_latency_ms": 190097, "minimum_quality": 0.8855}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000738", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.644, "estimated_latency_ms": 184310, "estimated_pass_rate": 0.7855, "estimated_quality": 0.9259, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.45, "max_latency_ms": 190097, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7855, "quality": 0.9259, "quality_target_met": true, "utility": 77.26215}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.45,\"max_latency_ms\":190097,\"minimum_quality\":0.8855},\"task\":{\"ambiguity\":0.6882,\"category\":\"review\",\"complexity\":0.72,\"consequence\":0.921,\"evidence_need\":0.7554,\"family\":\"policy_compliance\",\"input_tokens\":2960,\"reversibility\":0.6457,\"risk\":0.51,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.826,\"id\":\"worker_a\",\"p95_latency_ms\":27662,\"pass_rate\":0.5906,\"provider_group\":\"provider_y\",\"quality\":0.714,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.05,\"id\":\"worker_b\",\"p95_latency_ms\":70085,\"pass_rate\":0.8533,\"provider_group\":\"provider_z\",\"quality\":0.909,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.768,\"id\":\"worker_c\",\"p95_latency_ms\":86563,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.644,\"estimated_latency_ms\":184310,\"estimated_pass_rate\":0.7855,\"estimated_quality\":0.9259,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.45,\"max_latency_ms\":190097,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6882, "category": "review", "complexity": 0.72, "consequence": 0.921, "evidence_need": 0.7554, "family": "policy_compliance", "input_tokens": 2960, "request_id": "synthetic-policy_compliance-000738", "reversibility": 0.6457, "risk": 0.51, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.826, "id": "worker_a", "p95_latency_ms": 27662, "pass_rate": 0.5906, "provider_group": "provider_y", "quality": 0.714, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.05, "id": "worker_b", "p95_latency_ms": 70085, "pass_rate": 0.8533, "provider_group": "provider_z", "quality": 0.909, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.768, "id": "worker_c", "p95_latency_ms": 86563, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.883, "max_latency_ms": 178448, "minimum_quality": 0.833}, "family": "migration_planning", "id": "synthetic-migration_planning-000898", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.565, "estimated_latency_ms": 164328, "estimated_pass_rate": 0.7587, "estimated_quality": 0.8747, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.883, "max_latency_ms": 178448, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7587, "quality": 0.8747, "quality_target_met": true, "utility": 73.483266}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.883,\"max_latency_ms\":178448,\"minimum_quality\":0.833},\"task\":{\"ambiguity\":0.6972,\"category\":\"coding\",\"complexity\":0.8432,\"consequence\":0.5498,\"evidence_need\":0.5379,\"family\":\"migration_planning\",\"input_tokens\":7400,\"reversibility\":0.6151,\"risk\":0.6922,\"summary\":\"Review a staged migration for tenant-scoped caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":10.652,\"id\":\"worker_a\",\"p95_latency_ms\":75698,\"pass_rate\":0.9299,\"provider_group\":\"provider_x\",\"quality\":0.9461,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.999,\"id\":\"worker_b\",\"p95_latency_ms\":26280,\"pass_rate\":0.612,\"provider_group\":\"provider_y\",\"quality\":0.6904,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.914,\"id\":\"worker_c\",\"p95_latency_ms\":62350,\"pass_rate\":0.7708,\"provider_group\":\"provider_z\",\"quality\":0.9075,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.565,\"estimated_latency_ms\":164328,\"estimated_pass_rate\":0.7587,\"estimated_quality\":0.8747,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.883,\"max_latency_ms\":178448,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6972, "category": "coding", "complexity": 0.8432, "consequence": 0.5498, "evidence_need": 0.5379, "family": "migration_planning", "input_tokens": 7400, "request_id": "synthetic-migration_planning-000898", "reversibility": 0.6151, "risk": 0.6922, "summary": "Review a staged migration for tenant-scoped caching and identify irreversible steps."}, "workers": [{"credits_per_call": 10.652, "id": "worker_a", "p95_latency_ms": 75698, "pass_rate": 0.9299, "provider_group": "provider_x", "quality": 0.9461, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.999, "id": "worker_b", "p95_latency_ms": 26280, "pass_rate": 0.612, "provider_group": "provider_y", "quality": 0.6904, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.914, "id": "worker_c", "p95_latency_ms": 62350, "pass_rate": 0.7708, "provider_group": "provider_z", "quality": 0.9075, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.323, "max_latency_ms": 42006, "minimum_quality": 0.6814}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000074", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.006, "estimated_latency_ms": 27945, "estimated_pass_rate": 0.4124, "estimated_quality": 0.5445, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.323, "max_latency_ms": 42006, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4124, "quality": 0.5445, "quality_target_met": false, "utility": 29.177969}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.323,\"max_latency_ms\":42006,\"minimum_quality\":0.6814},\"task\":{\"ambiguity\":0.5525,\"category\":\"analysis\",\"complexity\":0.637,\"consequence\":0.5676,\"evidence_need\":0.3958,\"family\":\"performance_tuning\",\"input_tokens\":5480,\"reversibility\":0.6033,\"risk\":0.3444,\"summary\":\"Review a proposed optimization for schema evolution for hidden regressions.\"},\"workers\":[{\"credits_per_call\":2.996,\"id\":\"worker_a\",\"p95_latency_ms\":53564,\"pass_rate\":0.7967,\"provider_group\":\"provider_z\",\"quality\":0.9189,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.006,\"id\":\"worker_b\",\"p95_latency_ms\":27945,\"pass_rate\":0.5543,\"provider_group\":\"provider_y\",\"quality\":0.6416,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":13.297,\"id\":\"worker_c\",\"p95_latency_ms\":91763,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.006,\"estimated_latency_ms\":27945,\"estimated_pass_rate\":0.4124,\"estimated_quality\":0.5445,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.323,\"max_latency_ms\":42006,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5525, "category": "analysis", "complexity": 0.637, "consequence": 0.5676, "evidence_need": 0.3958, "family": "performance_tuning", "input_tokens": 5480, "request_id": "synthetic-performance_tuning-000074", "reversibility": 0.6033, "risk": 0.3444, "summary": "Review a proposed optimization for schema evolution for hidden regressions."}, "workers": [{"credits_per_call": 2.996, "id": "worker_a", "p95_latency_ms": 53564, "pass_rate": 0.7967, "provider_group": "provider_z", "quality": 0.9189, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.006, "id": "worker_b", "p95_latency_ms": 27945, "pass_rate": 0.5543, "provider_group": "provider_y", "quality": 0.6416, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 13.297, "id": "worker_c", "p95_latency_ms": 91763, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.991, "max_latency_ms": 147571, "minimum_quality": 0.7129}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000357", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.819, "estimated_latency_ms": 146217, "estimated_pass_rate": 0.9184, "estimated_quality": 0.9496, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.991, "max_latency_ms": 147571, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9184, "quality": 0.9496, "quality_target_met": true, "utility": 71.202602}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.991,\"max_latency_ms\":147571,\"minimum_quality\":0.7129},\"task\":{\"ambiguity\":0.5706,\"category\":\"review\",\"complexity\":0.405,\"consequence\":0.504,\"evidence_need\":0.8261,\"family\":\"policy_compliance\",\"input_tokens\":7960,\"reversibility\":0.4851,\"risk\":0.4709,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.142,\"id\":\"worker_a\",\"p95_latency_ms\":18604,\"pass_rate\":0.4675,\"provider_group\":\"provider_z\",\"quality\":0.7511,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.104,\"id\":\"worker_b\",\"p95_latency_ms\":69712,\"pass_rate\":0.815,\"provider_group\":\"provider_x\",\"quality\":0.9179,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.715,\"id\":\"worker_c\",\"p95_latency_ms\":76505,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9593,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.819,\"estimated_latency_ms\":146217,\"estimated_pass_rate\":0.9184,\"estimated_quality\":0.9496,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.991,\"max_latency_ms\":147571,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5706, "category": "review", "complexity": 0.405, "consequence": 0.504, "evidence_need": 0.8261, "family": "policy_compliance", "input_tokens": 7960, "request_id": "synthetic-policy_compliance-000357", "reversibility": 0.4851, "risk": 0.4709, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.142, "id": "worker_a", "p95_latency_ms": 18604, "pass_rate": 0.4675, "provider_group": "provider_z", "quality": 0.7511, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.104, "id": "worker_b", "p95_latency_ms": 69712, "pass_rate": 0.815, "provider_group": "provider_x", "quality": 0.9179, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.715, "id": "worker_c", "p95_latency_ms": 76505, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9593, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 4.693, "max_latency_ms": 241019, "minimum_quality": 0.7055}, "family": "migration_planning", "id": "synthetic-migration_planning-000619", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.749, "estimated_latency_ms": 101375, "estimated_pass_rate": 0.6656, "estimated_quality": 0.7938, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.693, "max_latency_ms": 241019, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6656, "quality": 0.7938, "quality_target_met": true, "utility": 55.52144}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":4.693,\"max_latency_ms\":241019,\"minimum_quality\":0.7055},\"task\":{\"ambiguity\":0.4206,\"category\":\"coding\",\"complexity\":0.7048,\"consequence\":0.6853,\"evidence_need\":0.5398,\"family\":\"migration_planning\",\"input_tokens\":3440,\"reversibility\":0.2801,\"risk\":0.3806,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.905,\"id\":\"worker_a\",\"p95_latency_ms\":82198,\"pass_rate\":0.7251,\"provider_group\":\"provider_y\",\"quality\":0.8398,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.844,\"id\":\"worker_b\",\"p95_latency_ms\":19177,\"pass_rate\":0.5864,\"provider_group\":\"provider_z\",\"quality\":0.7244,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.269,\"id\":\"worker_c\",\"p95_latency_ms\":79829,\"pass_rate\":0.8851,\"provider_group\":\"provider_x\",\"quality\":0.9935,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.749,\"estimated_latency_ms\":101375,\"estimated_pass_rate\":0.6656,\"estimated_quality\":0.7938,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.693,\"max_latency_ms\":241019,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4206, "category": "coding", "complexity": 0.7048, "consequence": 0.6853, "evidence_need": 0.5398, "family": "migration_planning", "input_tokens": 3440, "request_id": "synthetic-migration_planning-000619", "reversibility": 0.2801, "risk": 0.3806, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 2.905, "id": "worker_a", "p95_latency_ms": 82198, "pass_rate": 0.7251, "provider_group": "provider_y", "quality": 0.8398, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.844, "id": "worker_b", "p95_latency_ms": 19177, "pass_rate": 0.5864, "provider_group": "provider_z", "quality": 0.7244, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.269, "id": "worker_c", "p95_latency_ms": 79829, "pass_rate": 0.8851, "provider_group": "provider_x", "quality": 0.9935, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.489, "max_latency_ms": 172368, "minimum_quality": 0.8034}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000588", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.402, "estimated_latency_ms": 155392, "estimated_pass_rate": 0.8602, "estimated_quality": 0.9133, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.489, "max_latency_ms": 172368, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8602, "quality": 0.9133, "quality_target_met": true, "utility": 71.3251}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.489,\"max_latency_ms\":172368,\"minimum_quality\":0.8034},\"task\":{\"ambiguity\":0.6897,\"category\":\"review\",\"complexity\":0.7808,\"consequence\":0.7372,\"evidence_need\":0.6274,\"family\":\"policy_compliance\",\"input_tokens\":4120,\"reversibility\":0.5469,\"risk\":0.8369,\"summary\":\"Identify evidence required before approving database failover for production.\"},\"workers\":[{\"credits_per_call\":10.002,\"id\":\"worker_a\",\"p95_latency_ms\":74561,\"pass_rate\":0.9522,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.972,\"id\":\"worker_b\",\"p95_latency_ms\":31170,\"pass_rate\":0.5463,\"provider_group\":\"provider_z\",\"quality\":0.7781,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.4,\"id\":\"worker_c\",\"p95_latency_ms\":80831,\"pass_rate\":0.8538,\"provider_group\":\"provider_x\",\"quality\":0.8346,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.402,\"estimated_latency_ms\":155392,\"estimated_pass_rate\":0.8602,\"estimated_quality\":0.9133,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.489,\"max_latency_ms\":172368,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6897, "category": "review", "complexity": 0.7808, "consequence": 0.7372, "evidence_need": 0.6274, "family": "policy_compliance", "input_tokens": 4120, "request_id": "synthetic-policy_compliance-000588", "reversibility": 0.5469, "risk": 0.8369, "summary": "Identify evidence required before approving database failover for production."}, "workers": [{"credits_per_call": 10.002, "id": "worker_a", "p95_latency_ms": 74561, "pass_rate": 0.9522, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.972, "id": "worker_b", "p95_latency_ms": 31170, "pass_rate": 0.5463, "provider_group": "provider_z", "quality": 0.7781, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.4, "id": "worker_c", "p95_latency_ms": 80831, "pass_rate": 0.8538, "provider_group": "provider_x", "quality": 0.8346, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.514, "max_latency_ms": 34140, "minimum_quality": 0.7387}, "family": "migration_planning", "id": "synthetic-migration_planning-000703", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.126, "estimated_latency_ms": 21031, "estimated_pass_rate": 0.3535, "estimated_quality": 0.6083, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.514, "max_latency_ms": 34140, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3535, "quality": 0.6083, "quality_target_met": false, "utility": 35.353172}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.514,\"max_latency_ms\":34140,\"minimum_quality\":0.7387},\"task\":{\"ambiguity\":0.3452,\"category\":\"coding\",\"complexity\":0.9249,\"consequence\":0.5055,\"evidence_need\":0.6435,\"family\":\"migration_planning\",\"input_tokens\":7600,\"reversibility\":0.3886,\"risk\":0.6104,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":3.127,\"id\":\"worker_a\",\"p95_latency_ms\":80522,\"pass_rate\":0.8825,\"provider_group\":\"provider_y\",\"quality\":0.8466,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.126,\"id\":\"worker_b\",\"p95_latency_ms\":21031,\"pass_rate\":0.5339,\"provider_group\":\"provider_x\",\"quality\":0.7627,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.296,\"id\":\"worker_c\",\"p95_latency_ms\":81470,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9566,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.126,\"estimated_latency_ms\":21031,\"estimated_pass_rate\":0.3535,\"estimated_quality\":0.6083,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.514,\"max_latency_ms\":34140,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3452, "category": "coding", "complexity": 0.9249, "consequence": 0.5055, "evidence_need": 0.6435, "family": "migration_planning", "input_tokens": 7600, "request_id": "synthetic-migration_planning-000703", "reversibility": 0.3886, "risk": 0.6104, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 3.127, "id": "worker_a", "p95_latency_ms": 80522, "pass_rate": 0.8825, "provider_group": "provider_y", "quality": 0.8466, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.126, "id": "worker_b", "p95_latency_ms": 21031, "pass_rate": 0.5339, "provider_group": "provider_x", "quality": 0.7627, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.296, "id": "worker_c", "p95_latency_ms": 81470, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9566, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.898, "max_latency_ms": 234234, "minimum_quality": 0.8436}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000249", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.441, "estimated_latency_ms": 172384, "estimated_pass_rate": 0.7691, "estimated_quality": 0.9082, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.898, "max_latency_ms": 234234, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7691, "quality": 0.9082, "quality_target_met": true, "utility": 77.053514}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.898,\"max_latency_ms\":234234,\"minimum_quality\":0.8436},\"task\":{\"ambiguity\":0.6396,\"category\":\"review\",\"complexity\":0.564,\"consequence\":0.8439,\"evidence_need\":0.924,\"family\":\"policy_compliance\",\"input_tokens\":1480,\"reversibility\":0.6218,\"risk\":0.6949,\"summary\":\"Review whether private credentials satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.861,\"id\":\"worker_a\",\"p95_latency_ms\":22070,\"pass_rate\":0.5519,\"provider_group\":\"provider_y\",\"quality\":0.6888,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.321,\"id\":\"worker_b\",\"p95_latency_ms\":79302,\"pass_rate\":0.7781,\"provider_group\":\"provider_x\",\"quality\":0.9026,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.259,\"id\":\"worker_c\",\"p95_latency_ms\":71012,\"pass_rate\":0.9781,\"provider_group\":\"provider_z\",\"quality\":0.9633,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.441,\"estimated_latency_ms\":172384,\"estimated_pass_rate\":0.7691,\"estimated_quality\":0.9082,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.898,\"max_latency_ms\":234234,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6396, "category": "review", "complexity": 0.564, "consequence": 0.8439, "evidence_need": 0.924, "family": "policy_compliance", "input_tokens": 1480, "request_id": "synthetic-policy_compliance-000249", "reversibility": 0.6218, "risk": 0.6949, "summary": "Review whether private credentials satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.861, "id": "worker_a", "p95_latency_ms": 22070, "pass_rate": 0.5519, "provider_group": "provider_y", "quality": 0.6888, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.321, "id": "worker_b", "p95_latency_ms": 79302, "pass_rate": 0.7781, "provider_group": "provider_x", "quality": 0.9026, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.259, "id": "worker_c", "p95_latency_ms": 71012, "pass_rate": 0.9781, "provider_group": "provider_z", "quality": 0.9633, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.057, "max_latency_ms": 133288, "minimum_quality": 0.8561}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000257", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.809, "estimated_latency_ms": 100230, "estimated_pass_rate": 0.786, "estimated_quality": 0.8215, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.057, "max_latency_ms": 133288, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.786, "quality": 0.8215, "quality_target_met": false, "utility": 59.495819}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.057,\"max_latency_ms\":133288,\"minimum_quality\":0.8561},\"task\":{\"ambiguity\":0.7212,\"category\":\"analysis\",\"complexity\":0.7753,\"consequence\":0.6283,\"evidence_need\":0.6162,\"family\":\"performance_tuning\",\"input_tokens\":7080,\"reversibility\":0.8467,\"risk\":0.2511,\"summary\":\"Choose between throughput and tail-latency changes for provider failover.\"},\"workers\":[{\"credits_per_call\":2.876,\"id\":\"worker_a\",\"p95_latency_ms\":73034,\"pass_rate\":0.9138,\"provider_group\":\"provider_z\",\"quality\":0.915,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.251,\"id\":\"worker_b\",\"p95_latency_ms\":92141,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.933,\"id\":\"worker_c\",\"p95_latency_ms\":27196,\"pass_rate\":0.5463,\"provider_group\":\"provider_x\",\"quality\":0.6955,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.809,\"estimated_latency_ms\":100230,\"estimated_pass_rate\":0.786,\"estimated_quality\":0.8215,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.057,\"max_latency_ms\":133288,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7212, "category": "analysis", "complexity": 0.7753, "consequence": 0.6283, "evidence_need": 0.6162, "family": "performance_tuning", "input_tokens": 7080, "request_id": "synthetic-performance_tuning-000257", "reversibility": 0.8467, "risk": 0.2511, "summary": "Choose between throughput and tail-latency changes for provider failover."}, "workers": [{"credits_per_call": 2.876, "id": "worker_a", "p95_latency_ms": 73034, "pass_rate": 0.9138, "provider_group": "provider_z", "quality": 0.915, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.251, "id": "worker_b", "p95_latency_ms": 92141, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.933, "id": "worker_c", "p95_latency_ms": 27196, "pass_rate": 0.5463, "provider_group": "provider_x", "quality": 0.6955, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.62, "max_latency_ms": 159132, "minimum_quality": 0.6942}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000338", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.996, "estimated_latency_ms": 119884, "estimated_pass_rate": 0.8996, "estimated_quality": 0.9212, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.62, "max_latency_ms": 159132, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8996, "quality": 0.9212, "quality_target_met": true, "utility": 69.554313}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.62,\"max_latency_ms\":159132,\"minimum_quality\":0.6942},\"task\":{\"ambiguity\":0.5162,\"category\":\"analysis\",\"complexity\":0.8801,\"consequence\":0.6071,\"evidence_need\":0.4383,\"family\":\"performance_tuning\",\"input_tokens\":7760,\"reversibility\":0.5866,\"risk\":0.55,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":2.791,\"id\":\"worker_a\",\"p95_latency_ms\":59562,\"pass_rate\":0.9377,\"provider_group\":\"provider_x\",\"quality\":0.9128,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.958,\"id\":\"worker_b\",\"p95_latency_ms\":22470,\"pass_rate\":0.4615,\"provider_group\":\"provider_z\",\"quality\":0.6355,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.205,\"id\":\"worker_c\",\"p95_latency_ms\":60322,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.996,\"estimated_latency_ms\":119884,\"estimated_pass_rate\":0.8996,\"estimated_quality\":0.9212,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.62,\"max_latency_ms\":159132,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5162, "category": "analysis", "complexity": 0.8801, "consequence": 0.6071, "evidence_need": 0.4383, "family": "performance_tuning", "input_tokens": 7760, "request_id": "synthetic-performance_tuning-000338", "reversibility": 0.5866, "risk": 0.55, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 2.791, "id": "worker_a", "p95_latency_ms": 59562, "pass_rate": 0.9377, "provider_group": "provider_x", "quality": 0.9128, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.958, "id": "worker_b", "p95_latency_ms": 22470, "pass_rate": 0.4615, "provider_group": "provider_z", "quality": 0.6355, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.205, "id": "worker_c", "p95_latency_ms": 60322, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.675, "max_latency_ms": 165616, "minimum_quality": 0.9144}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000834", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.7, "estimated_latency_ms": 142405, "estimated_pass_rate": 0.7047, "estimated_quality": 0.9217, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.675, "max_latency_ms": 165616, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7047, "quality": 0.9217, "quality_target_met": true, "utility": 77.681188}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.675,\"max_latency_ms\":165616,\"minimum_quality\":0.9144},\"task\":{\"ambiguity\":0.391,\"category\":\"review\",\"complexity\":0.8071,\"consequence\":0.9429,\"evidence_need\":0.8474,\"family\":\"policy_compliance\",\"input_tokens\":840,\"reversibility\":0.4398,\"risk\":0.6459,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.374,\"id\":\"worker_a\",\"p95_latency_ms\":74221,\"pass_rate\":0.8983,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.977,\"id\":\"worker_b\",\"p95_latency_ms\":20277,\"pass_rate\":0.5365,\"provider_group\":\"provider_x\",\"quality\":0.8044,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.349,\"id\":\"worker_c\",\"p95_latency_ms\":47907,\"pass_rate\":0.7593,\"provider_group\":\"provider_y\",\"quality\":0.7987,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.7,\"estimated_latency_ms\":142405,\"estimated_pass_rate\":0.7047,\"estimated_quality\":0.9217,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.675,\"max_latency_ms\":165616,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.391, "category": "review", "complexity": 0.8071, "consequence": 0.9429, "evidence_need": 0.8474, "family": "policy_compliance", "input_tokens": 840, "request_id": "synthetic-policy_compliance-000834", "reversibility": 0.4398, "risk": 0.6459, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 11.374, "id": "worker_a", "p95_latency_ms": 74221, "pass_rate": 0.8983, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.977, "id": "worker_b", "p95_latency_ms": 20277, "pass_rate": 0.5365, "provider_group": "provider_x", "quality": 0.8044, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.349, "id": "worker_c", "p95_latency_ms": 47907, "pass_rate": 0.7593, "provider_group": "provider_y", "quality": 0.7987, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.102, "max_latency_ms": 141171, "minimum_quality": 0.8085}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001164", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.203, "estimated_latency_ms": 126347, "estimated_pass_rate": 0.8406, "estimated_quality": 0.9107, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.102, "max_latency_ms": 141171, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8406, "quality": 0.9107, "quality_target_met": true, "utility": 68.067988}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.102,\"max_latency_ms\":141171,\"minimum_quality\":0.8085},\"task\":{\"ambiguity\":0.4852,\"category\":\"review\",\"complexity\":0.5866,\"consequence\":0.674,\"evidence_need\":0.6446,\"family\":\"policy_compliance\",\"input_tokens\":1200,\"reversibility\":0.6336,\"risk\":0.5918,\"summary\":\"Identify evidence required before approving database failover for production.\"},\"workers\":[{\"credits_per_call\":1.121,\"id\":\"worker_a\",\"p95_latency_ms\":18223,\"pass_rate\":0.5876,\"provider_group\":\"provider_y\",\"quality\":0.7493,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.138,\"id\":\"worker_b\",\"p95_latency_ms\":50415,\"pass_rate\":0.8194,\"provider_group\":\"provider_x\",\"quality\":0.8012,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.065,\"id\":\"worker_c\",\"p95_latency_ms\":75932,\"pass_rate\":0.9013,\"provider_group\":\"provider_z\",\"quality\":0.9468,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.203,\"estimated_latency_ms\":126347,\"estimated_pass_rate\":0.8406,\"estimated_quality\":0.9107,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.102,\"max_latency_ms\":141171,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4852, "category": "review", "complexity": 0.5866, "consequence": 0.674, "evidence_need": 0.6446, "family": "policy_compliance", "input_tokens": 1200, "request_id": "synthetic-policy_compliance-001164", "reversibility": 0.6336, "risk": 0.5918, "summary": "Identify evidence required before approving database failover for production."}, "workers": [{"credits_per_call": 1.121, "id": "worker_a", "p95_latency_ms": 18223, "pass_rate": 0.5876, "provider_group": "provider_y", "quality": 0.7493, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.138, "id": "worker_b", "p95_latency_ms": 50415, "pass_rate": 0.8194, "provider_group": "provider_x", "quality": 0.8012, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.065, "id": "worker_c", "p95_latency_ms": 75932, "pass_rate": 0.9013, "provider_group": "provider_z", "quality": 0.9468, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.905, "max_latency_ms": 164742, "minimum_quality": 0.8785}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000915", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.025, "estimated_latency_ms": 69599, "estimated_pass_rate": 0.6851, "estimated_quality": 0.801, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.905, "max_latency_ms": 164742, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6851, "quality": 0.801, "quality_target_met": false, "utility": 60.310881}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.905,\"max_latency_ms\":164742,\"minimum_quality\":0.8785},\"task\":{\"ambiguity\":0.467,\"category\":\"review\",\"complexity\":0.6625,\"consequence\":0.6455,\"evidence_need\":0.8652,\"family\":\"policy_compliance\",\"input_tokens\":7520,\"reversibility\":0.6382,\"risk\":0.7735,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.981,\"id\":\"worker_a\",\"p95_latency_ms\":24974,\"pass_rate\":0.617,\"provider_group\":\"provider_z\",\"quality\":0.6805,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.439,\"id\":\"worker_b\",\"p95_latency_ms\":60684,\"pass_rate\":0.925,\"provider_group\":\"provider_x\",\"quality\":0.9534,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.044,\"id\":\"worker_c\",\"p95_latency_ms\":44625,\"pass_rate\":0.8027,\"provider_group\":\"provider_y\",\"quality\":0.8698,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.025,\"estimated_latency_ms\":69599,\"estimated_pass_rate\":0.6851,\"estimated_quality\":0.801,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.905,\"max_latency_ms\":164742,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.467, "category": "review", "complexity": 0.6625, "consequence": 0.6455, "evidence_need": 0.8652, "family": "policy_compliance", "input_tokens": 7520, "request_id": "synthetic-policy_compliance-000915", "reversibility": 0.6382, "risk": 0.7735, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.981, "id": "worker_a", "p95_latency_ms": 24974, "pass_rate": 0.617, "provider_group": "provider_z", "quality": 0.6805, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.439, "id": "worker_b", "p95_latency_ms": 60684, "pass_rate": 0.925, "provider_group": "provider_x", "quality": 0.9534, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.044, "id": "worker_c", "p95_latency_ms": 44625, "pass_rate": 0.8027, "provider_group": "provider_y", "quality": 0.8698, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.253, "max_latency_ms": 132729, "minimum_quality": 0.8373}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000687", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.85, "estimated_latency_ms": 95074, "estimated_pass_rate": 0.675, "estimated_quality": 0.8454, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.253, "max_latency_ms": 132729, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.675, "quality": 0.8454, "quality_target_met": true, "utility": 62.479197}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.253,\"max_latency_ms\":132729,\"minimum_quality\":0.8373},\"task\":{\"ambiguity\":0.5237,\"category\":\"review\",\"complexity\":0.5132,\"consequence\":0.8654,\"evidence_need\":0.761,\"family\":\"policy_compliance\",\"input_tokens\":4040,\"reversibility\":0.6438,\"risk\":0.8248,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.018,\"id\":\"worker_a\",\"p95_latency_ms\":88683,\"pass_rate\":0.9007,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.96,\"id\":\"worker_b\",\"p95_latency_ms\":19502,\"pass_rate\":0.4816,\"provider_group\":\"provider_x\",\"quality\":0.7364,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.89,\"id\":\"worker_c\",\"p95_latency_ms\":75572,\"pass_rate\":0.7504,\"provider_group\":\"provider_y\",\"quality\":0.9132,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.85,\"estimated_latency_ms\":95074,\"estimated_pass_rate\":0.675,\"estimated_quality\":0.8454,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.253,\"max_latency_ms\":132729,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5237, "category": "review", "complexity": 0.5132, "consequence": 0.8654, "evidence_need": 0.761, "family": "policy_compliance", "input_tokens": 4040, "request_id": "synthetic-policy_compliance-000687", "reversibility": 0.6438, "risk": 0.8248, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 10.018, "id": "worker_a", "p95_latency_ms": 88683, "pass_rate": 0.9007, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.96, "id": "worker_b", "p95_latency_ms": 19502, "pass_rate": 0.4816, "provider_group": "provider_x", "quality": 0.7364, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.89, "id": "worker_c", "p95_latency_ms": 75572, "pass_rate": 0.7504, "provider_group": "provider_y", "quality": 0.9132, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.645, "max_latency_ms": 202764, "minimum_quality": 0.8831}, "family": "migration_planning", "id": "synthetic-migration_planning-000433", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.146, "estimated_latency_ms": 169378, "estimated_pass_rate": 0.6909, "estimated_quality": 0.8961, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.645, "max_latency_ms": 202764, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6909, "quality": 0.8961, "quality_target_met": true, "utility": 69.550714}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.645,\"max_latency_ms\":202764,\"minimum_quality\":0.8831},\"task\":{\"ambiguity\":0.5959,\"category\":\"coding\",\"complexity\":0.854,\"consequence\":0.5001,\"evidence_need\":0.5444,\"family\":\"migration_planning\",\"input_tokens\":6840,\"reversibility\":0.4962,\"risk\":0.427,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.169,\"id\":\"worker_a\",\"p95_latency_ms\":22205,\"pass_rate\":0.4678,\"provider_group\":\"provider_z\",\"quality\":0.6849,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.115,\"id\":\"worker_b\",\"p95_latency_ms\":64066,\"pass_rate\":0.7536,\"provider_group\":\"provider_x\",\"quality\":0.8066,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.862,\"id\":\"worker_c\",\"p95_latency_ms\":83107,\"pass_rate\":0.9891,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.146,\"estimated_latency_ms\":169378,\"estimated_pass_rate\":0.6909,\"estimated_quality\":0.8961,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.645,\"max_latency_ms\":202764,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5959, "category": "coding", "complexity": 0.854, "consequence": 0.5001, "evidence_need": 0.5444, "family": "migration_planning", "input_tokens": 6840, "request_id": "synthetic-migration_planning-000433", "reversibility": 0.4962, "risk": 0.427, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 1.169, "id": "worker_a", "p95_latency_ms": 22205, "pass_rate": 0.4678, "provider_group": "provider_z", "quality": 0.6849, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.115, "id": "worker_b", "p95_latency_ms": 64066, "pass_rate": 0.7536, "provider_group": "provider_x", "quality": 0.8066, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.862, "id": "worker_c", "p95_latency_ms": 83107, "pass_rate": 0.9891, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.012, "max_latency_ms": 233320, "minimum_quality": 0.8068}, "family": "migration_planning", "id": "synthetic-migration_planning-000595", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.362, "estimated_latency_ms": 145194, "estimated_pass_rate": 0.8838, "estimated_quality": 0.8684, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.012, "max_latency_ms": 233320, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8838, "quality": 0.8684, "quality_target_met": true, "utility": 67.277952}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.012,\"max_latency_ms\":233320,\"minimum_quality\":0.8068},\"task\":{\"ambiguity\":0.607,\"category\":\"coding\",\"complexity\":0.8459,\"consequence\":0.5938,\"evidence_need\":0.4752,\"family\":\"migration_planning\",\"input_tokens\":160,\"reversibility\":0.6269,\"risk\":0.4957,\"summary\":\"Review a staged migration for access control and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.143,\"id\":\"worker_a\",\"p95_latency_ms\":64410,\"pass_rate\":0.8636,\"provider_group\":\"provider_z\",\"quality\":0.8424,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.219,\"id\":\"worker_b\",\"p95_latency_ms\":80784,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9213,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.036,\"id\":\"worker_c\",\"p95_latency_ms\":31094,\"pass_rate\":0.4949,\"provider_group\":\"provider_x\",\"quality\":0.7634,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.362,\"estimated_latency_ms\":145194,\"estimated_pass_rate\":0.8838,\"estimated_quality\":0.8684,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.012,\"max_latency_ms\":233320,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.607, "category": "coding", "complexity": 0.8459, "consequence": 0.5938, "evidence_need": 0.4752, "family": "migration_planning", "input_tokens": 160, "request_id": "synthetic-migration_planning-000595", "reversibility": 0.6269, "risk": 0.4957, "summary": "Review a staged migration for access control and identify irreversible steps."}, "workers": [{"credits_per_call": 3.143, "id": "worker_a", "p95_latency_ms": 64410, "pass_rate": 0.8636, "provider_group": "provider_z", "quality": 0.8424, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.219, "id": "worker_b", "p95_latency_ms": 80784, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9213, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.036, "id": "worker_c", "p95_latency_ms": 31094, "pass_rate": 0.4949, "provider_group": "provider_x", "quality": 0.7634, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.637, "max_latency_ms": 34917, "minimum_quality": 0.6717}, "family": "migration_planning", "id": "synthetic-migration_planning-000046", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.053, "estimated_latency_ms": 28772, "estimated_pass_rate": 0.3942, "estimated_quality": 0.6277, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.637, "max_latency_ms": 34917, "policy": "lean", "reason_codes": ["best_feasible_below_target", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3942, "quality": 0.6277, "quality_target_met": false, "utility": 39.632728}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.637,\"max_latency_ms\":34917,\"minimum_quality\":0.6717},\"task\":{\"ambiguity\":0.4067,\"category\":\"coding\",\"complexity\":0.615,\"consequence\":0.4031,\"evidence_need\":0.4967,\"family\":\"migration_planning\",\"input_tokens\":4760,\"reversibility\":0.3369,\"risk\":0.5815,\"summary\":\"Plan a zero-downtime migration of background jobs with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":10.876,\"id\":\"worker_a\",\"p95_latency_ms\":88017,\"pass_rate\":0.9801,\"provider_group\":\"provider_x\",\"quality\":0.9768,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.053,\"id\":\"worker_b\",\"p95_latency_ms\":28772,\"pass_rate\":0.5418,\"provider_group\":\"provider_y\",\"quality\":0.7281,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.411,\"id\":\"worker_c\",\"p95_latency_ms\":60573,\"pass_rate\":0.724,\"provider_group\":\"provider_z\",\"quality\":0.8988,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.053,\"estimated_latency_ms\":28772,\"estimated_pass_rate\":0.3942,\"estimated_quality\":0.6277,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.637,\"max_latency_ms\":34917,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4067, "category": "coding", "complexity": 0.615, "consequence": 0.4031, "evidence_need": 0.4967, "family": "migration_planning", "input_tokens": 4760, "request_id": "synthetic-migration_planning-000046", "reversibility": 0.3369, "risk": 0.5815, "summary": "Plan a zero-downtime migration of background jobs with rollback checkpoints."}, "workers": [{"credits_per_call": 10.876, "id": "worker_a", "p95_latency_ms": 88017, "pass_rate": 0.9801, "provider_group": "provider_x", "quality": 0.9768, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.053, "id": "worker_b", "p95_latency_ms": 28772, "pass_rate": 0.5418, "provider_group": "provider_y", "quality": 0.7281, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.411, "id": "worker_c", "p95_latency_ms": 60573, "pass_rate": 0.724, "provider_group": "provider_z", "quality": 0.8988, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.571, "max_latency_ms": 58490, "minimum_quality": 0.7683}, "family": "migration_planning", "id": "synthetic-migration_planning-000115", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 3.454, "estimated_latency_ms": 48939, "estimated_pass_rate": 0.6562, "estimated_quality": 0.7432, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.571, "max_latency_ms": 58490, "policy": "lean", "reason_codes": ["best_feasible_below_target", "credit_constrained", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 2, "pass_rate": 0.6562, "quality": 0.7432, "quality_target_met": false, "utility": 49.580965}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.571,\"max_latency_ms\":58490,\"minimum_quality\":0.7683},\"task\":{\"ambiguity\":0.3301,\"category\":\"coding\",\"complexity\":0.7081,\"consequence\":0.6254,\"evidence_need\":0.8051,\"family\":\"migration_planning\",\"input_tokens\":5080,\"reversibility\":0.5143,\"risk\":0.4294,\"summary\":\"Design compatibility testing for a webhook delivery schema migration.\"},\"workers\":[{\"credits_per_call\":3.454,\"id\":\"worker_a\",\"p95_latency_ms\":48939,\"pass_rate\":0.7962,\"provider_group\":\"provider_x\",\"quality\":0.8443,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.171,\"id\":\"worker_b\",\"p95_latency_ms\":31387,\"pass_rate\":0.4511,\"provider_group\":\"provider_y\",\"quality\":0.7322,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.535,\"id\":\"worker_c\",\"p95_latency_ms\":81942,\"pass_rate\":0.9362,\"provider_group\":\"provider_z\",\"quality\":0.9498,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":3.454,\"estimated_latency_ms\":48939,\"estimated_pass_rate\":0.6562,\"estimated_quality\":0.7432,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.571,\"max_latency_ms\":58490,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3301, "category": "coding", "complexity": 0.7081, "consequence": 0.6254, "evidence_need": 0.8051, "family": "migration_planning", "input_tokens": 5080, "request_id": "synthetic-migration_planning-000115", "reversibility": 0.5143, "risk": 0.4294, "summary": "Design compatibility testing for a webhook delivery schema migration."}, "workers": [{"credits_per_call": 3.454, "id": "worker_a", "p95_latency_ms": 48939, "pass_rate": 0.7962, "provider_group": "provider_x", "quality": 0.8443, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.171, "id": "worker_b", "p95_latency_ms": 31387, "pass_rate": 0.4511, "provider_group": "provider_y", "quality": 0.7322, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.535, "id": "worker_c", "p95_latency_ms": 81942, "pass_rate": 0.9362, "provider_group": "provider_z", "quality": 0.9498, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.221, "max_latency_ms": 38632, "minimum_quality": 0.6541}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000036", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.891, "estimated_latency_ms": 25107, "estimated_pass_rate": 0.4061, "estimated_quality": 0.5673, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.221, "max_latency_ms": 38632, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4061, "quality": 0.5673, "quality_target_met": false, "utility": 35.134548}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.221,\"max_latency_ms\":38632,\"minimum_quality\":0.6541},\"task\":{\"ambiguity\":0.7224,\"category\":\"review\",\"complexity\":0.539,\"consequence\":0.6612,\"evidence_need\":0.5864,\"family\":\"policy_compliance\",\"input_tokens\":6080,\"reversibility\":0.3161,\"risk\":0.7473,\"summary\":\"Identify evidence required before approving private credentials for production.\"},\"workers\":[{\"credits_per_call\":0.891,\"id\":\"worker_a\",\"p95_latency_ms\":25107,\"pass_rate\":0.5824,\"provider_group\":\"provider_x\",\"quality\":0.6871,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.402,\"id\":\"worker_b\",\"p95_latency_ms\":72272,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.953,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.606,\"id\":\"worker_c\",\"p95_latency_ms\":55694,\"pass_rate\":0.7845,\"provider_group\":\"provider_z\",\"quality\":0.8225,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.891,\"estimated_latency_ms\":25107,\"estimated_pass_rate\":0.4061,\"estimated_quality\":0.5673,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.221,\"max_latency_ms\":38632,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7224, "category": "review", "complexity": 0.539, "consequence": 0.6612, "evidence_need": 0.5864, "family": "policy_compliance", "input_tokens": 6080, "request_id": "synthetic-policy_compliance-000036", "reversibility": 0.3161, "risk": 0.7473, "summary": "Identify evidence required before approving private credentials for production."}, "workers": [{"credits_per_call": 0.891, "id": "worker_a", "p95_latency_ms": 25107, "pass_rate": 0.5824, "provider_group": "provider_x", "quality": 0.6871, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.402, "id": "worker_b", "p95_latency_ms": 72272, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.953, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.606, "id": "worker_c", "p95_latency_ms": 55694, "pass_rate": 0.7845, "provider_group": "provider_z", "quality": 0.8225, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.017, "max_latency_ms": 169165, "minimum_quality": 0.8129}, "family": "migration_planning", "id": "synthetic-migration_planning-000877", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.187, "estimated_latency_ms": 153857, "estimated_pass_rate": 0.6958, "estimated_quality": 0.8845, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.017, "max_latency_ms": 169165, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6958, "quality": 0.8845, "quality_target_met": true, "utility": 73.106217}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.017,\"max_latency_ms\":169165,\"minimum_quality\":0.8129},\"task\":{\"ambiguity\":0.5266,\"category\":\"coding\",\"complexity\":0.9476,\"consequence\":0.6003,\"evidence_need\":0.701,\"family\":\"migration_planning\",\"input_tokens\":3880,\"reversibility\":0.6163,\"risk\":0.6197,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":10.272,\"id\":\"worker_a\",\"p95_latency_ms\":72058,\"pass_rate\":0.907,\"provider_group\":\"provider_y\",\"quality\":0.9431,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.864,\"id\":\"worker_b\",\"p95_latency_ms\":30244,\"pass_rate\":0.5089,\"provider_group\":\"provider_x\",\"quality\":0.7612,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.051,\"id\":\"worker_c\",\"p95_latency_ms\":51555,\"pass_rate\":0.8195,\"provider_group\":\"provider_z\",\"quality\":0.794,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.187,\"estimated_latency_ms\":153857,\"estimated_pass_rate\":0.6958,\"estimated_quality\":0.8845,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.017,\"max_latency_ms\":169165,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5266, "category": "coding", "complexity": 0.9476, "consequence": 0.6003, "evidence_need": 0.701, "family": "migration_planning", "input_tokens": 3880, "request_id": "synthetic-migration_planning-000877", "reversibility": 0.6163, "risk": 0.6197, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 10.272, "id": "worker_a", "p95_latency_ms": 72058, "pass_rate": 0.907, "provider_group": "provider_y", "quality": 0.9431, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.864, "id": "worker_b", "p95_latency_ms": 30244, "pass_rate": 0.5089, "provider_group": "provider_x", "quality": 0.7612, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.051, "id": "worker_c", "p95_latency_ms": 51555, "pass_rate": 0.8195, "provider_group": "provider_z", "quality": 0.794, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.047, "max_latency_ms": 44791, "minimum_quality": 0.6208}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000077", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.989, "estimated_latency_ms": 31175, "estimated_pass_rate": 0.2647, "estimated_quality": 0.6098, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.047, "max_latency_ms": 44791, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2647, "quality": 0.6098, "quality_target_met": false, "utility": 36.919631}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.047,\"max_latency_ms\":44791,\"minimum_quality\":0.6208},\"task\":{\"ambiguity\":0.6931,\"category\":\"analysis\",\"complexity\":0.4954,\"consequence\":0.4292,\"evidence_need\":0.3188,\"family\":\"performance_tuning\",\"input_tokens\":7680,\"reversibility\":0.8795,\"risk\":0.2724,\"summary\":\"Choose between throughput and tail-latency changes for background jobs.\"},\"workers\":[{\"credits_per_call\":12.815,\"id\":\"worker_a\",\"p95_latency_ms\":63688,\"pass_rate\":0.9612,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.989,\"id\":\"worker_b\",\"p95_latency_ms\":31175,\"pass_rate\":0.3953,\"provider_group\":\"provider_y\",\"quality\":0.6856,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.813,\"id\":\"worker_c\",\"p95_latency_ms\":48229,\"pass_rate\":0.9273,\"provider_group\":\"provider_z\",\"quality\":0.8639,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.989,\"estimated_latency_ms\":31175,\"estimated_pass_rate\":0.2647,\"estimated_quality\":0.6098,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.047,\"max_latency_ms\":44791,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6931, "category": "analysis", "complexity": 0.4954, "consequence": 0.4292, "evidence_need": 0.3188, "family": "performance_tuning", "input_tokens": 7680, "request_id": "synthetic-performance_tuning-000077", "reversibility": 0.8795, "risk": 0.2724, "summary": "Choose between throughput and tail-latency changes for background jobs."}, "workers": [{"credits_per_call": 12.815, "id": "worker_a", "p95_latency_ms": 63688, "pass_rate": 0.9612, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.989, "id": "worker_b", "p95_latency_ms": 31175, "pass_rate": 0.3953, "provider_group": "provider_y", "quality": 0.6856, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.813, "id": "worker_c", "p95_latency_ms": 48229, "pass_rate": 0.9273, "provider_group": "provider_z", "quality": 0.8639, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.335, "max_latency_ms": 200523, "minimum_quality": 0.8528}, "family": "migration_planning", "id": "synthetic-migration_planning-000964", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.314, "estimated_latency_ms": 159078, "estimated_pass_rate": 0.7634, "estimated_quality": 0.9183, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.335, "max_latency_ms": 200523, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7634, "quality": 0.9183, "quality_target_met": true, "utility": 71.579435}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.335,\"max_latency_ms\":200523,\"minimum_quality\":0.8528},\"task\":{\"ambiguity\":0.3259,\"category\":\"coding\",\"complexity\":0.742,\"consequence\":0.7552,\"evidence_need\":0.4487,\"family\":\"migration_planning\",\"input_tokens\":5760,\"reversibility\":0.4713,\"risk\":0.5049,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.989,\"id\":\"worker_a\",\"p95_latency_ms\":26155,\"pass_rate\":0.5904,\"provider_group\":\"provider_x\",\"quality\":0.7901,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.13,\"id\":\"worker_b\",\"p95_latency_ms\":88402,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.195,\"id\":\"worker_c\",\"p95_latency_ms\":44521,\"pass_rate\":0.7568,\"provider_group\":\"provider_z\",\"quality\":0.8079,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.314,\"estimated_latency_ms\":159078,\"estimated_pass_rate\":0.7634,\"estimated_quality\":0.9183,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.335,\"max_latency_ms\":200523,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3259, "category": "coding", "complexity": 0.742, "consequence": 0.7552, "evidence_need": 0.4487, "family": "migration_planning", "input_tokens": 5760, "request_id": "synthetic-migration_planning-000964", "reversibility": 0.4713, "risk": 0.5049, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 0.989, "id": "worker_a", "p95_latency_ms": 26155, "pass_rate": 0.5904, "provider_group": "provider_x", "quality": 0.7901, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.13, "id": "worker_b", "p95_latency_ms": 88402, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.195, "id": "worker_c", "p95_latency_ms": 44521, "pass_rate": 0.7568, "provider_group": "provider_z", "quality": 0.8079, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.714, "max_latency_ms": 121940, "minimum_quality": 0.8038}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000057", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.001, "estimated_latency_ms": 82958, "estimated_pass_rate": 0.555, "estimated_quality": 0.8068, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.714, "max_latency_ms": 121940, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.555, "quality": 0.8068, "quality_target_met": true, "utility": 60.229947}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.714,\"max_latency_ms\":121940,\"minimum_quality\":0.8038},\"task\":{\"ambiguity\":0.4377,\"category\":\"review\",\"complexity\":0.7502,\"consequence\":0.5101,\"evidence_need\":0.8096,\"family\":\"policy_compliance\",\"input_tokens\":1640,\"reversibility\":0.6318,\"risk\":0.7518,\"summary\":\"Identify evidence required before approving database failover for production.\"},\"workers\":[{\"credits_per_call\":15.181,\"id\":\"worker_a\",\"p95_latency_ms\":58825,\"pass_rate\":0.9088,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.16,\"id\":\"worker_b\",\"p95_latency_ms\":64578,\"pass_rate\":0.7183,\"provider_group\":\"provider_x\",\"quality\":0.8588,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.841,\"id\":\"worker_c\",\"p95_latency_ms\":18380,\"pass_rate\":0.4626,\"provider_group\":\"provider_z\",\"quality\":0.7753,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.001,\"estimated_latency_ms\":82958,\"estimated_pass_rate\":0.555,\"estimated_quality\":0.8068,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.714,\"max_latency_ms\":121940,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4377, "category": "review", "complexity": 0.7502, "consequence": 0.5101, "evidence_need": 0.8096, "family": "policy_compliance", "input_tokens": 1640, "request_id": "synthetic-policy_compliance-000057", "reversibility": 0.6318, "risk": 0.7518, "summary": "Identify evidence required before approving database failover for production."}, "workers": [{"credits_per_call": 15.181, "id": "worker_a", "p95_latency_ms": 58825, "pass_rate": 0.9088, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.16, "id": "worker_b", "p95_latency_ms": 64578, "pass_rate": 0.7183, "provider_group": "provider_x", "quality": 0.8588, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.841, "id": "worker_c", "p95_latency_ms": 18380, "pass_rate": 0.4626, "provider_group": "provider_z", "quality": 0.7753, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.051, "max_latency_ms": 207876, "minimum_quality": 0.8644}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000066", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.877, "estimated_latency_ms": 156483, "estimated_pass_rate": 0.739, "estimated_quality": 0.9144, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.051, "max_latency_ms": 207876, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.739, "quality": 0.9144, "quality_target_met": true, "utility": 76.437771}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.051,\"max_latency_ms\":207876,\"minimum_quality\":0.8644},\"task\":{\"ambiguity\":0.7477,\"category\":\"review\",\"complexity\":0.5284,\"consequence\":0.5564,\"evidence_need\":0.8095,\"family\":\"policy_compliance\",\"input_tokens\":4280,\"reversibility\":0.5155,\"risk\":0.7826,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.919,\"id\":\"worker_a\",\"p95_latency_ms\":67703,\"pass_rate\":0.7739,\"provider_group\":\"provider_z\",\"quality\":0.8205,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.131,\"id\":\"worker_b\",\"p95_latency_ms\":29839,\"pass_rate\":0.4875,\"provider_group\":\"provider_y\",\"quality\":0.7095,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.827,\"id\":\"worker_c\",\"p95_latency_ms\":58941,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.877,\"estimated_latency_ms\":156483,\"estimated_pass_rate\":0.739,\"estimated_quality\":0.9144,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.051,\"max_latency_ms\":207876,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7477, "category": "review", "complexity": 0.5284, "consequence": 0.5564, "evidence_need": 0.8095, "family": "policy_compliance", "input_tokens": 4280, "request_id": "synthetic-policy_compliance-000066", "reversibility": 0.5155, "risk": 0.7826, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 2.919, "id": "worker_a", "p95_latency_ms": 67703, "pass_rate": 0.7739, "provider_group": "provider_z", "quality": 0.8205, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.131, "id": "worker_b", "p95_latency_ms": 29839, "pass_rate": 0.4875, "provider_group": "provider_y", "quality": 0.7095, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.827, "id": "worker_c", "p95_latency_ms": 58941, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.355, "max_latency_ms": 212357, "minimum_quality": 0.8888}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001092", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.208, "estimated_latency_ms": 88494, "estimated_pass_rate": 0.6672, "estimated_quality": 0.8318, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.355, "max_latency_ms": 212357, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6672, "quality": 0.8318, "quality_target_met": false, "utility": 56.029679}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.355,\"max_latency_ms\":212357,\"minimum_quality\":0.8888},\"task\":{\"ambiguity\":0.4187,\"category\":\"review\",\"complexity\":0.7459,\"consequence\":0.8028,\"evidence_need\":0.6275,\"family\":\"policy_compliance\",\"input_tokens\":3240,\"reversibility\":0.3394,\"risk\":0.568,\"summary\":\"Audit a workflow for streaming inference against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.122,\"id\":\"worker_a\",\"p95_latency_ms\":69523,\"pass_rate\":0.7518,\"provider_group\":\"provider_z\",\"quality\":0.9181,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.086,\"id\":\"worker_b\",\"p95_latency_ms\":18971,\"pass_rate\":0.4911,\"provider_group\":\"provider_x\",\"quality\":0.737,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.953,\"id\":\"worker_c\",\"p95_latency_ms\":89594,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9153,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.208,\"estimated_latency_ms\":88494,\"estimated_pass_rate\":0.6672,\"estimated_quality\":0.8318,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.355,\"max_latency_ms\":212357,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4187, "category": "review", "complexity": 0.7459, "consequence": 0.8028, "evidence_need": 0.6275, "family": "policy_compliance", "input_tokens": 3240, "request_id": "synthetic-policy_compliance-001092", "reversibility": 0.3394, "risk": 0.568, "summary": "Audit a workflow for streaming inference against explicit access requirements."}, "workers": [{"credits_per_call": 4.122, "id": "worker_a", "p95_latency_ms": 69523, "pass_rate": 0.7518, "provider_group": "provider_z", "quality": 0.9181, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.086, "id": "worker_b", "p95_latency_ms": 18971, "pass_rate": 0.4911, "provider_group": "provider_x", "quality": 0.737, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.953, "id": "worker_c", "p95_latency_ms": 89594, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9153, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.488, "max_latency_ms": 202524, "minimum_quality": 0.9444}, "family": "migration_planning", "id": "synthetic-migration_planning-000721", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.305, "estimated_latency_ms": 185748, "estimated_pass_rate": 0.7436, "estimated_quality": 0.9103, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 22.488, "max_latency_ms": 202524, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7436, "quality": 0.9103, "quality_target_met": false, "utility": 70.250241}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.488,\"max_latency_ms\":202524,\"minimum_quality\":0.9444},\"task\":{\"ambiguity\":0.4102,\"category\":\"coding\",\"complexity\":0.7677,\"consequence\":0.4576,\"evidence_need\":0.7488,\"family\":\"migration_planning\",\"input_tokens\":5680,\"reversibility\":0.2646,\"risk\":0.5629,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.653,\"id\":\"worker_a\",\"p95_latency_ms\":71201,\"pass_rate\":0.8202,\"provider_group\":\"provider_z\",\"quality\":0.8453,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.533,\"id\":\"worker_b\",\"p95_latency_ms\":86557,\"pass_rate\":0.9145,\"provider_group\":\"provider_y\",\"quality\":0.958,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.119,\"id\":\"worker_c\",\"p95_latency_ms\":27990,\"pass_rate\":0.58,\"provider_group\":\"provider_x\",\"quality\":0.7925,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.305,\"estimated_latency_ms\":185748,\"estimated_pass_rate\":0.7436,\"estimated_quality\":0.9103,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.488,\"max_latency_ms\":202524,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4102, "category": "coding", "complexity": 0.7677, "consequence": 0.4576, "evidence_need": 0.7488, "family": "migration_planning", "input_tokens": 5680, "request_id": "synthetic-migration_planning-000721", "reversibility": 0.2646, "risk": 0.5629, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 3.653, "id": "worker_a", "p95_latency_ms": 71201, "pass_rate": 0.8202, "provider_group": "provider_z", "quality": 0.8453, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.533, "id": "worker_b", "p95_latency_ms": 86557, "pass_rate": 0.9145, "provider_group": "provider_y", "quality": 0.958, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.119, "id": "worker_c", "p95_latency_ms": 27990, "pass_rate": 0.58, "provider_group": "provider_x", "quality": 0.7925, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.23, "max_latency_ms": 31210, "minimum_quality": 0.5467}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000561", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.935, "estimated_latency_ms": 26128, "estimated_pass_rate": 0.4549, "estimated_quality": 0.5877, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.23, "max_latency_ms": 31210, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4549, "quality": 0.5877, "quality_target_met": true, "utility": 28.003231}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.23,\"max_latency_ms\":31210,\"minimum_quality\":0.5467},\"task\":{\"ambiguity\":0.6985,\"category\":\"review\",\"complexity\":0.5492,\"consequence\":0.8576,\"evidence_need\":0.746,\"family\":\"policy_compliance\",\"input_tokens\":7360,\"reversibility\":0.619,\"risk\":0.6161,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.21,\"id\":\"worker_a\",\"p95_latency_ms\":63837,\"pass_rate\":0.9642,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.935,\"id\":\"worker_b\",\"p95_latency_ms\":26128,\"pass_rate\":0.6199,\"provider_group\":\"provider_x\",\"quality\":0.6983,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.152,\"id\":\"worker_c\",\"p95_latency_ms\":53443,\"pass_rate\":0.847,\"provider_group\":\"provider_z\",\"quality\":0.7948,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.935,\"estimated_latency_ms\":26128,\"estimated_pass_rate\":0.4549,\"estimated_quality\":0.5877,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.23,\"max_latency_ms\":31210,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6985, "category": "review", "complexity": 0.5492, "consequence": 0.8576, "evidence_need": 0.746, "family": "policy_compliance", "input_tokens": 7360, "request_id": "synthetic-policy_compliance-000561", "reversibility": 0.619, "risk": 0.6161, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 15.21, "id": "worker_a", "p95_latency_ms": 63837, "pass_rate": 0.9642, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.935, "id": "worker_b", "p95_latency_ms": 26128, "pass_rate": 0.6199, "provider_group": "provider_x", "quality": 0.6983, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.152, "id": "worker_c", "p95_latency_ms": 53443, "pass_rate": 0.847, "provider_group": "provider_z", "quality": 0.7948, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.058, "max_latency_ms": 207745, "minimum_quality": 0.7048}, "family": "migration_planning", "id": "synthetic-migration_planning-001069", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.623, "estimated_latency_ms": 73833, "estimated_pass_rate": 0.7147, "estimated_quality": 0.7758, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.058, "max_latency_ms": 207745, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7147, "quality": 0.7758, "quality_target_met": true, "utility": 60.495946}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.058,\"max_latency_ms\":207745,\"minimum_quality\":0.7048},\"task\":{\"ambiguity\":0.5385,\"category\":\"coding\",\"complexity\":0.7023,\"consequence\":0.7841,\"evidence_need\":0.515,\"family\":\"migration_planning\",\"input_tokens\":6400,\"reversibility\":0.2652,\"risk\":0.6751,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.704,\"id\":\"worker_a\",\"p95_latency_ms\":49553,\"pass_rate\":0.8083,\"provider_group\":\"provider_y\",\"quality\":0.8447,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.919,\"id\":\"worker_b\",\"p95_latency_ms\":24280,\"pass_rate\":0.5641,\"provider_group\":\"provider_x\",\"quality\":0.7082,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.771,\"id\":\"worker_c\",\"p95_latency_ms\":68386,\"pass_rate\":0.9099,\"provider_group\":\"provider_z\",\"quality\":0.9402,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.623,\"estimated_latency_ms\":73833,\"estimated_pass_rate\":0.7147,\"estimated_quality\":0.7758,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.058,\"max_latency_ms\":207745,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5385, "category": "coding", "complexity": 0.7023, "consequence": 0.7841, "evidence_need": 0.515, "family": "migration_planning", "input_tokens": 6400, "request_id": "synthetic-migration_planning-001069", "reversibility": 0.2652, "risk": 0.6751, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 3.704, "id": "worker_a", "p95_latency_ms": 49553, "pass_rate": 0.8083, "provider_group": "provider_y", "quality": 0.8447, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.919, "id": "worker_b", "p95_latency_ms": 24280, "pass_rate": 0.5641, "provider_group": "provider_x", "quality": 0.7082, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.771, "id": "worker_c", "p95_latency_ms": 68386, "pass_rate": 0.9099, "provider_group": "provider_z", "quality": 0.9402, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.644, "max_latency_ms": 54930, "minimum_quality": 0.6926}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000884", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.074, "estimated_latency_ms": 27738, "estimated_pass_rate": 0.4309, "estimated_quality": 0.626, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.644, "max_latency_ms": 54930, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4309, "quality": 0.626, "quality_target_met": false, "utility": 40.809716}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.644,\"max_latency_ms\":54930,\"minimum_quality\":0.6926},\"task\":{\"ambiguity\":0.363,\"category\":\"analysis\",\"complexity\":0.5213,\"consequence\":0.4014,\"evidence_need\":0.4376,\"family\":\"performance_tuning\",\"input_tokens\":1840,\"reversibility\":0.796,\"risk\":0.3792,\"summary\":\"Review a proposed optimization for schema evolution for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.074,\"id\":\"worker_a\",\"p95_latency_ms\":27738,\"pass_rate\":0.5482,\"provider_group\":\"provider_y\",\"quality\":0.6919,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.51,\"id\":\"worker_b\",\"p95_latency_ms\":66381,\"pass_rate\":0.9293,\"provider_group\":\"provider_z\",\"quality\":0.9359,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.476,\"id\":\"worker_c\",\"p95_latency_ms\":78042,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.074,\"estimated_latency_ms\":27738,\"estimated_pass_rate\":0.4309,\"estimated_quality\":0.626,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.644,\"max_latency_ms\":54930,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.363, "category": "analysis", "complexity": 0.5213, "consequence": 0.4014, "evidence_need": 0.4376, "family": "performance_tuning", "input_tokens": 1840, "request_id": "synthetic-performance_tuning-000884", "reversibility": 0.796, "risk": 0.3792, "summary": "Review a proposed optimization for schema evolution for hidden regressions."}, "workers": [{"credits_per_call": 1.074, "id": "worker_a", "p95_latency_ms": 27738, "pass_rate": 0.5482, "provider_group": "provider_y", "quality": 0.6919, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.51, "id": "worker_b", "p95_latency_ms": 66381, "pass_rate": 0.9293, "provider_group": "provider_z", "quality": 0.9359, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.476, "id": "worker_c", "p95_latency_ms": 78042, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.276, "max_latency_ms": 195149, "minimum_quality": 0.7315}, "family": "migration_planning", "id": "synthetic-migration_planning-000892", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.858, "estimated_latency_ms": 91898, "estimated_pass_rate": 0.6277, "estimated_quality": 0.8413, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.276, "max_latency_ms": 195149, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6277, "quality": 0.8413, "quality_target_met": true, "utility": 61.982424}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.276,\"max_latency_ms\":195149,\"minimum_quality\":0.7315},\"task\":{\"ambiguity\":0.5509,\"category\":\"coding\",\"complexity\":0.5871,\"consequence\":0.7878,\"evidence_need\":0.7808,\"family\":\"migration_planning\",\"input_tokens\":1440,\"reversibility\":0.4236,\"risk\":0.6395,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":3.821,\"id\":\"worker_a\",\"p95_latency_ms\":65224,\"pass_rate\":0.823,\"provider_group\":\"provider_x\",\"quality\":0.9226,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.037,\"id\":\"worker_b\",\"p95_latency_ms\":26674,\"pass_rate\":0.4975,\"provider_group\":\"provider_z\",\"quality\":0.7495,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.261,\"id\":\"worker_c\",\"p95_latency_ms\":73035,\"pass_rate\":0.9835,\"provider_group\":\"provider_y\",\"quality\":0.9362,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.858,\"estimated_latency_ms\":91898,\"estimated_pass_rate\":0.6277,\"estimated_quality\":0.8413,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.276,\"max_latency_ms\":195149,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5509, "category": "coding", "complexity": 0.5871, "consequence": 0.7878, "evidence_need": 0.7808, "family": "migration_planning", "input_tokens": 1440, "request_id": "synthetic-migration_planning-000892", "reversibility": 0.4236, "risk": 0.6395, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 3.821, "id": "worker_a", "p95_latency_ms": 65224, "pass_rate": 0.823, "provider_group": "provider_x", "quality": 0.9226, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.037, "id": "worker_b", "p95_latency_ms": 26674, "pass_rate": 0.4975, "provider_group": "provider_z", "quality": 0.7495, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.261, "id": "worker_c", "p95_latency_ms": 73035, "pass_rate": 0.9835, "provider_group": "provider_y", "quality": 0.9362, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.927, "max_latency_ms": 22986, "minimum_quality": 0.725}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000006", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.832, "estimated_latency_ms": 18179, "estimated_pass_rate": 0.4239, "estimated_quality": 0.6841, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.927, "max_latency_ms": 22986, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4239, "quality": 0.6841, "quality_target_met": false, "utility": 43.864858}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.927,\"max_latency_ms\":22986,\"minimum_quality\":0.725},\"task\":{\"ambiguity\":0.4541,\"category\":\"review\",\"complexity\":0.5645,\"consequence\":0.7227,\"evidence_need\":0.719,\"family\":\"policy_compliance\",\"input_tokens\":400,\"reversibility\":0.3797,\"risk\":0.6309,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":3.371,\"id\":\"worker_a\",\"p95_latency_ms\":68118,\"pass_rate\":0.7422,\"provider_group\":\"provider_y\",\"quality\":0.9194,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.832,\"id\":\"worker_b\",\"p95_latency_ms\":18179,\"pass_rate\":0.5733,\"provider_group\":\"provider_x\",\"quality\":0.7821,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.484,\"id\":\"worker_c\",\"p95_latency_ms\":59661,\"pass_rate\":0.8933,\"provider_group\":\"provider_z\",\"quality\":0.9919,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.832,\"estimated_latency_ms\":18179,\"estimated_pass_rate\":0.4239,\"estimated_quality\":0.6841,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.927,\"max_latency_ms\":22986,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4541, "category": "review", "complexity": 0.5645, "consequence": 0.7227, "evidence_need": 0.719, "family": "policy_compliance", "input_tokens": 400, "request_id": "synthetic-policy_compliance-000006", "reversibility": 0.3797, "risk": 0.6309, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 3.371, "id": "worker_a", "p95_latency_ms": 68118, "pass_rate": 0.7422, "provider_group": "provider_y", "quality": 0.9194, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.832, "id": "worker_b", "p95_latency_ms": 18179, "pass_rate": 0.5733, "provider_group": "provider_x", "quality": 0.7821, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.484, "id": "worker_c", "p95_latency_ms": 59661, "pass_rate": 0.8933, "provider_group": "provider_z", "quality": 0.9919, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.907, "max_latency_ms": 187327, "minimum_quality": 0.8615}, "family": "migration_planning", "id": "synthetic-migration_planning-000874", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.438, "estimated_latency_ms": 182448, "estimated_pass_rate": 0.7706, "estimated_quality": 0.8798, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.907, "max_latency_ms": 187327, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7706, "quality": 0.8798, "quality_target_met": true, "utility": 69.759626}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.907,\"max_latency_ms\":187327,\"minimum_quality\":0.8615},\"task\":{\"ambiguity\":0.463,\"category\":\"coding\",\"complexity\":0.6482,\"consequence\":0.6458,\"evidence_need\":0.5641,\"family\":\"migration_planning\",\"input_tokens\":1520,\"reversibility\":0.5072,\"risk\":0.6228,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.885,\"id\":\"worker_a\",\"p95_latency_ms\":30693,\"pass_rate\":0.5642,\"provider_group\":\"provider_x\",\"quality\":0.7298,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.474,\"id\":\"worker_b\",\"p95_latency_ms\":73938,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9223,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.079,\"id\":\"worker_c\",\"p95_latency_ms\":77817,\"pass_rate\":0.7722,\"provider_group\":\"provider_z\",\"quality\":0.8703,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.438,\"estimated_latency_ms\":182448,\"estimated_pass_rate\":0.7706,\"estimated_quality\":0.8798,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.907,\"max_latency_ms\":187327,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.463, "category": "coding", "complexity": 0.6482, "consequence": 0.6458, "evidence_need": 0.5641, "family": "migration_planning", "input_tokens": 1520, "request_id": "synthetic-migration_planning-000874", "reversibility": 0.5072, "risk": 0.6228, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 0.885, "id": "worker_a", "p95_latency_ms": 30693, "pass_rate": 0.5642, "provider_group": "provider_x", "quality": 0.7298, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.474, "id": "worker_b", "p95_latency_ms": 73938, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9223, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.079, "id": "worker_c", "p95_latency_ms": 77817, "pass_rate": 0.7722, "provider_group": "provider_z", "quality": 0.8703, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.88, "max_latency_ms": 193388, "minimum_quality": 0.7968}, "family": "migration_planning", "id": "synthetic-migration_planning-000409", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.976, "estimated_latency_ms": 167716, "estimated_pass_rate": 0.7677, "estimated_quality": 0.9031, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.88, "max_latency_ms": 193388, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7677, "quality": 0.9031, "quality_target_met": true, "utility": 72.242223}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.88,\"max_latency_ms\":193388,\"minimum_quality\":0.7968},\"task\":{\"ambiguity\":0.5348,\"category\":\"coding\",\"complexity\":0.7989,\"consequence\":0.4068,\"evidence_need\":0.7049,\"family\":\"migration_planning\",\"input_tokens\":6080,\"reversibility\":0.3967,\"risk\":0.4679,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.413,\"id\":\"worker_a\",\"p95_latency_ms\":78783,\"pass_rate\":0.8766,\"provider_group\":\"provider_y\",\"quality\":0.8802,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.708,\"id\":\"worker_b\",\"p95_latency_ms\":62597,\"pass_rate\":0.9763,\"provider_group\":\"provider_x\",\"quality\":0.963,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.855,\"id\":\"worker_c\",\"p95_latency_ms\":26336,\"pass_rate\":0.5688,\"provider_group\":\"provider_z\",\"quality\":0.69,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.976,\"estimated_latency_ms\":167716,\"estimated_pass_rate\":0.7677,\"estimated_quality\":0.9031,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.88,\"max_latency_ms\":193388,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5348, "category": "coding", "complexity": 0.7989, "consequence": 0.4068, "evidence_need": 0.7049, "family": "migration_planning", "input_tokens": 6080, "request_id": "synthetic-migration_planning-000409", "reversibility": 0.3967, "risk": 0.4679, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 3.413, "id": "worker_a", "p95_latency_ms": 78783, "pass_rate": 0.8766, "provider_group": "provider_y", "quality": 0.8802, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.708, "id": "worker_b", "p95_latency_ms": 62597, "pass_rate": 0.9763, "provider_group": "provider_x", "quality": 0.963, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.855, "id": "worker_c", "p95_latency_ms": 26336, "pass_rate": 0.5688, "provider_group": "provider_z", "quality": 0.69, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.114, "max_latency_ms": 250514, "minimum_quality": 0.7898}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000798", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.63, "estimated_latency_ms": 173687, "estimated_pass_rate": 0.7536, "estimated_quality": 0.9006, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.114, "max_latency_ms": 250514, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7536, "quality": 0.9006, "quality_target_met": true, "utility": 74.329677}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.114,\"max_latency_ms\":250514,\"minimum_quality\":0.7898},\"task\":{\"ambiguity\":0.5093,\"category\":\"review\",\"complexity\":0.6292,\"consequence\":0.5778,\"evidence_need\":0.7398,\"family\":\"policy_compliance\",\"input_tokens\":1000,\"reversibility\":0.5841,\"risk\":0.6696,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":11.708,\"id\":\"worker_a\",\"p95_latency_ms\":82587,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9402,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.001,\"id\":\"worker_b\",\"p95_latency_ms\":62067,\"pass_rate\":0.7242,\"provider_group\":\"provider_y\",\"quality\":0.8587,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.921,\"id\":\"worker_c\",\"p95_latency_ms\":29033,\"pass_rate\":0.5441,\"provider_group\":\"provider_z\",\"quality\":0.7915,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.63,\"estimated_latency_ms\":173687,\"estimated_pass_rate\":0.7536,\"estimated_quality\":0.9006,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.114,\"max_latency_ms\":250514,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5093, "category": "review", "complexity": 0.6292, "consequence": 0.5778, "evidence_need": 0.7398, "family": "policy_compliance", "input_tokens": 1000, "request_id": "synthetic-policy_compliance-000798", "reversibility": 0.5841, "risk": 0.6696, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 11.708, "id": "worker_a", "p95_latency_ms": 82587, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9402, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.001, "id": "worker_b", "p95_latency_ms": 62067, "pass_rate": 0.7242, "provider_group": "provider_y", "quality": 0.8587, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.921, "id": "worker_c", "p95_latency_ms": 29033, "pass_rate": 0.5441, "provider_group": "provider_z", "quality": 0.7915, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.647, "max_latency_ms": 166546, "minimum_quality": 0.8118}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001191", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.801, "estimated_latency_ms": 152318, "estimated_pass_rate": 0.7379, "estimated_quality": 0.8979, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 15.647, "max_latency_ms": 166546, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7379, "quality": 0.8979, "quality_target_met": true, "utility": 74.564028}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.647,\"max_latency_ms\":166546,\"minimum_quality\":0.8118},\"task\":{\"ambiguity\":0.7733,\"category\":\"review\",\"complexity\":0.7228,\"consequence\":0.6055,\"evidence_need\":0.5967,\"family\":\"policy_compliance\",\"input_tokens\":400,\"reversibility\":0.6692,\"risk\":0.7534,\"summary\":\"Identify evidence required before approving database failover for production.\"},\"workers\":[{\"credits_per_call\":10.152,\"id\":\"worker_a\",\"p95_latency_ms\":80882,\"pass_rate\":0.9486,\"provider_group\":\"provider_y\",\"quality\":0.9564,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.816,\"id\":\"worker_b\",\"p95_latency_ms\":53792,\"pass_rate\":0.8496,\"provider_group\":\"provider_x\",\"quality\":0.8115,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.833,\"id\":\"worker_c\",\"p95_latency_ms\":17644,\"pass_rate\":0.5224,\"provider_group\":\"provider_z\",\"quality\":0.771,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.801,\"estimated_latency_ms\":152318,\"estimated_pass_rate\":0.7379,\"estimated_quality\":0.8979,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":15.647,\"max_latency_ms\":166546,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7733, "category": "review", "complexity": 0.7228, "consequence": 0.6055, "evidence_need": 0.5967, "family": "policy_compliance", "input_tokens": 400, "request_id": "synthetic-policy_compliance-001191", "reversibility": 0.6692, "risk": 0.7534, "summary": "Identify evidence required before approving database failover for production."}, "workers": [{"credits_per_call": 10.152, "id": "worker_a", "p95_latency_ms": 80882, "pass_rate": 0.9486, "provider_group": "provider_y", "quality": 0.9564, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.816, "id": "worker_b", "p95_latency_ms": 53792, "pass_rate": 0.8496, "provider_group": "provider_x", "quality": 0.8115, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.833, "id": "worker_c", "p95_latency_ms": 17644, "pass_rate": 0.5224, "provider_group": "provider_z", "quality": 0.771, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.609, "max_latency_ms": 35058, "minimum_quality": 0.5636}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000023", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.151, "estimated_latency_ms": 19298, "estimated_pass_rate": 0.4176, "estimated_quality": 0.6667, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.609, "max_latency_ms": 35058, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4176, "quality": 0.6667, "quality_target_met": true, "utility": 46.230659}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.609,\"max_latency_ms\":35058,\"minimum_quality\":0.5636},\"task\":{\"ambiguity\":0.4373,\"category\":\"analysis\",\"complexity\":0.49,\"consequence\":0.4455,\"evidence_need\":0.6982,\"family\":\"performance_tuning\",\"input_tokens\":1760,\"reversibility\":0.8322,\"risk\":0.2683,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.151,\"id\":\"worker_a\",\"p95_latency_ms\":19298,\"pass_rate\":0.528,\"provider_group\":\"provider_x\",\"quality\":0.724,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.209,\"id\":\"worker_b\",\"p95_latency_ms\":60095,\"pass_rate\":0.8164,\"provider_group\":\"provider_y\",\"quality\":0.9415,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.257,\"id\":\"worker_c\",\"p95_latency_ms\":69980,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.151,\"estimated_latency_ms\":19298,\"estimated_pass_rate\":0.4176,\"estimated_quality\":0.6667,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.609,\"max_latency_ms\":35058,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4373, "category": "analysis", "complexity": 0.49, "consequence": 0.4455, "evidence_need": 0.6982, "family": "performance_tuning", "input_tokens": 1760, "request_id": "synthetic-performance_tuning-000023", "reversibility": 0.8322, "risk": 0.2683, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 1.151, "id": "worker_a", "p95_latency_ms": 19298, "pass_rate": 0.528, "provider_group": "provider_x", "quality": 0.724, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.209, "id": "worker_b", "p95_latency_ms": 60095, "pass_rate": 0.8164, "provider_group": "provider_y", "quality": 0.9415, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.257, "id": "worker_c", "p95_latency_ms": 69980, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.639, "max_latency_ms": 32469, "minimum_quality": 0.5198}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000137", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.19, "estimated_latency_ms": 22604, "estimated_pass_rate": 0.3032, "estimated_quality": 0.5447, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.639, "max_latency_ms": 32469, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3032, "quality": 0.5447, "quality_target_met": true, "utility": 35.323137}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.639,\"max_latency_ms\":32469,\"minimum_quality\":0.5198},\"task\":{\"ambiguity\":0.6384,\"category\":\"analysis\",\"complexity\":0.5741,\"consequence\":0.6542,\"evidence_need\":0.562,\"family\":\"performance_tuning\",\"input_tokens\":1360,\"reversibility\":0.8,\"risk\":0.339,\"summary\":\"Diagnose latency variance in database failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.406,\"id\":\"worker_a\",\"p95_latency_ms\":75200,\"pass_rate\":0.9599,\"provider_group\":\"provider_y\",\"quality\":0.9751,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.646,\"id\":\"worker_b\",\"p95_latency_ms\":66459,\"pass_rate\":0.8666,\"provider_group\":\"provider_z\",\"quality\":0.9022,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.19,\"id\":\"worker_c\",\"p95_latency_ms\":22604,\"pass_rate\":0.444,\"provider_group\":\"provider_x\",\"quality\":0.6358,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.19,\"estimated_latency_ms\":22604,\"estimated_pass_rate\":0.3032,\"estimated_quality\":0.5447,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.639,\"max_latency_ms\":32469,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6384, "category": "analysis", "complexity": 0.5741, "consequence": 0.6542, "evidence_need": 0.562, "family": "performance_tuning", "input_tokens": 1360, "request_id": "synthetic-performance_tuning-000137", "reversibility": 0.8, "risk": 0.339, "summary": "Diagnose latency variance in database failover and rank the next experiments."}, "workers": [{"credits_per_call": 11.406, "id": "worker_a", "p95_latency_ms": 75200, "pass_rate": 0.9599, "provider_group": "provider_y", "quality": 0.9751, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.646, "id": "worker_b", "p95_latency_ms": 66459, "pass_rate": 0.8666, "provider_group": "provider_z", "quality": 0.9022, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.19, "id": "worker_c", "p95_latency_ms": 22604, "pass_rate": 0.444, "provider_group": "provider_x", "quality": 0.6358, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.768, "max_latency_ms": 159429, "minimum_quality": 0.8983}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000909", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.051, "estimated_latency_ms": 130123, "estimated_pass_rate": 0.7824, "estimated_quality": 0.9356, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.768, "max_latency_ms": 159429, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7824, "quality": 0.9356, "quality_target_met": true, "utility": 75.349577}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.768,\"max_latency_ms\":159429,\"minimum_quality\":0.8983},\"task\":{\"ambiguity\":0.767,\"category\":\"review\",\"complexity\":0.4742,\"consequence\":0.5063,\"evidence_need\":0.7182,\"family\":\"policy_compliance\",\"input_tokens\":5720,\"reversibility\":0.418,\"risk\":0.6973,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":11.745,\"id\":\"worker_a\",\"p95_latency_ms\":54686,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.044,\"id\":\"worker_b\",\"p95_latency_ms\":25534,\"pass_rate\":0.5432,\"provider_group\":\"provider_z\",\"quality\":0.7814,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.262,\"id\":\"worker_c\",\"p95_latency_ms\":49903,\"pass_rate\":0.834,\"provider_group\":\"provider_y\",\"quality\":0.8918,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.051,\"estimated_latency_ms\":130123,\"estimated_pass_rate\":0.7824,\"estimated_quality\":0.9356,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.768,\"max_latency_ms\":159429,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.767, "category": "review", "complexity": 0.4742, "consequence": 0.5063, "evidence_need": 0.7182, "family": "policy_compliance", "input_tokens": 5720, "request_id": "synthetic-policy_compliance-000909", "reversibility": 0.418, "risk": 0.6973, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 11.745, "id": "worker_a", "p95_latency_ms": 54686, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.044, "id": "worker_b", "p95_latency_ms": 25534, "pass_rate": 0.5432, "provider_group": "provider_z", "quality": 0.7814, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.262, "id": "worker_c", "p95_latency_ms": 49903, "pass_rate": 0.834, "provider_group": "provider_y", "quality": 0.8918, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.595, "max_latency_ms": 199626, "minimum_quality": 0.8105}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001268", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.266, "estimated_latency_ms": 146555, "estimated_pass_rate": 0.8907, "estimated_quality": 0.9363, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.595, "max_latency_ms": 199626, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8907, "quality": 0.9363, "quality_target_met": true, "utility": 69.500421}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.595,\"max_latency_ms\":199626,\"minimum_quality\":0.8105},\"task\":{\"ambiguity\":0.5208,\"category\":\"analysis\",\"complexity\":0.674,\"consequence\":0.6526,\"evidence_need\":0.4257,\"family\":\"performance_tuning\",\"input_tokens\":4560,\"reversibility\":0.6863,\"risk\":0.2349,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.371,\"id\":\"worker_a\",\"p95_latency_ms\":81599,\"pass_rate\":0.8096,\"provider_group\":\"provider_y\",\"quality\":0.8928,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.895,\"id\":\"worker_b\",\"p95_latency_ms\":64956,\"pass_rate\":0.9761,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.884,\"id\":\"worker_c\",\"p95_latency_ms\":30315,\"pass_rate\":0.474,\"provider_group\":\"provider_x\",\"quality\":0.6156,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.266,\"estimated_latency_ms\":146555,\"estimated_pass_rate\":0.8907,\"estimated_quality\":0.9363,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.595,\"max_latency_ms\":199626,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5208, "category": "analysis", "complexity": 0.674, "consequence": 0.6526, "evidence_need": 0.4257, "family": "performance_tuning", "input_tokens": 4560, "request_id": "synthetic-performance_tuning-001268", "reversibility": 0.6863, "risk": 0.2349, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 3.371, "id": "worker_a", "p95_latency_ms": 81599, "pass_rate": 0.8096, "provider_group": "provider_y", "quality": 0.8928, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.895, "id": "worker_b", "p95_latency_ms": 64956, "pass_rate": 0.9761, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.884, "id": "worker_c", "p95_latency_ms": 30315, "pass_rate": 0.474, "provider_group": "provider_x", "quality": 0.6156, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 4.806, "max_latency_ms": 125522, "minimum_quality": 0.8181}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000099", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.208, "estimated_latency_ms": 85165, "estimated_pass_rate": 0.5754, "estimated_quality": 0.8251, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.806, "max_latency_ms": 125522, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5754, "quality": 0.8251, "quality_target_met": true, "utility": 57.40757}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":4.806,\"max_latency_ms\":125522,\"minimum_quality\":0.8181},\"task\":{\"ambiguity\":0.6904,\"category\":\"review\",\"complexity\":0.7854,\"consequence\":0.6318,\"evidence_need\":0.8554,\"family\":\"policy_compliance\",\"input_tokens\":1960,\"reversibility\":0.5748,\"risk\":0.5219,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.153,\"id\":\"worker_a\",\"p95_latency_ms\":23446,\"pass_rate\":0.5193,\"provider_group\":\"provider_y\",\"quality\":0.8008,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.055,\"id\":\"worker_b\",\"p95_latency_ms\":61719,\"pass_rate\":0.7887,\"provider_group\":\"provider_x\",\"quality\":0.8536,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.398,\"id\":\"worker_c\",\"p95_latency_ms\":86825,\"pass_rate\":0.984,\"provider_group\":\"provider_z\",\"quality\":0.9736,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.208,\"estimated_latency_ms\":85165,\"estimated_pass_rate\":0.5754,\"estimated_quality\":0.8251,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.806,\"max_latency_ms\":125522,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6904, "category": "review", "complexity": 0.7854, "consequence": 0.6318, "evidence_need": 0.8554, "family": "policy_compliance", "input_tokens": 1960, "request_id": "synthetic-policy_compliance-000099", "reversibility": 0.5748, "risk": 0.5219, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 1.153, "id": "worker_a", "p95_latency_ms": 23446, "pass_rate": 0.5193, "provider_group": "provider_y", "quality": 0.8008, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.055, "id": "worker_b", "p95_latency_ms": 61719, "pass_rate": 0.7887, "provider_group": "provider_x", "quality": 0.8536, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.398, "id": "worker_c", "p95_latency_ms": 86825, "pass_rate": 0.984, "provider_group": "provider_z", "quality": 0.9736, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.768, "max_latency_ms": 199231, "minimum_quality": 0.8649}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000953", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.16, "estimated_latency_ms": 172833, "estimated_pass_rate": 0.7623, "estimated_quality": 0.9244, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.768, "max_latency_ms": 199231, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7623, "quality": 0.9244, "quality_target_met": true, "utility": 70.482803}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.768,\"max_latency_ms\":199231,\"minimum_quality\":0.8649},\"task\":{\"ambiguity\":0.4547,\"category\":\"analysis\",\"complexity\":0.7676,\"consequence\":0.3803,\"evidence_need\":0.7485,\"family\":\"performance_tuning\",\"input_tokens\":6040,\"reversibility\":0.622,\"risk\":0.3064,\"summary\":\"Review a proposed optimization for model routing for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.695,\"id\":\"worker_a\",\"p95_latency_ms\":72930,\"pass_rate\":0.9817,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.874,\"id\":\"worker_b\",\"p95_latency_ms\":17773,\"pass_rate\":0.5371,\"provider_group\":\"provider_z\",\"quality\":0.6627,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.591,\"id\":\"worker_c\",\"p95_latency_ms\":82130,\"pass_rate\":0.9414,\"provider_group\":\"provider_x\",\"quality\":0.8909,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.16,\"estimated_latency_ms\":172833,\"estimated_pass_rate\":0.7623,\"estimated_quality\":0.9244,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.768,\"max_latency_ms\":199231,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4547, "category": "analysis", "complexity": 0.7676, "consequence": 0.3803, "evidence_need": 0.7485, "family": "performance_tuning", "input_tokens": 6040, "request_id": "synthetic-performance_tuning-000953", "reversibility": 0.622, "risk": 0.3064, "summary": "Review a proposed optimization for model routing for hidden regressions."}, "workers": [{"credits_per_call": 11.695, "id": "worker_a", "p95_latency_ms": 72930, "pass_rate": 0.9817, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.874, "id": "worker_b", "p95_latency_ms": 17773, "pass_rate": 0.5371, "provider_group": "provider_z", "quality": 0.6627, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.591, "id": "worker_c", "p95_latency_ms": 82130, "pass_rate": 0.9414, "provider_group": "provider_x", "quality": 0.8909, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.605, "max_latency_ms": 206792, "minimum_quality": 0.9426}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000819", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.231, "estimated_latency_ms": 169597, "estimated_pass_rate": 0.7441, "estimated_quality": 0.9311, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.605, "max_latency_ms": 206792, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7441, "quality": 0.9311, "quality_target_met": false, "utility": 75.403302}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.605,\"max_latency_ms\":206792,\"minimum_quality\":0.9426},\"task\":{\"ambiguity\":0.5616,\"category\":\"review\",\"complexity\":0.6661,\"consequence\":0.7675,\"evidence_need\":0.7636,\"family\":\"policy_compliance\",\"input_tokens\":4800,\"reversibility\":0.3978,\"risk\":0.6038,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.25,\"id\":\"worker_a\",\"p95_latency_ms\":91477,\"pass_rate\":0.92,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.912,\"id\":\"worker_b\",\"p95_latency_ms\":27734,\"pass_rate\":0.5597,\"provider_group\":\"provider_z\",\"quality\":0.7951,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.069,\"id\":\"worker_c\",\"p95_latency_ms\":50386,\"pass_rate\":0.7831,\"provider_group\":\"provider_y\",\"quality\":0.9077,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.231,\"estimated_latency_ms\":169597,\"estimated_pass_rate\":0.7441,\"estimated_quality\":0.9311,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.605,\"max_latency_ms\":206792,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5616, "category": "review", "complexity": 0.6661, "consequence": 0.7675, "evidence_need": 0.7636, "family": "policy_compliance", "input_tokens": 4800, "request_id": "synthetic-policy_compliance-000819", "reversibility": 0.3978, "risk": 0.6038, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 10.25, "id": "worker_a", "p95_latency_ms": 91477, "pass_rate": 0.92, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.912, "id": "worker_b", "p95_latency_ms": 27734, "pass_rate": 0.5597, "provider_group": "provider_z", "quality": 0.7951, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.069, "id": "worker_c", "p95_latency_ms": 50386, "pass_rate": 0.7831, "provider_group": "provider_y", "quality": 0.9077, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.891, "max_latency_ms": 226934, "minimum_quality": 0.9316}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001200", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.19, "estimated_latency_ms": 156855, "estimated_pass_rate": 0.7082, "estimated_quality": 0.9092, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.891, "max_latency_ms": 226934, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7082, "quality": 0.9092, "quality_target_met": false, "utility": 73.795791}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.891,\"max_latency_ms\":226934,\"minimum_quality\":0.9316},\"task\":{\"ambiguity\":0.7808,\"category\":\"review\",\"complexity\":0.7033,\"consequence\":0.5231,\"evidence_need\":0.7533,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.6674,\"risk\":0.5468,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":3.342,\"id\":\"worker_a\",\"p95_latency_ms\":70959,\"pass_rate\":0.7671,\"provider_group\":\"provider_z\",\"quality\":0.822,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.97,\"id\":\"worker_b\",\"p95_latency_ms\":17676,\"pass_rate\":0.5409,\"provider_group\":\"provider_y\",\"quality\":0.7913,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.878,\"id\":\"worker_c\",\"p95_latency_ms\":68220,\"pass_rate\":0.921,\"provider_group\":\"provider_x\",\"quality\":0.9715,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.19,\"estimated_latency_ms\":156855,\"estimated_pass_rate\":0.7082,\"estimated_quality\":0.9092,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.891,\"max_latency_ms\":226934,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7808, "category": "review", "complexity": 0.7033, "consequence": 0.5231, "evidence_need": 0.7533, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-001200", "reversibility": 0.6674, "risk": 0.5468, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 3.342, "id": "worker_a", "p95_latency_ms": 70959, "pass_rate": 0.7671, "provider_group": "provider_z", "quality": 0.822, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.97, "id": "worker_b", "p95_latency_ms": 17676, "pass_rate": 0.5409, "provider_group": "provider_y", "quality": 0.7913, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.878, "id": "worker_c", "p95_latency_ms": 68220, "pass_rate": 0.921, "provider_group": "provider_x", "quality": 0.9715, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.359, "max_latency_ms": 213021, "minimum_quality": 0.8082}, "family": "migration_planning", "id": "synthetic-migration_planning-001135", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 10.906, "estimated_latency_ms": 107975, "estimated_pass_rate": 0.7835, "estimated_quality": 0.9113, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.359, "max_latency_ms": 213021, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7835, "quality": 0.9113, "quality_target_met": true, "utility": 64.766183}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.359,\"max_latency_ms\":213021,\"minimum_quality\":0.8082},\"task\":{\"ambiguity\":0.5791,\"category\":\"coding\",\"complexity\":0.5586,\"consequence\":0.8401,\"evidence_need\":0.8195,\"family\":\"migration_planning\",\"input_tokens\":3120,\"reversibility\":0.5866,\"risk\":0.5167,\"summary\":\"Plan a zero-downtime migration of rate limiting with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.964,\"id\":\"worker_a\",\"p95_latency_ms\":55498,\"pass_rate\":0.7495,\"provider_group\":\"provider_y\",\"quality\":0.8488,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.959,\"id\":\"worker_b\",\"p95_latency_ms\":27009,\"pass_rate\":0.5092,\"provider_group\":\"provider_x\",\"quality\":0.7384,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.947,\"id\":\"worker_c\",\"p95_latency_ms\":80966,\"pass_rate\":0.8853,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":10.906,\"estimated_latency_ms\":107975,\"estimated_pass_rate\":0.7835,\"estimated_quality\":0.9113,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.359,\"max_latency_ms\":213021,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5791, "category": "coding", "complexity": 0.5586, "consequence": 0.8401, "evidence_need": 0.8195, "family": "migration_planning", "input_tokens": 3120, "request_id": "synthetic-migration_planning-001135", "reversibility": 0.5866, "risk": 0.5167, "summary": "Plan a zero-downtime migration of rate limiting with rollback checkpoints."}, "workers": [{"credits_per_call": 3.964, "id": "worker_a", "p95_latency_ms": 55498, "pass_rate": 0.7495, "provider_group": "provider_y", "quality": 0.8488, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.959, "id": "worker_b", "p95_latency_ms": 27009, "pass_rate": 0.5092, "provider_group": "provider_x", "quality": 0.7384, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.947, "id": "worker_c", "p95_latency_ms": 80966, "pass_rate": 0.8853, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.678, "max_latency_ms": 129803, "minimum_quality": 0.8592}, "family": "migration_planning", "id": "synthetic-migration_planning-001138", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.889, "estimated_latency_ms": 126609, "estimated_pass_rate": 0.8833, "estimated_quality": 0.9278, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.678, "max_latency_ms": 129803, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8833, "quality": 0.9278, "quality_target_met": true, "utility": 70.406268}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.678,\"max_latency_ms\":129803,\"minimum_quality\":0.8592},\"task\":{\"ambiguity\":0.7452,\"category\":\"coding\",\"complexity\":0.6335,\"consequence\":0.7333,\"evidence_need\":0.7086,\"family\":\"migration_planning\",\"input_tokens\":7920,\"reversibility\":0.5977,\"risk\":0.6154,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":0.871,\"id\":\"worker_a\",\"p95_latency_ms\":30998,\"pass_rate\":0.5296,\"provider_group\":\"provider_x\",\"quality\":0.731,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.815,\"id\":\"worker_b\",\"p95_latency_ms\":77838,\"pass_rate\":0.9889,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.074,\"id\":\"worker_c\",\"p95_latency_ms\":48771,\"pass_rate\":0.8045,\"provider_group\":\"provider_y\",\"quality\":0.8546,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.889,\"estimated_latency_ms\":126609,\"estimated_pass_rate\":0.8833,\"estimated_quality\":0.9278,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.678,\"max_latency_ms\":129803,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7452, "category": "coding", "complexity": 0.6335, "consequence": 0.7333, "evidence_need": 0.7086, "family": "migration_planning", "input_tokens": 7920, "request_id": "synthetic-migration_planning-001138", "reversibility": 0.5977, "risk": 0.6154, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 0.871, "id": "worker_a", "p95_latency_ms": 30998, "pass_rate": 0.5296, "provider_group": "provider_x", "quality": 0.731, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.815, "id": "worker_b", "p95_latency_ms": 77838, "pass_rate": 0.9889, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.074, "id": "worker_c", "p95_latency_ms": 48771, "pass_rate": 0.8045, "provider_group": "provider_y", "quality": 0.8546, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.636, "max_latency_ms": 85534, "minimum_quality": 0.7146}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001002", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.914, "estimated_latency_ms": 75711, "estimated_pass_rate": 0.7404, "estimated_quality": 0.9009, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.636, "max_latency_ms": 85534, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 5, "pass_rate": 0.7404, "quality": 0.9009, "quality_target_met": true, "utility": 68.118127}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.636,\"max_latency_ms\":85534,\"minimum_quality\":0.7146},\"task\":{\"ambiguity\":0.4622,\"category\":\"review\",\"complexity\":0.5057,\"consequence\":0.8418,\"evidence_need\":0.7351,\"family\":\"policy_compliance\",\"input_tokens\":6640,\"reversibility\":0.3483,\"risk\":0.8223,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.076,\"id\":\"worker_a\",\"p95_latency_ms\":52265,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.26,\"id\":\"worker_b\",\"p95_latency_ms\":70977,\"pass_rate\":0.8001,\"provider_group\":\"provider_z\",\"quality\":0.7978,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.838,\"id\":\"worker_c\",\"p95_latency_ms\":23446,\"pass_rate\":0.4512,\"provider_group\":\"provider_x\",\"quality\":0.7024,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.914,\"estimated_latency_ms\":75711,\"estimated_pass_rate\":0.7404,\"estimated_quality\":0.9009,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.636,\"max_latency_ms\":85534,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4622, "category": "review", "complexity": 0.5057, "consequence": 0.8418, "evidence_need": 0.7351, "family": "policy_compliance", "input_tokens": 6640, "request_id": "synthetic-policy_compliance-001002", "reversibility": 0.3483, "risk": 0.8223, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 11.076, "id": "worker_a", "p95_latency_ms": 52265, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.26, "id": "worker_b", "p95_latency_ms": 70977, "pass_rate": 0.8001, "provider_group": "provider_z", "quality": 0.7978, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.838, "id": "worker_c", "p95_latency_ms": 23446, "pass_rate": 0.4512, "provider_group": "provider_x", "quality": 0.7024, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.758, "max_latency_ms": 124724, "minimum_quality": 0.8168}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000021", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.425, "estimated_latency_ms": 121264, "estimated_pass_rate": 0.7151, "estimated_quality": 0.9298, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.758, "max_latency_ms": 124724, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7151, "quality": 0.9298, "quality_target_met": true, "utility": 72.599916}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.758,\"max_latency_ms\":124724,\"minimum_quality\":0.8168},\"task\":{\"ambiguity\":0.3638,\"category\":\"review\",\"complexity\":0.6432,\"consequence\":0.7152,\"evidence_need\":0.7466,\"family\":\"policy_compliance\",\"input_tokens\":2920,\"reversibility\":0.4996,\"risk\":0.5909,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.084,\"id\":\"worker_a\",\"p95_latency_ms\":19194,\"pass_rate\":0.4753,\"provider_group\":\"provider_x\",\"quality\":0.7853,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.238,\"id\":\"worker_b\",\"p95_latency_ms\":44165,\"pass_rate\":0.8163,\"provider_group\":\"provider_y\",\"quality\":0.8366,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.103,\"id\":\"worker_c\",\"p95_latency_ms\":57905,\"pass_rate\":0.9287,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.425,\"estimated_latency_ms\":121264,\"estimated_pass_rate\":0.7151,\"estimated_quality\":0.9298,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.758,\"max_latency_ms\":124724,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3638, "category": "review", "complexity": 0.6432, "consequence": 0.7152, "evidence_need": 0.7466, "family": "policy_compliance", "input_tokens": 2920, "request_id": "synthetic-policy_compliance-000021", "reversibility": 0.4996, "risk": 0.5909, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 1.084, "id": "worker_a", "p95_latency_ms": 19194, "pass_rate": 0.4753, "provider_group": "provider_x", "quality": 0.7853, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.238, "id": "worker_b", "p95_latency_ms": 44165, "pass_rate": 0.8163, "provider_group": "provider_y", "quality": 0.8366, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 11.103, "id": "worker_c", "p95_latency_ms": 57905, "pass_rate": 0.9287, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.975, "max_latency_ms": 225073, "minimum_quality": 0.7855}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000156", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.488, "estimated_latency_ms": 171960, "estimated_pass_rate": 0.7505, "estimated_quality": 0.9092, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.975, "max_latency_ms": 225073, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7505, "quality": 0.9092, "quality_target_met": true, "utility": 74.587777}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.975,\"max_latency_ms\":225073,\"minimum_quality\":0.7855},\"task\":{\"ambiguity\":0.5273,\"category\":\"review\",\"complexity\":0.7093,\"consequence\":0.7528,\"evidence_need\":0.6266,\"family\":\"policy_compliance\",\"input_tokens\":1200,\"reversibility\":0.5512,\"risk\":0.5884,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":10.326,\"id\":\"worker_a\",\"p95_latency_ms\":74157,\"pass_rate\":0.9679,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.828,\"id\":\"worker_b\",\"p95_latency_ms\":31393,\"pass_rate\":0.5533,\"provider_group\":\"provider_y\",\"quality\":0.7056,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.334,\"id\":\"worker_c\",\"p95_latency_ms\":66410,\"pass_rate\":0.8262,\"provider_group\":\"provider_z\",\"quality\":0.8178,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.488,\"estimated_latency_ms\":171960,\"estimated_pass_rate\":0.7505,\"estimated_quality\":0.9092,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.975,\"max_latency_ms\":225073,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5273, "category": "review", "complexity": 0.7093, "consequence": 0.7528, "evidence_need": 0.6266, "family": "policy_compliance", "input_tokens": 1200, "request_id": "synthetic-policy_compliance-000156", "reversibility": 0.5512, "risk": 0.5884, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 10.326, "id": "worker_a", "p95_latency_ms": 74157, "pass_rate": 0.9679, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.828, "id": "worker_b", "p95_latency_ms": 31393, "pass_rate": 0.5533, "provider_group": "provider_y", "quality": 0.7056, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.334, "id": "worker_c", "p95_latency_ms": 66410, "pass_rate": 0.8262, "provider_group": "provider_z", "quality": 0.8178, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.668, "max_latency_ms": 204266, "minimum_quality": 0.9202}, "family": "migration_planning", "id": "synthetic-migration_planning-000472", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.44, "estimated_latency_ms": 152947, "estimated_pass_rate": 0.7151, "estimated_quality": 0.8958, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.668, "max_latency_ms": 204266, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7151, "quality": 0.8958, "quality_target_met": false, "utility": 68.912469}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.668,\"max_latency_ms\":204266,\"minimum_quality\":0.9202},\"task\":{\"ambiguity\":0.4611,\"category\":\"coding\",\"complexity\":0.8435,\"consequence\":0.7111,\"evidence_need\":0.5271,\"family\":\"migration_planning\",\"input_tokens\":6200,\"reversibility\":0.5609,\"risk\":0.3778,\"summary\":\"Review a staged migration for observability and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.159,\"id\":\"worker_a\",\"p95_latency_ms\":28936,\"pass_rate\":0.5926,\"provider_group\":\"provider_y\",\"quality\":0.7161,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.061,\"id\":\"worker_b\",\"p95_latency_ms\":69047,\"pass_rate\":0.7419,\"provider_group\":\"provider_z\",\"quality\":0.8133,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.22,\"id\":\"worker_c\",\"p95_latency_ms\":54964,\"pass_rate\":0.9064,\"provider_group\":\"provider_x\",\"quality\":0.9937,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.44,\"estimated_latency_ms\":152947,\"estimated_pass_rate\":0.7151,\"estimated_quality\":0.8958,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.668,\"max_latency_ms\":204266,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4611, "category": "coding", "complexity": 0.8435, "consequence": 0.7111, "evidence_need": 0.5271, "family": "migration_planning", "input_tokens": 6200, "request_id": "synthetic-migration_planning-000472", "reversibility": 0.5609, "risk": 0.3778, "summary": "Review a staged migration for observability and identify irreversible steps."}, "workers": [{"credits_per_call": 1.159, "id": "worker_a", "p95_latency_ms": 28936, "pass_rate": 0.5926, "provider_group": "provider_y", "quality": 0.7161, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.061, "id": "worker_b", "p95_latency_ms": 69047, "pass_rate": 0.7419, "provider_group": "provider_z", "quality": 0.8133, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.22, "id": "worker_c", "p95_latency_ms": 54964, "pass_rate": 0.9064, "provider_group": "provider_x", "quality": 0.9937, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.693, "max_latency_ms": 98358, "minimum_quality": 0.8545}, "family": "migration_planning", "id": "synthetic-migration_planning-001180", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.839, "estimated_latency_ms": 88851, "estimated_pass_rate": 0.6198, "estimated_quality": 0.7827, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.693, "max_latency_ms": 98358, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6198, "quality": 0.7827, "quality_target_met": false, "utility": 52.837844}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.693,\"max_latency_ms\":98358,\"minimum_quality\":0.8545},\"task\":{\"ambiguity\":0.5356,\"category\":\"coding\",\"complexity\":0.7813,\"consequence\":0.727,\"evidence_need\":0.4164,\"family\":\"migration_planning\",\"input_tokens\":160,\"reversibility\":0.4396,\"risk\":0.5217,\"summary\":\"Design compatibility testing for a prompt caching schema migration.\"},\"workers\":[{\"credits_per_call\":1.014,\"id\":\"worker_a\",\"p95_latency_ms\":29623,\"pass_rate\":0.5405,\"provider_group\":\"provider_y\",\"quality\":0.7577,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.825,\"id\":\"worker_b\",\"p95_latency_ms\":59228,\"pass_rate\":0.871,\"provider_group\":\"provider_x\",\"quality\":0.8503,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.579,\"id\":\"worker_c\",\"p95_latency_ms\":78578,\"pass_rate\":0.9391,\"provider_group\":\"provider_z\",\"quality\":0.9389,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.839,\"estimated_latency_ms\":88851,\"estimated_pass_rate\":0.6198,\"estimated_quality\":0.7827,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.693,\"max_latency_ms\":98358,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5356, "category": "coding", "complexity": 0.7813, "consequence": 0.727, "evidence_need": 0.4164, "family": "migration_planning", "input_tokens": 160, "request_id": "synthetic-migration_planning-001180", "reversibility": 0.4396, "risk": 0.5217, "summary": "Design compatibility testing for a prompt caching schema migration."}, "workers": [{"credits_per_call": 1.014, "id": "worker_a", "p95_latency_ms": 29623, "pass_rate": 0.5405, "provider_group": "provider_y", "quality": 0.7577, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.825, "id": "worker_b", "p95_latency_ms": 59228, "pass_rate": 0.871, "provider_group": "provider_x", "quality": 0.8503, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.579, "id": "worker_c", "p95_latency_ms": 78578, "pass_rate": 0.9391, "provider_group": "provider_z", "quality": 0.9389, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.674, "max_latency_ms": 152878, "minimum_quality": 0.8716}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000306", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.262, "estimated_latency_ms": 135064, "estimated_pass_rate": 0.7294, "estimated_quality": 0.8954, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.674, "max_latency_ms": 152878, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7294, "quality": 0.8954, "quality_target_met": true, "utility": 70.9917}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.674,\"max_latency_ms\":152878,\"minimum_quality\":0.8716},\"task\":{\"ambiguity\":0.7074,\"category\":\"review\",\"complexity\":0.5401,\"consequence\":0.819,\"evidence_need\":0.7124,\"family\":\"policy_compliance\",\"input_tokens\":880,\"reversibility\":0.3861,\"risk\":0.4966,\"summary\":\"Audit a workflow for model routing against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.854,\"id\":\"worker_a\",\"p95_latency_ms\":49534,\"pass_rate\":0.8029,\"provider_group\":\"provider_y\",\"quality\":0.8248,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.384,\"id\":\"worker_b\",\"p95_latency_ms\":61515,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.936,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.024,\"id\":\"worker_c\",\"p95_latency_ms\":24015,\"pass_rate\":0.4892,\"provider_group\":\"provider_x\",\"quality\":0.7458,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.262,\"estimated_latency_ms\":135064,\"estimated_pass_rate\":0.7294,\"estimated_quality\":0.8954,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.674,\"max_latency_ms\":152878,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7074, "category": "review", "complexity": 0.5401, "consequence": 0.819, "evidence_need": 0.7124, "family": "policy_compliance", "input_tokens": 880, "request_id": "synthetic-policy_compliance-000306", "reversibility": 0.3861, "risk": 0.4966, "summary": "Audit a workflow for model routing against explicit access requirements."}, "workers": [{"credits_per_call": 3.854, "id": "worker_a", "p95_latency_ms": 49534, "pass_rate": 0.8029, "provider_group": "provider_y", "quality": 0.8248, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.384, "id": "worker_b", "p95_latency_ms": 61515, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.936, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.024, "id": "worker_c", "p95_latency_ms": 24015, "pass_rate": 0.4892, "provider_group": "provider_x", "quality": 0.7458, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.801, "max_latency_ms": 148569, "minimum_quality": 0.7938}, "family": "migration_planning", "id": "synthetic-migration_planning-000325", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.155, "estimated_latency_ms": 73268, "estimated_pass_rate": 0.6115, "estimated_quality": 0.7878, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.801, "max_latency_ms": 148569, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6115, "quality": 0.7878, "quality_target_met": false, "utility": 62.479934}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.801,\"max_latency_ms\":148569,\"minimum_quality\":0.7938},\"task\":{\"ambiguity\":0.5311,\"category\":\"coding\",\"complexity\":0.9313,\"consequence\":0.7473,\"evidence_need\":0.8398,\"family\":\"migration_planning\",\"input_tokens\":360,\"reversibility\":0.2596,\"risk\":0.7494,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":10.941,\"id\":\"worker_a\",\"p95_latency_ms\":62661,\"pass_rate\":0.915,\"provider_group\":\"provider_x\",\"quality\":0.9679,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.254,\"id\":\"worker_b\",\"p95_latency_ms\":52651,\"pass_rate\":0.8482,\"provider_group\":\"provider_y\",\"quality\":0.8335,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.901,\"id\":\"worker_c\",\"p95_latency_ms\":20617,\"pass_rate\":0.5202,\"provider_group\":\"provider_z\",\"quality\":0.7053,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.155,\"estimated_latency_ms\":73268,\"estimated_pass_rate\":0.6115,\"estimated_quality\":0.7878,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.801,\"max_latency_ms\":148569,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5311, "category": "coding", "complexity": 0.9313, "consequence": 0.7473, "evidence_need": 0.8398, "family": "migration_planning", "input_tokens": 360, "request_id": "synthetic-migration_planning-000325", "reversibility": 0.2596, "risk": 0.7494, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 10.941, "id": "worker_a", "p95_latency_ms": 62661, "pass_rate": 0.915, "provider_group": "provider_x", "quality": 0.9679, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.254, "id": "worker_b", "p95_latency_ms": 52651, "pass_rate": 0.8482, "provider_group": "provider_y", "quality": 0.8335, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.901, "id": "worker_c", "p95_latency_ms": 20617, "pass_rate": 0.5202, "provider_group": "provider_z", "quality": 0.7053, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.6, "max_latency_ms": 29216, "minimum_quality": 0.5491}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000056", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.154, "estimated_latency_ms": 20665, "estimated_pass_rate": 0.4099, "estimated_quality": 0.6324, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.6, "max_latency_ms": 29216, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4099, "quality": 0.6324, "quality_target_met": true, "utility": 43.209681}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.6,\"max_latency_ms\":29216,\"minimum_quality\":0.5491},\"task\":{\"ambiguity\":0.5007,\"category\":\"analysis\",\"complexity\":0.7161,\"consequence\":0.6448,\"evidence_need\":0.5377,\"family\":\"performance_tuning\",\"input_tokens\":3720,\"reversibility\":0.5537,\"risk\":0.2458,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":14.444,\"id\":\"worker_a\",\"p95_latency_ms\":88490,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9808,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.154,\"id\":\"worker_b\",\"p95_latency_ms\":20665,\"pass_rate\":0.5489,\"provider_group\":\"provider_x\",\"quality\":0.7332,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.473,\"id\":\"worker_c\",\"p95_latency_ms\":77766,\"pass_rate\":0.8514,\"provider_group\":\"provider_z\",\"quality\":0.9546,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.154,\"estimated_latency_ms\":20665,\"estimated_pass_rate\":0.4099,\"estimated_quality\":0.6324,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.6,\"max_latency_ms\":29216,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5007, "category": "analysis", "complexity": 0.7161, "consequence": 0.6448, "evidence_need": 0.5377, "family": "performance_tuning", "input_tokens": 3720, "request_id": "synthetic-performance_tuning-000056", "reversibility": 0.5537, "risk": 0.2458, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 14.444, "id": "worker_a", "p95_latency_ms": 88490, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9808, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.154, "id": "worker_b", "p95_latency_ms": 20665, "pass_rate": 0.5489, "provider_group": "provider_x", "quality": 0.7332, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.473, "id": "worker_c", "p95_latency_ms": 77766, "pass_rate": 0.8514, "provider_group": "provider_z", "quality": 0.9546, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.428, "max_latency_ms": 45104, "minimum_quality": 0.7197}, "family": "migration_planning", "id": "synthetic-migration_planning-000445", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.107, "estimated_latency_ms": 27165, "estimated_pass_rate": 0.2869, "estimated_quality": 0.6373, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.428, "max_latency_ms": 45104, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2869, "quality": 0.6373, "quality_target_met": false, "utility": 36.748748}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.428,\"max_latency_ms\":45104,\"minimum_quality\":0.7197},\"task\":{\"ambiguity\":0.5744,\"category\":\"coding\",\"complexity\":0.8454,\"consequence\":0.452,\"evidence_need\":0.4891,\"family\":\"migration_planning\",\"input_tokens\":4440,\"reversibility\":0.5892,\"risk\":0.487,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.949,\"id\":\"worker_a\",\"p95_latency_ms\":67192,\"pass_rate\":0.7285,\"provider_group\":\"provider_x\",\"quality\":0.8207,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.107,\"id\":\"worker_b\",\"p95_latency_ms\":27165,\"pass_rate\":0.4656,\"provider_group\":\"provider_z\",\"quality\":0.7838,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.798,\"id\":\"worker_c\",\"p95_latency_ms\":53499,\"pass_rate\":0.9028,\"provider_group\":\"provider_y\",\"quality\":0.987,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.107,\"estimated_latency_ms\":27165,\"estimated_pass_rate\":0.2869,\"estimated_quality\":0.6373,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.428,\"max_latency_ms\":45104,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5744, "category": "coding", "complexity": 0.8454, "consequence": 0.452, "evidence_need": 0.4891, "family": "migration_planning", "input_tokens": 4440, "request_id": "synthetic-migration_planning-000445", "reversibility": 0.5892, "risk": 0.487, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 2.949, "id": "worker_a", "p95_latency_ms": 67192, "pass_rate": 0.7285, "provider_group": "provider_x", "quality": 0.8207, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.107, "id": "worker_b", "p95_latency_ms": 27165, "pass_rate": 0.4656, "provider_group": "provider_z", "quality": 0.7838, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.798, "id": "worker_c", "p95_latency_ms": 53499, "pass_rate": 0.9028, "provider_group": "provider_y", "quality": 0.987, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.576, "max_latency_ms": 219444, "minimum_quality": 0.9273}, "family": "migration_planning", "id": "synthetic-migration_planning-000352", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.15, "estimated_latency_ms": 164896, "estimated_pass_rate": 0.6898, "estimated_quality": 0.8991, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.576, "max_latency_ms": 219444, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6898, "quality": 0.8991, "quality_target_met": false, "utility": 72.97219}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.576,\"max_latency_ms\":219444,\"minimum_quality\":0.9273},\"task\":{\"ambiguity\":0.7157,\"category\":\"coding\",\"complexity\":0.9085,\"consequence\":0.6957,\"evidence_need\":0.4682,\"family\":\"migration_planning\",\"input_tokens\":2040,\"reversibility\":0.4004,\"risk\":0.5268,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.874,\"id\":\"worker_a\",\"p95_latency_ms\":31231,\"pass_rate\":0.4589,\"provider_group\":\"provider_x\",\"quality\":0.792,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.591,\"id\":\"worker_b\",\"p95_latency_ms\":46498,\"pass_rate\":0.8344,\"provider_group\":\"provider_y\",\"quality\":0.8337,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.685,\"id\":\"worker_c\",\"p95_latency_ms\":87167,\"pass_rate\":0.94,\"provider_group\":\"provider_z\",\"quality\":0.9643,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.15,\"estimated_latency_ms\":164896,\"estimated_pass_rate\":0.6898,\"estimated_quality\":0.8991,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.576,\"max_latency_ms\":219444,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7157, "category": "coding", "complexity": 0.9085, "consequence": 0.6957, "evidence_need": 0.4682, "family": "migration_planning", "input_tokens": 2040, "request_id": "synthetic-migration_planning-000352", "reversibility": 0.4004, "risk": 0.5268, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 0.874, "id": "worker_a", "p95_latency_ms": 31231, "pass_rate": 0.4589, "provider_group": "provider_x", "quality": 0.792, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.591, "id": "worker_b", "p95_latency_ms": 46498, "pass_rate": 0.8344, "provider_group": "provider_y", "quality": 0.8337, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 14.685, "id": "worker_c", "p95_latency_ms": 87167, "pass_rate": 0.94, "provider_group": "provider_z", "quality": 0.9643, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.307, "max_latency_ms": 33358, "minimum_quality": 0.6488}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000650", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.021, "estimated_latency_ms": 18381, "estimated_pass_rate": 0.2809, "estimated_quality": 0.5691, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.307, "max_latency_ms": 33358, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2809, "quality": 0.5691, "quality_target_met": false, "utility": 33.756727}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.307,\"max_latency_ms\":33358,\"minimum_quality\":0.6488},\"task\":{\"ambiguity\":0.4971,\"category\":\"analysis\",\"complexity\":0.8432,\"consequence\":0.2912,\"evidence_need\":0.3461,\"family\":\"performance_tuning\",\"input_tokens\":7000,\"reversibility\":0.8534,\"risk\":0.5009,\"summary\":\"Diagnose latency variance in database failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":14.524,\"id\":\"worker_a\",\"p95_latency_ms\":53820,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.021,\"id\":\"worker_b\",\"p95_latency_ms\":18381,\"pass_rate\":0.4547,\"provider_group\":\"provider_y\",\"quality\":0.711,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.339,\"id\":\"worker_c\",\"p95_latency_ms\":75152,\"pass_rate\":0.8196,\"provider_group\":\"provider_x\",\"quality\":0.9081,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.021,\"estimated_latency_ms\":18381,\"estimated_pass_rate\":0.2809,\"estimated_quality\":0.5691,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.307,\"max_latency_ms\":33358,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4971, "category": "analysis", "complexity": 0.8432, "consequence": 0.2912, "evidence_need": 0.3461, "family": "performance_tuning", "input_tokens": 7000, "request_id": "synthetic-performance_tuning-000650", "reversibility": 0.8534, "risk": 0.5009, "summary": "Diagnose latency variance in database failover and rank the next experiments."}, "workers": [{"credits_per_call": 14.524, "id": "worker_a", "p95_latency_ms": 53820, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.021, "id": "worker_b", "p95_latency_ms": 18381, "pass_rate": 0.4547, "provider_group": "provider_y", "quality": 0.711, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.339, "id": "worker_c", "p95_latency_ms": 75152, "pass_rate": 0.8196, "provider_group": "provider_x", "quality": 0.9081, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.357, "max_latency_ms": 30049, "minimum_quality": 0.7747}, "family": "migration_planning", "id": "synthetic-migration_planning-000154", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.965, "estimated_latency_ms": 24916, "estimated_pass_rate": 0.3783, "estimated_quality": 0.5107, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.357, "max_latency_ms": 30049, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3783, "quality": 0.5107, "quality_target_met": false, "utility": 1.197713}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.357,\"max_latency_ms\":30049,\"minimum_quality\":0.7747},\"task\":{\"ambiguity\":0.7341,\"category\":\"coding\",\"complexity\":0.9356,\"consequence\":0.6508,\"evidence_need\":0.8453,\"family\":\"migration_planning\",\"input_tokens\":1400,\"reversibility\":0.2986,\"risk\":0.7001,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.965,\"id\":\"worker_a\",\"p95_latency_ms\":24916,\"pass_rate\":0.5966,\"provider_group\":\"provider_z\",\"quality\":0.6996,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.178,\"id\":\"worker_b\",\"p95_latency_ms\":59288,\"pass_rate\":0.7558,\"provider_group\":\"provider_y\",\"quality\":0.8077,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.727,\"id\":\"worker_c\",\"p95_latency_ms\":61156,\"pass_rate\":0.9102,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.965,\"estimated_latency_ms\":24916,\"estimated_pass_rate\":0.3783,\"estimated_quality\":0.5107,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.357,\"max_latency_ms\":30049,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7341, "category": "coding", "complexity": 0.9356, "consequence": 0.6508, "evidence_need": 0.8453, "family": "migration_planning", "input_tokens": 1400, "request_id": "synthetic-migration_planning-000154", "reversibility": 0.2986, "risk": 0.7001, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 0.965, "id": "worker_a", "p95_latency_ms": 24916, "pass_rate": 0.5966, "provider_group": "provider_z", "quality": 0.6996, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.178, "id": "worker_b", "p95_latency_ms": 59288, "pass_rate": 0.7558, "provider_group": "provider_y", "quality": 0.8077, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.727, "id": "worker_c", "p95_latency_ms": 61156, "pass_rate": 0.9102, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.84, "max_latency_ms": 220568, "minimum_quality": 0.7975}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001232", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.047, "estimated_latency_ms": 176709, "estimated_pass_rate": 0.7296, "estimated_quality": 0.9153, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.84, "max_latency_ms": 220568, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7296, "quality": 0.9153, "quality_target_met": true, "utility": 69.31617}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.84,\"max_latency_ms\":220568,\"minimum_quality\":0.7975},\"task\":{\"ambiguity\":0.5229,\"category\":\"analysis\",\"complexity\":0.8358,\"consequence\":0.3701,\"evidence_need\":0.553,\"family\":\"performance_tuning\",\"input_tokens\":5920,\"reversibility\":0.6486,\"risk\":0.3307,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":14.183,\"id\":\"worker_a\",\"p95_latency_ms\":78175,\"pass_rate\":0.9904,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.098,\"id\":\"worker_b\",\"p95_latency_ms\":21126,\"pass_rate\":0.4796,\"provider_group\":\"provider_x\",\"quality\":0.699,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.766,\"id\":\"worker_c\",\"p95_latency_ms\":77408,\"pass_rate\":0.8719,\"provider_group\":\"provider_y\",\"quality\":0.8961,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.047,\"estimated_latency_ms\":176709,\"estimated_pass_rate\":0.7296,\"estimated_quality\":0.9153,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.84,\"max_latency_ms\":220568,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5229, "category": "analysis", "complexity": 0.8358, "consequence": 0.3701, "evidence_need": 0.553, "family": "performance_tuning", "input_tokens": 5920, "request_id": "synthetic-performance_tuning-001232", "reversibility": 0.6486, "risk": 0.3307, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 14.183, "id": "worker_a", "p95_latency_ms": 78175, "pass_rate": 0.9904, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.098, "id": "worker_b", "p95_latency_ms": 21126, "pass_rate": 0.4796, "provider_group": "provider_x", "quality": 0.699, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.766, "id": "worker_c", "p95_latency_ms": 77408, "pass_rate": 0.8719, "provider_group": "provider_y", "quality": 0.8961, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.797, "max_latency_ms": 50523, "minimum_quality": 0.747}, "family": "migration_planning", "id": "synthetic-migration_planning-000760", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 3.438, "estimated_latency_ms": 48366, "estimated_pass_rate": 0.6274, "estimated_quality": 0.7105, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.797, "max_latency_ms": 50523, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 2, "pass_rate": 0.6274, "quality": 0.7105, "quality_target_met": false, "utility": 38.802159}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.797,\"max_latency_ms\":50523,\"minimum_quality\":0.747},\"task\":{\"ambiguity\":0.5204,\"category\":\"coding\",\"complexity\":0.8018,\"consequence\":0.525,\"evidence_need\":0.7607,\"family\":\"migration_planning\",\"input_tokens\":4360,\"reversibility\":0.2606,\"risk\":0.7129,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.438,\"id\":\"worker_a\",\"p95_latency_ms\":48366,\"pass_rate\":0.8154,\"provider_group\":\"provider_x\",\"quality\":0.8618,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.905,\"id\":\"worker_b\",\"p95_latency_ms\":71094,\"pass_rate\":0.9852,\"provider_group\":\"provider_y\",\"quality\":0.9667,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.179,\"id\":\"worker_c\",\"p95_latency_ms\":30323,\"pass_rate\":0.4714,\"provider_group\":\"provider_z\",\"quality\":0.7984,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":3.438,\"estimated_latency_ms\":48366,\"estimated_pass_rate\":0.6274,\"estimated_quality\":0.7105,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.797,\"max_latency_ms\":50523,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5204, "category": "coding", "complexity": 0.8018, "consequence": 0.525, "evidence_need": 0.7607, "family": "migration_planning", "input_tokens": 4360, "request_id": "synthetic-migration_planning-000760", "reversibility": 0.2606, "risk": 0.7129, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 3.438, "id": "worker_a", "p95_latency_ms": 48366, "pass_rate": 0.8154, "provider_group": "provider_x", "quality": 0.8618, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.905, "id": "worker_b", "p95_latency_ms": 71094, "pass_rate": 0.9852, "provider_group": "provider_y", "quality": 0.9667, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.179, "id": "worker_c", "p95_latency_ms": 30323, "pass_rate": 0.4714, "provider_group": "provider_z", "quality": 0.7984, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.231, "max_latency_ms": 196770, "minimum_quality": 0.7945}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000786", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.035, "estimated_latency_ms": 156995, "estimated_pass_rate": 0.7624, "estimated_quality": 0.9277, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.231, "max_latency_ms": 196770, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7624, "quality": 0.9277, "quality_target_met": true, "utility": 72.79205}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.231,\"max_latency_ms\":196770,\"minimum_quality\":0.7945},\"task\":{\"ambiguity\":0.6624,\"category\":\"review\",\"complexity\":0.4229,\"consequence\":0.6633,\"evidence_need\":0.6546,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.6672,\"risk\":0.6035,\"summary\":\"Identify evidence required before approving streaming inference for production.\"},\"workers\":[{\"credits_per_call\":1.105,\"id\":\"worker_a\",\"p95_latency_ms\":18096,\"pass_rate\":0.6111,\"provider_group\":\"provider_y\",\"quality\":0.7576,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.674,\"id\":\"worker_b\",\"p95_latency_ms\":81815,\"pass_rate\":0.8898,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.256,\"id\":\"worker_c\",\"p95_latency_ms\":57084,\"pass_rate\":0.8726,\"provider_group\":\"provider_z\",\"quality\":0.8357,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.035,\"estimated_latency_ms\":156995,\"estimated_pass_rate\":0.7624,\"estimated_quality\":0.9277,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.231,\"max_latency_ms\":196770,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6624, "category": "review", "complexity": 0.4229, "consequence": 0.6633, "evidence_need": 0.6546, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-000786", "reversibility": 0.6672, "risk": 0.6035, "summary": "Identify evidence required before approving streaming inference for production."}, "workers": [{"credits_per_call": 1.105, "id": "worker_a", "p95_latency_ms": 18096, "pass_rate": 0.6111, "provider_group": "provider_y", "quality": 0.7576, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.674, "id": "worker_b", "p95_latency_ms": 81815, "pass_rate": 0.8898, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.256, "id": "worker_c", "p95_latency_ms": 57084, "pass_rate": 0.8726, "provider_group": "provider_z", "quality": 0.8357, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.919, "max_latency_ms": 33427, "minimum_quality": 0.7592}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000350", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.031, "estimated_latency_ms": 18023, "estimated_pass_rate": 0.3026, "estimated_quality": 0.5484, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.919, "max_latency_ms": 33427, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3026, "quality": 0.5484, "quality_target_met": false, "utility": 25.9199}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.919,\"max_latency_ms\":33427,\"minimum_quality\":0.7592},\"task\":{\"ambiguity\":0.3361,\"category\":\"analysis\",\"complexity\":0.6715,\"consequence\":0.3808,\"evidence_need\":0.6154,\"family\":\"performance_tuning\",\"input_tokens\":5240,\"reversibility\":0.6066,\"risk\":0.4914,\"summary\":\"Review a proposed optimization for credit reservation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":10.196,\"id\":\"worker_a\",\"p95_latency_ms\":63372,\"pass_rate\":0.9673,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.195,\"id\":\"worker_b\",\"p95_latency_ms\":64382,\"pass_rate\":0.829,\"provider_group\":\"provider_x\",\"quality\":0.9239,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.031,\"id\":\"worker_c\",\"p95_latency_ms\":18023,\"pass_rate\":0.444,\"provider_group\":\"provider_y\",\"quality\":0.6478,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.031,\"estimated_latency_ms\":18023,\"estimated_pass_rate\":0.3026,\"estimated_quality\":0.5484,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.919,\"max_latency_ms\":33427,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3361, "category": "analysis", "complexity": 0.6715, "consequence": 0.3808, "evidence_need": 0.6154, "family": "performance_tuning", "input_tokens": 5240, "request_id": "synthetic-performance_tuning-000350", "reversibility": 0.6066, "risk": 0.4914, "summary": "Review a proposed optimization for credit reservation for hidden regressions."}, "workers": [{"credits_per_call": 10.196, "id": "worker_a", "p95_latency_ms": 63372, "pass_rate": 0.9673, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.195, "id": "worker_b", "p95_latency_ms": 64382, "pass_rate": 0.829, "provider_group": "provider_x", "quality": 0.9239, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.031, "id": "worker_c", "p95_latency_ms": 18023, "pass_rate": 0.444, "provider_group": "provider_y", "quality": 0.6478, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.818, "max_latency_ms": 40222, "minimum_quality": 0.6695}, "family": "migration_planning", "id": "synthetic-migration_planning-000055", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.186, "estimated_latency_ms": 20265, "estimated_pass_rate": 0.3779, "estimated_quality": 0.5664, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.818, "max_latency_ms": 40222, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3779, "quality": 0.5664, "quality_target_met": false, "utility": 35.036634}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.818,\"max_latency_ms\":40222,\"minimum_quality\":0.6695},\"task\":{\"ambiguity\":0.5823,\"category\":\"coding\",\"complexity\":0.7669,\"consequence\":0.4194,\"evidence_need\":0.7351,\"family\":\"migration_planning\",\"input_tokens\":3240,\"reversibility\":0.339,\"risk\":0.7308,\"summary\":\"Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.186,\"id\":\"worker_a\",\"p95_latency_ms\":20265,\"pass_rate\":0.5681,\"provider_group\":\"provider_y\",\"quality\":0.7168,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.02,\"id\":\"worker_b\",\"p95_latency_ms\":79634,\"pass_rate\":0.8577,\"provider_group\":\"provider_x\",\"quality\":0.7998,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.961,\"id\":\"worker_c\",\"p95_latency_ms\":57401,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9224,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.186,\"estimated_latency_ms\":20265,\"estimated_pass_rate\":0.3779,\"estimated_quality\":0.5664,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.818,\"max_latency_ms\":40222,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5823, "category": "coding", "complexity": 0.7669, "consequence": 0.4194, "evidence_need": 0.7351, "family": "migration_planning", "input_tokens": 3240, "request_id": "synthetic-migration_planning-000055", "reversibility": 0.339, "risk": 0.7308, "summary": "Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints."}, "workers": [{"credits_per_call": 1.186, "id": "worker_a", "p95_latency_ms": 20265, "pass_rate": 0.5681, "provider_group": "provider_y", "quality": 0.7168, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.02, "id": "worker_b", "p95_latency_ms": 79634, "pass_rate": 0.8577, "provider_group": "provider_x", "quality": 0.7998, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.961, "id": "worker_c", "p95_latency_ms": 57401, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9224, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.836, "max_latency_ms": 28845, "minimum_quality": 0.6369}, "family": "migration_planning", "id": "synthetic-migration_planning-000367", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.978, "estimated_latency_ms": 23411, "estimated_pass_rate": 0.2829, "estimated_quality": 0.5273, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.836, "max_latency_ms": 28845, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2829, "quality": 0.5273, "quality_target_met": false, "utility": 29.159166}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.836,\"max_latency_ms\":28845,\"minimum_quality\":0.6369},\"task\":{\"ambiguity\":0.5648,\"category\":\"coding\",\"complexity\":0.8266,\"consequence\":0.4749,\"evidence_need\":0.8416,\"family\":\"migration_planning\",\"input_tokens\":7120,\"reversibility\":0.5161,\"risk\":0.7169,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":0.978,\"id\":\"worker_a\",\"p95_latency_ms\":23411,\"pass_rate\":0.4774,\"provider_group\":\"provider_z\",\"quality\":0.6863,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.966,\"id\":\"worker_b\",\"p95_latency_ms\":85523,\"pass_rate\":0.8914,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.047,\"id\":\"worker_c\",\"p95_latency_ms\":55422,\"pass_rate\":0.8027,\"provider_group\":\"provider_x\",\"quality\":0.8394,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.978,\"estimated_latency_ms\":23411,\"estimated_pass_rate\":0.2829,\"estimated_quality\":0.5273,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.836,\"max_latency_ms\":28845,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5648, "category": "coding", "complexity": 0.8266, "consequence": 0.4749, "evidence_need": 0.8416, "family": "migration_planning", "input_tokens": 7120, "request_id": "synthetic-migration_planning-000367", "reversibility": 0.5161, "risk": 0.7169, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 0.978, "id": "worker_a", "p95_latency_ms": 23411, "pass_rate": 0.4774, "provider_group": "provider_z", "quality": 0.6863, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.966, "id": "worker_b", "p95_latency_ms": 85523, "pass_rate": 0.8914, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.047, "id": "worker_c", "p95_latency_ms": 55422, "pass_rate": 0.8027, "provider_group": "provider_x", "quality": 0.8394, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.078, "max_latency_ms": 35369, "minimum_quality": 0.5571}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000090", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.824, "estimated_latency_ms": 24154, "estimated_pass_rate": 0.3141, "estimated_quality": 0.5624, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.078, "max_latency_ms": 35369, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3141, "quality": 0.5624, "quality_target_met": true, "utility": 21.284109}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.078,\"max_latency_ms\":35369,\"minimum_quality\":0.5571},\"task\":{\"ambiguity\":0.563,\"category\":\"review\",\"complexity\":0.7153,\"consequence\":0.7443,\"evidence_need\":0.8409,\"family\":\"policy_compliance\",\"input_tokens\":2960,\"reversibility\":0.4878,\"risk\":0.7732,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":2.953,\"id\":\"worker_a\",\"p95_latency_ms\":50946,\"pass_rate\":0.875,\"provider_group\":\"provider_x\",\"quality\":0.8128,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.302,\"id\":\"worker_b\",\"p95_latency_ms\":62609,\"pass_rate\":0.9291,\"provider_group\":\"provider_z\",\"quality\":0.9722,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.824,\"id\":\"worker_c\",\"p95_latency_ms\":24154,\"pass_rate\":0.5004,\"provider_group\":\"provider_y\",\"quality\":0.7052,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.824,\"estimated_latency_ms\":24154,\"estimated_pass_rate\":0.3141,\"estimated_quality\":0.5624,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.078,\"max_latency_ms\":35369,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.563, "category": "review", "complexity": 0.7153, "consequence": 0.7443, "evidence_need": 0.8409, "family": "policy_compliance", "input_tokens": 2960, "request_id": "synthetic-policy_compliance-000090", "reversibility": 0.4878, "risk": 0.7732, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 2.953, "id": "worker_a", "p95_latency_ms": 50946, "pass_rate": 0.875, "provider_group": "provider_x", "quality": 0.8128, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.302, "id": "worker_b", "p95_latency_ms": 62609, "pass_rate": 0.9291, "provider_group": "provider_z", "quality": 0.9722, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.824, "id": "worker_c", "p95_latency_ms": 24154, "pass_rate": 0.5004, "provider_group": "provider_y", "quality": 0.7052, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.278, "max_latency_ms": 211421, "minimum_quality": 0.8152}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000150", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.049, "estimated_latency_ms": 170839, "estimated_pass_rate": 0.6896, "estimated_quality": 0.8814, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.278, "max_latency_ms": 211421, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6896, "quality": 0.8814, "quality_target_met": true, "utility": 73.51927}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.278,\"max_latency_ms\":211421,\"minimum_quality\":0.8152},\"task\":{\"ambiguity\":0.6245,\"category\":\"review\",\"complexity\":0.8227,\"consequence\":0.9171,\"evidence_need\":0.853,\"family\":\"policy_compliance\",\"input_tokens\":7200,\"reversibility\":0.3394,\"risk\":0.4725,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":3.287,\"id\":\"worker_a\",\"p95_latency_ms\":53004,\"pass_rate\":0.8272,\"provider_group\":\"provider_y\",\"quality\":0.8334,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.924,\"id\":\"worker_b\",\"p95_latency_ms\":31177,\"pass_rate\":0.4755,\"provider_group\":\"provider_x\",\"quality\":0.718,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.838,\"id\":\"worker_c\",\"p95_latency_ms\":86658,\"pass_rate\":0.9196,\"provider_group\":\"provider_z\",\"quality\":0.923,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.049,\"estimated_latency_ms\":170839,\"estimated_pass_rate\":0.6896,\"estimated_quality\":0.8814,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.278,\"max_latency_ms\":211421,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6245, "category": "review", "complexity": 0.8227, "consequence": 0.9171, "evidence_need": 0.853, "family": "policy_compliance", "input_tokens": 7200, "request_id": "synthetic-policy_compliance-000150", "reversibility": 0.3394, "risk": 0.4725, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 3.287, "id": "worker_a", "p95_latency_ms": 53004, "pass_rate": 0.8272, "provider_group": "provider_y", "quality": 0.8334, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.924, "id": "worker_b", "p95_latency_ms": 31177, "pass_rate": 0.4755, "provider_group": "provider_x", "quality": 0.718, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.838, "id": "worker_c", "p95_latency_ms": 86658, "pass_rate": 0.9196, "provider_group": "provider_z", "quality": 0.923, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.147, "max_latency_ms": 35295, "minimum_quality": 0.6957}, "family": "migration_planning", "id": "synthetic-migration_planning-000658", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.071, "estimated_latency_ms": 30357, "estimated_pass_rate": 0.2945, "estimated_quality": 0.5647, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.147, "max_latency_ms": 35295, "policy": "lean", "reason_codes": ["best_feasible_below_target", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2945, "quality": 0.5647, "quality_target_met": false, "utility": 28.92063}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.147,\"max_latency_ms\":35295,\"minimum_quality\":0.6957},\"task\":{\"ambiguity\":0.5473,\"category\":\"coding\",\"complexity\":0.6446,\"consequence\":0.6844,\"evidence_need\":0.4499,\"family\":\"migration_planning\",\"input_tokens\":520,\"reversibility\":0.492,\"risk\":0.6569,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.071,\"id\":\"worker_a\",\"p95_latency_ms\":30357,\"pass_rate\":0.4622,\"provider_group\":\"provider_y\",\"quality\":0.6853,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.085,\"id\":\"worker_b\",\"p95_latency_ms\":54780,\"pass_rate\":0.8875,\"provider_group\":\"provider_x\",\"quality\":0.8406,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.977,\"id\":\"worker_c\",\"p95_latency_ms\":88248,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.957,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.071,\"estimated_latency_ms\":30357,\"estimated_pass_rate\":0.2945,\"estimated_quality\":0.5647,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.147,\"max_latency_ms\":35295,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5473, "category": "coding", "complexity": 0.6446, "consequence": 0.6844, "evidence_need": 0.4499, "family": "migration_planning", "input_tokens": 520, "request_id": "synthetic-migration_planning-000658", "reversibility": 0.492, "risk": 0.6569, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 1.071, "id": "worker_a", "p95_latency_ms": 30357, "pass_rate": 0.4622, "provider_group": "provider_y", "quality": 0.6853, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.085, "id": "worker_b", "p95_latency_ms": 54780, "pass_rate": 0.8875, "provider_group": "provider_x", "quality": 0.8406, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.977, "id": "worker_c", "p95_latency_ms": 88248, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.957, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.276, "max_latency_ms": 49925, "minimum_quality": 0.4809}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000308", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.152, "estimated_latency_ms": 30803, "estimated_pass_rate": 0.2811, "estimated_quality": 0.581, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.276, "max_latency_ms": 49925, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2811, "quality": 0.581, "quality_target_met": true, "utility": 37.249327}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.276,\"max_latency_ms\":49925,\"minimum_quality\":0.4809},\"task\":{\"ambiguity\":0.6288,\"category\":\"analysis\",\"complexity\":0.487,\"consequence\":0.5554,\"evidence_need\":0.599,\"family\":\"performance_tuning\",\"input_tokens\":4840,\"reversibility\":0.8898,\"risk\":0.4532,\"summary\":\"Diagnose latency variance in streaming inference and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.743,\"id\":\"worker_a\",\"p95_latency_ms\":86502,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.612,\"id\":\"worker_b\",\"p95_latency_ms\":64907,\"pass_rate\":0.8202,\"provider_group\":\"provider_z\",\"quality\":0.9707,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.152,\"id\":\"worker_c\",\"p95_latency_ms\":30803,\"pass_rate\":0.4206,\"provider_group\":\"provider_y\",\"quality\":0.6639,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.152,\"estimated_latency_ms\":30803,\"estimated_pass_rate\":0.2811,\"estimated_quality\":0.581,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.276,\"max_latency_ms\":49925,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6288, "category": "analysis", "complexity": 0.487, "consequence": 0.5554, "evidence_need": 0.599, "family": "performance_tuning", "input_tokens": 4840, "request_id": "synthetic-performance_tuning-000308", "reversibility": 0.8898, "risk": 0.4532, "summary": "Diagnose latency variance in streaming inference and rank the next experiments."}, "workers": [{"credits_per_call": 11.743, "id": "worker_a", "p95_latency_ms": 86502, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.612, "id": "worker_b", "p95_latency_ms": 64907, "pass_rate": 0.8202, "provider_group": "provider_z", "quality": 0.9707, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.152, "id": "worker_c", "p95_latency_ms": 30803, "pass_rate": 0.4206, "provider_group": "provider_y", "quality": 0.6639, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.973, "max_latency_ms": 40991, "minimum_quality": 0.5726}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000248", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.875, "estimated_latency_ms": 26529, "estimated_pass_rate": 0.273, "estimated_quality": 0.5856, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.973, "max_latency_ms": 40991, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.273, "quality": 0.5856, "quality_target_met": true, "utility": 36.569432}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.973,\"max_latency_ms\":40991,\"minimum_quality\":0.5726},\"task\":{\"ambiguity\":0.5543,\"category\":\"analysis\",\"complexity\":0.4675,\"consequence\":0.3615,\"evidence_need\":0.6373,\"family\":\"performance_tuning\",\"input_tokens\":4720,\"reversibility\":0.5787,\"risk\":0.4856,\"summary\":\"Choose between throughput and tail-latency changes for deployment rollback.\"},\"workers\":[{\"credits_per_call\":3.478,\"id\":\"worker_a\",\"p95_latency_ms\":82801,\"pass_rate\":0.9377,\"provider_group\":\"provider_z\",\"quality\":0.9064,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.875,\"id\":\"worker_b\",\"p95_latency_ms\":26529,\"pass_rate\":0.4073,\"provider_group\":\"provider_x\",\"quality\":0.6623,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.769,\"id\":\"worker_c\",\"p95_latency_ms\":63051,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.875,\"estimated_latency_ms\":26529,\"estimated_pass_rate\":0.273,\"estimated_quality\":0.5856,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.973,\"max_latency_ms\":40991,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5543, "category": "analysis", "complexity": 0.4675, "consequence": 0.3615, "evidence_need": 0.6373, "family": "performance_tuning", "input_tokens": 4720, "request_id": "synthetic-performance_tuning-000248", "reversibility": 0.5787, "risk": 0.4856, "summary": "Choose between throughput and tail-latency changes for deployment rollback."}, "workers": [{"credits_per_call": 3.478, "id": "worker_a", "p95_latency_ms": 82801, "pass_rate": 0.9377, "provider_group": "provider_z", "quality": 0.9064, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.875, "id": "worker_b", "p95_latency_ms": 26529, "pass_rate": 0.4073, "provider_group": "provider_x", "quality": 0.6623, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.769, "id": "worker_c", "p95_latency_ms": 63051, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.153, "max_latency_ms": 37745, "minimum_quality": 0.5039}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000591", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.076, "estimated_latency_ms": 26377, "estimated_pass_rate": 0.4397, "estimated_quality": 0.62, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.153, "max_latency_ms": 37745, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4397, "quality": 0.62, "quality_target_met": true, "utility": 30.235362}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.153,\"max_latency_ms\":37745,\"minimum_quality\":0.5039},\"task\":{\"ambiguity\":0.5458,\"category\":\"review\",\"complexity\":0.5329,\"consequence\":0.7465,\"evidence_need\":0.8076,\"family\":\"policy_compliance\",\"input_tokens\":5760,\"reversibility\":0.5749,\"risk\":0.6696,\"summary\":\"Review whether provider failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.938,\"id\":\"worker_a\",\"p95_latency_ms\":52185,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.755,\"id\":\"worker_b\",\"p95_latency_ms\":48831,\"pass_rate\":0.8183,\"provider_group\":\"provider_x\",\"quality\":0.8582,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.076,\"id\":\"worker_c\",\"p95_latency_ms\":26377,\"pass_rate\":0.5956,\"provider_group\":\"provider_z\",\"quality\":0.7212,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.076,\"estimated_latency_ms\":26377,\"estimated_pass_rate\":0.4397,\"estimated_quality\":0.62,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.153,\"max_latency_ms\":37745,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5458, "category": "review", "complexity": 0.5329, "consequence": 0.7465, "evidence_need": 0.8076, "family": "policy_compliance", "input_tokens": 5760, "request_id": "synthetic-policy_compliance-000591", "reversibility": 0.5749, "risk": 0.6696, "summary": "Review whether provider failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.938, "id": "worker_a", "p95_latency_ms": 52185, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.755, "id": "worker_b", "p95_latency_ms": 48831, "pass_rate": 0.8183, "provider_group": "provider_x", "quality": 0.8582, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.076, "id": "worker_c", "p95_latency_ms": 26377, "pass_rate": 0.5956, "provider_group": "provider_z", "quality": 0.7212, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.729, "max_latency_ms": 210471, "minimum_quality": 0.7829}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000263", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.315, "estimated_latency_ms": 186420, "estimated_pass_rate": 0.7363, "estimated_quality": 0.9025, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.729, "max_latency_ms": 210471, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7363, "quality": 0.9025, "quality_target_met": true, "utility": 71.742635}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.729,\"max_latency_ms\":210471,\"minimum_quality\":0.7829},\"task\":{\"ambiguity\":0.6815,\"category\":\"analysis\",\"complexity\":0.8099,\"consequence\":0.5194,\"evidence_need\":0.7225,\"family\":\"performance_tuning\",\"input_tokens\":2360,\"reversibility\":0.6496,\"risk\":0.3067,\"summary\":\"Choose between throughput and tail-latency changes for observability.\"},\"workers\":[{\"credits_per_call\":12.061,\"id\":\"worker_a\",\"p95_latency_ms\":88388,\"pass_rate\":0.9752,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.289,\"id\":\"worker_b\",\"p95_latency_ms\":74450,\"pass_rate\":0.8127,\"provider_group\":\"provider_z\",\"quality\":0.89,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.965,\"id\":\"worker_c\",\"p95_latency_ms\":23582,\"pass_rate\":0.5515,\"provider_group\":\"provider_y\",\"quality\":0.6159,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.315,\"estimated_latency_ms\":186420,\"estimated_pass_rate\":0.7363,\"estimated_quality\":0.9025,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.729,\"max_latency_ms\":210471,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6815, "category": "analysis", "complexity": 0.8099, "consequence": 0.5194, "evidence_need": 0.7225, "family": "performance_tuning", "input_tokens": 2360, "request_id": "synthetic-performance_tuning-000263", "reversibility": 0.6496, "risk": 0.3067, "summary": "Choose between throughput and tail-latency changes for observability."}, "workers": [{"credits_per_call": 12.061, "id": "worker_a", "p95_latency_ms": 88388, "pass_rate": 0.9752, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.289, "id": "worker_b", "p95_latency_ms": 74450, "pass_rate": 0.8127, "provider_group": "provider_z", "quality": 0.89, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.965, "id": "worker_c", "p95_latency_ms": 23582, "pass_rate": 0.5515, "provider_group": "provider_y", "quality": 0.6159, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 4.27, "max_latency_ms": 205706, "minimum_quality": 0.7869}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000075", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.559, "estimated_latency_ms": 83071, "estimated_pass_rate": 0.6918, "estimated_quality": 0.7856, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.27, "max_latency_ms": 205706, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6918, "quality": 0.7856, "quality_target_met": false, "utility": 56.040146}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":4.27,\"max_latency_ms\":205706,\"minimum_quality\":0.7869},\"task\":{\"ambiguity\":0.7875,\"category\":\"review\",\"complexity\":0.8141,\"consequence\":0.6608,\"evidence_need\":0.5973,\"family\":\"policy_compliance\",\"input_tokens\":2760,\"reversibility\":0.434,\"risk\":0.6419,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.816,\"id\":\"worker_a\",\"p95_latency_ms\":32840,\"pass_rate\":0.5208,\"provider_group\":\"provider_z\",\"quality\":0.6883,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.743,\"id\":\"worker_b\",\"p95_latency_ms\":50231,\"pass_rate\":0.8234,\"provider_group\":\"provider_y\",\"quality\":0.9049,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.09,\"id\":\"worker_c\",\"p95_latency_ms\":83264,\"pass_rate\":0.9936,\"provider_group\":\"provider_x\",\"quality\":0.9337,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.559,\"estimated_latency_ms\":83071,\"estimated_pass_rate\":0.6918,\"estimated_quality\":0.7856,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.27,\"max_latency_ms\":205706,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7875, "category": "review", "complexity": 0.8141, "consequence": 0.6608, "evidence_need": 0.5973, "family": "policy_compliance", "input_tokens": 2760, "request_id": "synthetic-policy_compliance-000075", "reversibility": 0.434, "risk": 0.6419, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 0.816, "id": "worker_a", "p95_latency_ms": 32840, "pass_rate": 0.5208, "provider_group": "provider_z", "quality": 0.6883, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.743, "id": "worker_b", "p95_latency_ms": 50231, "pass_rate": 0.8234, "provider_group": "provider_y", "quality": 0.9049, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 10.09, "id": "worker_c", "p95_latency_ms": 83264, "pass_rate": 0.9936, "provider_group": "provider_x", "quality": 0.9337, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.548, "max_latency_ms": 179941, "minimum_quality": 0.8855}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000891", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.142, "estimated_latency_ms": 94252, "estimated_pass_rate": 0.6051, "estimated_quality": 0.8291, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.548, "max_latency_ms": 179941, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6051, "quality": 0.8291, "quality_target_met": false, "utility": 60.503923}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.548,\"max_latency_ms\":179941,\"minimum_quality\":0.8855},\"task\":{\"ambiguity\":0.797,\"category\":\"review\",\"complexity\":0.5073,\"consequence\":0.7173,\"evidence_need\":0.8387,\"family\":\"policy_compliance\",\"input_tokens\":2760,\"reversibility\":0.3407,\"risk\":0.7039,\"summary\":\"Review whether usage metering satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.265,\"id\":\"worker_a\",\"p95_latency_ms\":66677,\"pass_rate\":0.81,\"provider_group\":\"provider_y\",\"quality\":0.8371,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.032,\"id\":\"worker_b\",\"p95_latency_ms\":55557,\"pass_rate\":0.9115,\"provider_group\":\"provider_x\",\"quality\":0.9442,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.877,\"id\":\"worker_c\",\"p95_latency_ms\":27575,\"pass_rate\":0.5109,\"provider_group\":\"provider_z\",\"quality\":0.762,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.142,\"estimated_latency_ms\":94252,\"estimated_pass_rate\":0.6051,\"estimated_quality\":0.8291,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.548,\"max_latency_ms\":179941,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.797, "category": "review", "complexity": 0.5073, "consequence": 0.7173, "evidence_need": 0.8387, "family": "policy_compliance", "input_tokens": 2760, "request_id": "synthetic-policy_compliance-000891", "reversibility": 0.3407, "risk": 0.7039, "summary": "Review whether usage metering satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.265, "id": "worker_a", "p95_latency_ms": 66677, "pass_rate": 0.81, "provider_group": "provider_y", "quality": 0.8371, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.032, "id": "worker_b", "p95_latency_ms": 55557, "pass_rate": 0.9115, "provider_group": "provider_x", "quality": 0.9442, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.877, "id": "worker_c", "p95_latency_ms": 27575, "pass_rate": 0.5109, "provider_group": "provider_z", "quality": 0.762, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.702, "max_latency_ms": 158712, "minimum_quality": 0.7805}, "family": "migration_planning", "id": "synthetic-migration_planning-000565", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.406, "estimated_latency_ms": 126811, "estimated_pass_rate": 0.7088, "estimated_quality": 0.8674, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.702, "max_latency_ms": 158712, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7088, "quality": 0.8674, "quality_target_met": true, "utility": 69.698983}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.702,\"max_latency_ms\":158712,\"minimum_quality\":0.7805},\"task\":{\"ambiguity\":0.5997,\"category\":\"coding\",\"complexity\":0.6188,\"consequence\":0.5747,\"evidence_need\":0.4507,\"family\":\"migration_planning\",\"input_tokens\":3920,\"reversibility\":0.4057,\"risk\":0.7196,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.662,\"id\":\"worker_a\",\"p95_latency_ms\":56019,\"pass_rate\":0.8954,\"provider_group\":\"provider_x\",\"quality\":0.9149,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.89,\"id\":\"worker_b\",\"p95_latency_ms\":22588,\"pass_rate\":0.4546,\"provider_group\":\"provider_y\",\"quality\":0.7136,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.854,\"id\":\"worker_c\",\"p95_latency_ms\":48204,\"pass_rate\":0.7961,\"provider_group\":\"provider_z\",\"quality\":0.907,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.406,\"estimated_latency_ms\":126811,\"estimated_pass_rate\":0.7088,\"estimated_quality\":0.8674,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.702,\"max_latency_ms\":158712,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5997, "category": "coding", "complexity": 0.6188, "consequence": 0.5747, "evidence_need": 0.4507, "family": "migration_planning", "input_tokens": 3920, "request_id": "synthetic-migration_planning-000565", "reversibility": 0.4057, "risk": 0.7196, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 12.662, "id": "worker_a", "p95_latency_ms": 56019, "pass_rate": 0.8954, "provider_group": "provider_x", "quality": 0.9149, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.89, "id": "worker_b", "p95_latency_ms": 22588, "pass_rate": 0.4546, "provider_group": "provider_y", "quality": 0.7136, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.854, "id": "worker_c", "p95_latency_ms": 48204, "pass_rate": 0.7961, "provider_group": "provider_z", "quality": 0.907, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.685, "max_latency_ms": 188659, "minimum_quality": 0.8445}, "family": "migration_planning", "id": "synthetic-migration_planning-000280", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.185, "estimated_latency_ms": 130562, "estimated_pass_rate": 0.7325, "estimated_quality": 0.9157, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 22.685, "max_latency_ms": 188659, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7325, "quality": 0.9157, "quality_target_met": true, "utility": 78.278101}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.685,\"max_latency_ms\":188659,\"minimum_quality\":0.8445},\"task\":{\"ambiguity\":0.453,\"category\":\"coding\",\"complexity\":0.9216,\"consequence\":0.65,\"evidence_need\":0.7942,\"family\":\"migration_planning\",\"input_tokens\":7320,\"reversibility\":0.4015,\"risk\":0.6769,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.166,\"id\":\"worker_a\",\"p95_latency_ms\":26660,\"pass_rate\":0.5361,\"provider_group\":\"provider_z\",\"quality\":0.7289,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.972,\"id\":\"worker_b\",\"p95_latency_ms\":56994,\"pass_rate\":0.9173,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.047,\"id\":\"worker_c\",\"p95_latency_ms\":46908,\"pass_rate\":0.8042,\"provider_group\":\"provider_x\",\"quality\":0.8882,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.185,\"estimated_latency_ms\":130562,\"estimated_pass_rate\":0.7325,\"estimated_quality\":0.9157,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.685,\"max_latency_ms\":188659,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.453, "category": "coding", "complexity": 0.9216, "consequence": 0.65, "evidence_need": 0.7942, "family": "migration_planning", "input_tokens": 7320, "request_id": "synthetic-migration_planning-000280", "reversibility": 0.4015, "risk": 0.6769, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 1.166, "id": "worker_a", "p95_latency_ms": 26660, "pass_rate": 0.5361, "provider_group": "provider_z", "quality": 0.7289, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.972, "id": "worker_b", "p95_latency_ms": 56994, "pass_rate": 0.9173, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.047, "id": "worker_c", "p95_latency_ms": 46908, "pass_rate": 0.8042, "provider_group": "provider_x", "quality": 0.8882, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.167, "max_latency_ms": 170799, "minimum_quality": 0.7627}, "family": "migration_planning", "id": "synthetic-migration_planning-001270", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.143, "estimated_latency_ms": 70985, "estimated_pass_rate": 0.5811, "estimated_quality": 0.7818, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.167, "max_latency_ms": 170799, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5811, "quality": 0.7818, "quality_target_met": true, "utility": 61.347989}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.167,\"max_latency_ms\":170799,\"minimum_quality\":0.7627},\"task\":{\"ambiguity\":0.4135,\"category\":\"coding\",\"complexity\":0.7318,\"consequence\":0.8113,\"evidence_need\":0.6818,\"family\":\"migration_planning\",\"input_tokens\":360,\"reversibility\":0.6005,\"risk\":0.7319,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":1.089,\"id\":\"worker_a\",\"p95_latency_ms\":24172,\"pass_rate\":0.4974,\"provider_group\":\"provider_y\",\"quality\":0.7401,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.054,\"id\":\"worker_b\",\"p95_latency_ms\":46813,\"pass_rate\":0.7559,\"provider_group\":\"provider_z\",\"quality\":0.8291,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.335,\"id\":\"worker_c\",\"p95_latency_ms\":61071,\"pass_rate\":0.8893,\"provider_group\":\"provider_x\",\"quality\":0.9239,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.143,\"estimated_latency_ms\":70985,\"estimated_pass_rate\":0.5811,\"estimated_quality\":0.7818,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.167,\"max_latency_ms\":170799,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4135, "category": "coding", "complexity": 0.7318, "consequence": 0.8113, "evidence_need": 0.6818, "family": "migration_planning", "input_tokens": 360, "request_id": "synthetic-migration_planning-001270", "reversibility": 0.6005, "risk": 0.7319, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 1.089, "id": "worker_a", "p95_latency_ms": 24172, "pass_rate": 0.4974, "provider_group": "provider_y", "quality": 0.7401, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.054, "id": "worker_b", "p95_latency_ms": 46813, "pass_rate": 0.7559, "provider_group": "provider_z", "quality": 0.8291, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.335, "id": "worker_c", "p95_latency_ms": 61071, "pass_rate": 0.8893, "provider_group": "provider_x", "quality": 0.9239, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.927, "max_latency_ms": 165801, "minimum_quality": 0.8977}, "family": "migration_planning", "id": "synthetic-migration_planning-000922", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.935, "estimated_latency_ms": 138684, "estimated_pass_rate": 0.6893, "estimated_quality": 0.9093, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.927, "max_latency_ms": 165801, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6893, "quality": 0.9093, "quality_target_met": true, "utility": 67.701021}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.927,\"max_latency_ms\":165801,\"minimum_quality\":0.8977},\"task\":{\"ambiguity\":0.5184,\"category\":\"coding\",\"complexity\":0.7638,\"consequence\":0.6391,\"evidence_need\":0.4286,\"family\":\"migration_planning\",\"input_tokens\":1560,\"reversibility\":0.4127,\"risk\":0.3712,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.841,\"id\":\"worker_a\",\"p95_latency_ms\":29794,\"pass_rate\":0.5262,\"provider_group\":\"provider_x\",\"quality\":0.7782,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.043,\"id\":\"worker_b\",\"p95_latency_ms\":57152,\"pass_rate\":0.7322,\"provider_group\":\"provider_z\",\"quality\":0.8185,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.051,\"id\":\"worker_c\",\"p95_latency_ms\":51738,\"pass_rate\":0.9122,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.935,\"estimated_latency_ms\":138684,\"estimated_pass_rate\":0.6893,\"estimated_quality\":0.9093,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.927,\"max_latency_ms\":165801,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5184, "category": "coding", "complexity": 0.7638, "consequence": 0.6391, "evidence_need": 0.4286, "family": "migration_planning", "input_tokens": 1560, "request_id": "synthetic-migration_planning-000922", "reversibility": 0.4127, "risk": 0.3712, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 0.841, "id": "worker_a", "p95_latency_ms": 29794, "pass_rate": 0.5262, "provider_group": "provider_x", "quality": 0.7782, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.043, "id": "worker_b", "p95_latency_ms": 57152, "pass_rate": 0.7322, "provider_group": "provider_z", "quality": 0.8185, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.051, "id": "worker_c", "p95_latency_ms": 51738, "pass_rate": 0.9122, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.358, "max_latency_ms": 207638, "minimum_quality": 0.8522}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000573", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.683, "estimated_latency_ms": 178075, "estimated_pass_rate": 0.7192, "estimated_quality": 0.8746, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 16.358, "max_latency_ms": 207638, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7192, "quality": 0.8746, "quality_target_met": true, "utility": 72.049197}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.358,\"max_latency_ms\":207638,\"minimum_quality\":0.8522},\"task\":{\"ambiguity\":0.7147,\"category\":\"review\",\"complexity\":0.5934,\"consequence\":0.5495,\"evidence_need\":0.719,\"family\":\"policy_compliance\",\"input_tokens\":8000,\"reversibility\":0.3974,\"risk\":0.8194,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":0.885,\"id\":\"worker_a\",\"p95_latency_ms\":31494,\"pass_rate\":0.5074,\"provider_group\":\"provider_y\",\"quality\":0.6779,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.8,\"id\":\"worker_b\",\"p95_latency_ms\":70016,\"pass_rate\":0.9608,\"provider_group\":\"provider_z\",\"quality\":0.9511,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.998,\"id\":\"worker_c\",\"p95_latency_ms\":76565,\"pass_rate\":0.7216,\"provider_group\":\"provider_x\",\"quality\":0.8057,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.683,\"estimated_latency_ms\":178075,\"estimated_pass_rate\":0.7192,\"estimated_quality\":0.8746,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.358,\"max_latency_ms\":207638,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7147, "category": "review", "complexity": 0.5934, "consequence": 0.5495, "evidence_need": 0.719, "family": "policy_compliance", "input_tokens": 8000, "request_id": "synthetic-policy_compliance-000573", "reversibility": 0.3974, "risk": 0.8194, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 0.885, "id": "worker_a", "p95_latency_ms": 31494, "pass_rate": 0.5074, "provider_group": "provider_y", "quality": 0.6779, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.8, "id": "worker_b", "p95_latency_ms": 70016, "pass_rate": 0.9608, "provider_group": "provider_z", "quality": 0.9511, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.998, "id": "worker_c", "p95_latency_ms": 76565, "pass_rate": 0.7216, "provider_group": "provider_x", "quality": 0.8057, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.321, "max_latency_ms": 203885, "minimum_quality": 0.8538}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000660", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.354, "estimated_latency_ms": 156649, "estimated_pass_rate": 0.6799, "estimated_quality": 0.8968, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 26.321, "max_latency_ms": 203885, "policy": "strict", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6799, "quality": 0.8968, "quality_target_met": true, "utility": 76.552766}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.321,\"max_latency_ms\":203885,\"minimum_quality\":0.8538},\"task\":{\"ambiguity\":0.6566,\"category\":\"review\",\"complexity\":0.6325,\"consequence\":0.6942,\"evidence_need\":0.7693,\"family\":\"policy_compliance\",\"input_tokens\":4920,\"reversibility\":0.5756,\"risk\":0.8497,\"summary\":\"Identify evidence required before approving streaming inference for production.\"},\"workers\":[{\"credits_per_call\":3.819,\"id\":\"worker_a\",\"p95_latency_ms\":60475,\"pass_rate\":0.7589,\"provider_group\":\"provider_x\",\"quality\":0.8069,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.352,\"id\":\"worker_b\",\"p95_latency_ms\":71113,\"pass_rate\":0.8851,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.183,\"id\":\"worker_c\",\"p95_latency_ms\":25061,\"pass_rate\":0.4517,\"provider_group\":\"provider_y\",\"quality\":0.682,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.354,\"estimated_latency_ms\":156649,\"estimated_pass_rate\":0.6799,\"estimated_quality\":0.8968,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.321,\"max_latency_ms\":203885,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6566, "category": "review", "complexity": 0.6325, "consequence": 0.6942, "evidence_need": 0.7693, "family": "policy_compliance", "input_tokens": 4920, "request_id": "synthetic-policy_compliance-000660", "reversibility": 0.5756, "risk": 0.8497, "summary": "Identify evidence required before approving streaming inference for production."}, "workers": [{"credits_per_call": 3.819, "id": "worker_a", "p95_latency_ms": 60475, "pass_rate": 0.7589, "provider_group": "provider_x", "quality": 0.8069, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 14.352, "id": "worker_b", "p95_latency_ms": 71113, "pass_rate": 0.8851, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.183, "id": "worker_c", "p95_latency_ms": 25061, "pass_rate": 0.4517, "provider_group": "provider_y", "quality": 0.682, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.732, "max_latency_ms": 135951, "minimum_quality": 0.7361}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000906", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.145, "estimated_latency_ms": 109312, "estimated_pass_rate": 0.713, "estimated_quality": 0.8903, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.732, "max_latency_ms": 135951, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.713, "quality": 0.8903, "quality_target_met": true, "utility": 66.568157}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.732,\"max_latency_ms\":135951,\"minimum_quality\":0.7361},\"task\":{\"ambiguity\":0.5553,\"category\":\"review\",\"complexity\":0.8464,\"consequence\":0.712,\"evidence_need\":0.7518,\"family\":\"policy_compliance\",\"input_tokens\":4800,\"reversibility\":0.5566,\"risk\":0.7914,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":1.209,\"id\":\"worker_a\",\"p95_latency_ms\":25332,\"pass_rate\":0.4841,\"provider_group\":\"provider_x\",\"quality\":0.7688,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.936,\"id\":\"worker_b\",\"p95_latency_ms\":83980,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.959,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.877,\"id\":\"worker_c\",\"p95_latency_ms\":80696,\"pass_rate\":0.8025,\"provider_group\":\"provider_y\",\"quality\":0.8668,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.145,\"estimated_latency_ms\":109312,\"estimated_pass_rate\":0.713,\"estimated_quality\":0.8903,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.732,\"max_latency_ms\":135951,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5553, "category": "review", "complexity": 0.8464, "consequence": 0.712, "evidence_need": 0.7518, "family": "policy_compliance", "input_tokens": 4800, "request_id": "synthetic-policy_compliance-000906", "reversibility": 0.5566, "risk": 0.7914, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 1.209, "id": "worker_a", "p95_latency_ms": 25332, "pass_rate": 0.4841, "provider_group": "provider_x", "quality": 0.7688, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.936, "id": "worker_b", "p95_latency_ms": 83980, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.959, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.877, "id": "worker_c", "p95_latency_ms": 80696, "pass_rate": 0.8025, "provider_group": "provider_y", "quality": 0.8668, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.071, "max_latency_ms": 217550, "minimum_quality": 0.7457}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000957", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.238, "estimated_latency_ms": 149930, "estimated_pass_rate": 0.859, "estimated_quality": 0.9144, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.071, "max_latency_ms": 217550, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.859, "quality": 0.9144, "quality_target_met": true, "utility": 69.46855}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.071,\"max_latency_ms\":217550,\"minimum_quality\":0.7457},\"task\":{\"ambiguity\":0.799,\"category\":\"review\",\"complexity\":0.7987,\"consequence\":0.7734,\"evidence_need\":0.8233,\"family\":\"policy_compliance\",\"input_tokens\":5800,\"reversibility\":0.3796,\"risk\":0.4923,\"summary\":\"Identify evidence required before approving background jobs for production.\"},\"workers\":[{\"credits_per_call\":3.355,\"id\":\"worker_a\",\"p95_latency_ms\":75365,\"pass_rate\":0.7648,\"provider_group\":\"provider_x\",\"quality\":0.8701,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.874,\"id\":\"worker_b\",\"p95_latency_ms\":18759,\"pass_rate\":0.5971,\"provider_group\":\"provider_z\",\"quality\":0.7584,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.883,\"id\":\"worker_c\",\"p95_latency_ms\":74565,\"pass_rate\":0.9762,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.238,\"estimated_latency_ms\":149930,\"estimated_pass_rate\":0.859,\"estimated_quality\":0.9144,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.071,\"max_latency_ms\":217550,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.799, "category": "review", "complexity": 0.7987, "consequence": 0.7734, "evidence_need": 0.8233, "family": "policy_compliance", "input_tokens": 5800, "request_id": "synthetic-policy_compliance-000957", "reversibility": 0.3796, "risk": 0.4923, "summary": "Identify evidence required before approving background jobs for production."}, "workers": [{"credits_per_call": 3.355, "id": "worker_a", "p95_latency_ms": 75365, "pass_rate": 0.7648, "provider_group": "provider_x", "quality": 0.8701, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.874, "id": "worker_b", "p95_latency_ms": 18759, "pass_rate": 0.5971, "provider_group": "provider_z", "quality": 0.7584, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.883, "id": "worker_c", "p95_latency_ms": 74565, "pass_rate": 0.9762, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.177, "max_latency_ms": 46169, "minimum_quality": 0.5599}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000299", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.889, "estimated_latency_ms": 23355, "estimated_pass_rate": 0.3936, "estimated_quality": 0.5299, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.177, "max_latency_ms": 46169, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3936, "quality": 0.5299, "quality_target_met": false, "utility": 36.411469}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.177,\"max_latency_ms\":46169,\"minimum_quality\":0.5599},\"task\":{\"ambiguity\":0.3508,\"category\":\"analysis\",\"complexity\":0.8395,\"consequence\":0.4967,\"evidence_need\":0.7408,\"family\":\"performance_tuning\",\"input_tokens\":3760,\"reversibility\":0.795,\"risk\":0.3428,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.564,\"id\":\"worker_a\",\"p95_latency_ms\":54210,\"pass_rate\":0.9245,\"provider_group\":\"provider_y\",\"quality\":0.9208,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.889,\"id\":\"worker_b\",\"p95_latency_ms\":23355,\"pass_rate\":0.5431,\"provider_group\":\"provider_z\",\"quality\":0.6499,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.201,\"id\":\"worker_c\",\"p95_latency_ms\":51844,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.889,\"estimated_latency_ms\":23355,\"estimated_pass_rate\":0.3936,\"estimated_quality\":0.5299,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.177,\"max_latency_ms\":46169,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3508, "category": "analysis", "complexity": 0.8395, "consequence": 0.4967, "evidence_need": 0.7408, "family": "performance_tuning", "input_tokens": 3760, "request_id": "synthetic-performance_tuning-000299", "reversibility": 0.795, "risk": 0.3428, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 3.564, "id": "worker_a", "p95_latency_ms": 54210, "pass_rate": 0.9245, "provider_group": "provider_y", "quality": 0.9208, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.889, "id": "worker_b", "p95_latency_ms": 23355, "pass_rate": 0.5431, "provider_group": "provider_z", "quality": 0.6499, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.201, "id": "worker_c", "p95_latency_ms": 51844, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.374, "max_latency_ms": 269006, "minimum_quality": 0.8698}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001125", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.471, "estimated_latency_ms": 195702, "estimated_pass_rate": 0.7338, "estimated_quality": 0.8922, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.374, "max_latency_ms": 269006, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7338, "quality": 0.8922, "quality_target_met": true, "utility": 73.267947}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.374,\"max_latency_ms\":269006,\"minimum_quality\":0.8698},\"task\":{\"ambiguity\":0.3648,\"category\":\"review\",\"complexity\":0.7635,\"consequence\":0.6636,\"evidence_need\":0.8729,\"family\":\"policy_compliance\",\"input_tokens\":7400,\"reversibility\":0.6841,\"risk\":0.617,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.065,\"id\":\"worker_a\",\"p95_latency_ms\":26514,\"pass_rate\":0.4547,\"provider_group\":\"provider_y\",\"quality\":0.7041,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.554,\"id\":\"worker_b\",\"p95_latency_ms\":82096,\"pass_rate\":0.8585,\"provider_group\":\"provider_x\",\"quality\":0.8774,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.852,\"id\":\"worker_c\",\"p95_latency_ms\":87092,\"pass_rate\":0.9606,\"provider_group\":\"provider_z\",\"quality\":0.922,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.471,\"estimated_latency_ms\":195702,\"estimated_pass_rate\":0.7338,\"estimated_quality\":0.8922,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.374,\"max_latency_ms\":269006,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3648, "category": "review", "complexity": 0.7635, "consequence": 0.6636, "evidence_need": 0.8729, "family": "policy_compliance", "input_tokens": 7400, "request_id": "synthetic-policy_compliance-001125", "reversibility": 0.6841, "risk": 0.617, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.065, "id": "worker_a", "p95_latency_ms": 26514, "pass_rate": 0.4547, "provider_group": "provider_y", "quality": 0.7041, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.554, "id": "worker_b", "p95_latency_ms": 82096, "pass_rate": 0.8585, "provider_group": "provider_x", "quality": 0.8774, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.852, "id": "worker_c", "p95_latency_ms": 87092, "pass_rate": 0.9606, "provider_group": "provider_z", "quality": 0.922, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.085, "max_latency_ms": 172111, "minimum_quality": 0.8983}, "family": "migration_planning", "id": "synthetic-migration_planning-000358", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.287, "estimated_latency_ms": 119010, "estimated_pass_rate": 0.7313, "estimated_quality": 0.9052, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.085, "max_latency_ms": 172111, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7313, "quality": 0.9052, "quality_target_met": true, "utility": 72.541308}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.085,\"max_latency_ms\":172111,\"minimum_quality\":0.8983},\"task\":{\"ambiguity\":0.548,\"category\":\"coding\",\"complexity\":0.5778,\"consequence\":0.4507,\"evidence_need\":0.6258,\"family\":\"migration_planning\",\"input_tokens\":4280,\"reversibility\":0.5204,\"risk\":0.7074,\"summary\":\"Plan a zero-downtime migration of provider failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":10.417,\"id\":\"worker_a\",\"p95_latency_ms\":55967,\"pass_rate\":0.9189,\"provider_group\":\"provider_x\",\"quality\":0.9714,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.852,\"id\":\"worker_b\",\"p95_latency_ms\":17897,\"pass_rate\":0.4891,\"provider_group\":\"provider_y\",\"quality\":0.7216,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.018,\"id\":\"worker_c\",\"p95_latency_ms\":45146,\"pass_rate\":0.8488,\"provider_group\":\"provider_z\",\"quality\":0.8482,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.287,\"estimated_latency_ms\":119010,\"estimated_pass_rate\":0.7313,\"estimated_quality\":0.9052,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.085,\"max_latency_ms\":172111,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.548, "category": "coding", "complexity": 0.5778, "consequence": 0.4507, "evidence_need": 0.6258, "family": "migration_planning", "input_tokens": 4280, "request_id": "synthetic-migration_planning-000358", "reversibility": 0.5204, "risk": 0.7074, "summary": "Plan a zero-downtime migration of provider failover with rollback checkpoints."}, "workers": [{"credits_per_call": 10.417, "id": "worker_a", "p95_latency_ms": 55967, "pass_rate": 0.9189, "provider_group": "provider_x", "quality": 0.9714, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.852, "id": "worker_b", "p95_latency_ms": 17897, "pass_rate": 0.4891, "provider_group": "provider_y", "quality": 0.7216, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.018, "id": "worker_c", "p95_latency_ms": 45146, "pass_rate": 0.8488, "provider_group": "provider_z", "quality": 0.8482, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.03, "max_latency_ms": 177820, "minimum_quality": 0.7804}, "family": "migration_planning", "id": "synthetic-migration_planning-000541", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.939, "estimated_latency_ms": 132999, "estimated_pass_rate": 0.7345, "estimated_quality": 0.917, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.03, "max_latency_ms": 177820, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7345, "quality": 0.917, "quality_target_met": true, "utility": 70.480216}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.03,\"max_latency_ms\":177820,\"minimum_quality\":0.7804},\"task\":{\"ambiguity\":0.3343,\"category\":\"coding\",\"complexity\":0.8099,\"consequence\":0.4777,\"evidence_need\":0.6651,\"family\":\"migration_planning\",\"input_tokens\":7280,\"reversibility\":0.4346,\"risk\":0.3565,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.195,\"id\":\"worker_a\",\"p95_latency_ms\":24678,\"pass_rate\":0.5386,\"provider_group\":\"provider_z\",\"quality\":0.7106,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.441,\"id\":\"worker_b\",\"p95_latency_ms\":56136,\"pass_rate\":0.8387,\"provider_group\":\"provider_y\",\"quality\":0.8875,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.303,\"id\":\"worker_c\",\"p95_latency_ms\":52185,\"pass_rate\":0.9379,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.939,\"estimated_latency_ms\":132999,\"estimated_pass_rate\":0.7345,\"estimated_quality\":0.917,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.03,\"max_latency_ms\":177820,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3343, "category": "coding", "complexity": 0.8099, "consequence": 0.4777, "evidence_need": 0.6651, "family": "migration_planning", "input_tokens": 7280, "request_id": "synthetic-migration_planning-000541", "reversibility": 0.4346, "risk": 0.3565, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 1.195, "id": "worker_a", "p95_latency_ms": 24678, "pass_rate": 0.5386, "provider_group": "provider_z", "quality": 0.7106, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.441, "id": "worker_b", "p95_latency_ms": 56136, "pass_rate": 0.8387, "provider_group": "provider_y", "quality": 0.8875, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.303, "id": "worker_c", "p95_latency_ms": 52185, "pass_rate": 0.9379, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.135, "max_latency_ms": 45917, "minimum_quality": 0.7251}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000818", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.09, "estimated_latency_ms": 32416, "estimated_pass_rate": 0.382, "estimated_quality": 0.5998, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.135, "max_latency_ms": 45917, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.382, "quality": 0.5998, "quality_target_met": false, "utility": 35.100326}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.135,\"max_latency_ms\":45917,\"minimum_quality\":0.7251},\"task\":{\"ambiguity\":0.3354,\"category\":\"analysis\",\"complexity\":0.8307,\"consequence\":0.4333,\"evidence_need\":0.5308,\"family\":\"performance_tuning\",\"input_tokens\":3720,\"reversibility\":0.8077,\"risk\":0.5073,\"summary\":\"Diagnose latency variance in usage metering and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":12.047,\"id\":\"worker_a\",\"p95_latency_ms\":71868,\"pass_rate\":0.9784,\"provider_group\":\"provider_x\",\"quality\":0.9913,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.264,\"id\":\"worker_b\",\"p95_latency_ms\":81599,\"pass_rate\":0.8961,\"provider_group\":\"provider_z\",\"quality\":0.8737,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.09,\"id\":\"worker_c\",\"p95_latency_ms\":32416,\"pass_rate\":0.5426,\"provider_group\":\"provider_y\",\"quality\":0.7291,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.09,\"estimated_latency_ms\":32416,\"estimated_pass_rate\":0.382,\"estimated_quality\":0.5998,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.135,\"max_latency_ms\":45917,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3354, "category": "analysis", "complexity": 0.8307, "consequence": 0.4333, "evidence_need": 0.5308, "family": "performance_tuning", "input_tokens": 3720, "request_id": "synthetic-performance_tuning-000818", "reversibility": 0.8077, "risk": 0.5073, "summary": "Diagnose latency variance in usage metering and rank the next experiments."}, "workers": [{"credits_per_call": 12.047, "id": "worker_a", "p95_latency_ms": 71868, "pass_rate": 0.9784, "provider_group": "provider_x", "quality": 0.9913, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.264, "id": "worker_b", "p95_latency_ms": 81599, "pass_rate": 0.8961, "provider_group": "provider_z", "quality": 0.8737, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.09, "id": "worker_c", "p95_latency_ms": 32416, "pass_rate": 0.5426, "provider_group": "provider_y", "quality": 0.7291, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.264, "max_latency_ms": 39773, "minimum_quality": 0.7763}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000510", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.055, "estimated_latency_ms": 26815, "estimated_pass_rate": 0.3916, "estimated_quality": 0.7311, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.264, "max_latency_ms": 39773, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3916, "quality": 0.7311, "quality_target_met": false, "utility": 34.751024}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.264,\"max_latency_ms\":39773,\"minimum_quality\":0.7763},\"task\":{\"ambiguity\":0.4061,\"category\":\"review\",\"complexity\":0.4075,\"consequence\":0.7107,\"evidence_need\":0.8455,\"family\":\"policy_compliance\",\"input_tokens\":5640,\"reversibility\":0.5754,\"risk\":0.6691,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.264,\"id\":\"worker_a\",\"p95_latency_ms\":46659,\"pass_rate\":0.8049,\"provider_group\":\"provider_y\",\"quality\":0.8885,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.919,\"id\":\"worker_b\",\"p95_latency_ms\":69282,\"pass_rate\":0.9662,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.055,\"id\":\"worker_c\",\"p95_latency_ms\":26815,\"pass_rate\":0.5227,\"provider_group\":\"provider_x\",\"quality\":0.8002,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.055,\"estimated_latency_ms\":26815,\"estimated_pass_rate\":0.3916,\"estimated_quality\":0.7311,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.264,\"max_latency_ms\":39773,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4061, "category": "review", "complexity": 0.4075, "consequence": 0.7107, "evidence_need": 0.8455, "family": "policy_compliance", "input_tokens": 5640, "request_id": "synthetic-policy_compliance-000510", "reversibility": 0.5754, "risk": 0.6691, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.264, "id": "worker_a", "p95_latency_ms": 46659, "pass_rate": 0.8049, "provider_group": "provider_y", "quality": 0.8885, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.919, "id": "worker_b", "p95_latency_ms": 69282, "pass_rate": 0.9662, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.055, "id": "worker_c", "p95_latency_ms": 26815, "pass_rate": 0.5227, "provider_group": "provider_x", "quality": 0.8002, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.479, "max_latency_ms": 34672, "minimum_quality": 0.7671}, "family": "migration_planning", "id": "synthetic-migration_planning-000439", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.075, "estimated_latency_ms": 31674, "estimated_pass_rate": 0.3838, "estimated_quality": 0.6242, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.479, "max_latency_ms": 34672, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3838, "quality": 0.6242, "quality_target_met": false, "utility": 34.379605}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.479,\"max_latency_ms\":34672,\"minimum_quality\":0.7671},\"task\":{\"ambiguity\":0.7014,\"category\":\"coding\",\"complexity\":0.8367,\"consequence\":0.7915,\"evidence_need\":0.8262,\"family\":\"migration_planning\",\"input_tokens\":4200,\"reversibility\":0.419,\"risk\":0.4557,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.075,\"id\":\"worker_a\",\"p95_latency_ms\":31674,\"pass_rate\":0.5686,\"provider_group\":\"provider_y\",\"quality\":0.7754,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.306,\"id\":\"worker_b\",\"p95_latency_ms\":48936,\"pass_rate\":0.7883,\"provider_group\":\"provider_z\",\"quality\":0.8241,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.629,\"id\":\"worker_c\",\"p95_latency_ms\":58668,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9381,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.075,\"estimated_latency_ms\":31674,\"estimated_pass_rate\":0.3838,\"estimated_quality\":0.6242,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.479,\"max_latency_ms\":34672,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7014, "category": "coding", "complexity": 0.8367, "consequence": 0.7915, "evidence_need": 0.8262, "family": "migration_planning", "input_tokens": 4200, "request_id": "synthetic-migration_planning-000439", "reversibility": 0.419, "risk": 0.4557, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 1.075, "id": "worker_a", "p95_latency_ms": 31674, "pass_rate": 0.5686, "provider_group": "provider_y", "quality": 0.7754, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.306, "id": "worker_b", "p95_latency_ms": 48936, "pass_rate": 0.7883, "provider_group": "provider_z", "quality": 0.8241, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.629, "id": "worker_c", "p95_latency_ms": 58668, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9381, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.46, "max_latency_ms": 31547, "minimum_quality": 0.7198}, "family": "migration_planning", "id": "synthetic-migration_planning-000394", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.156, "estimated_latency_ms": 19736, "estimated_pass_rate": 0.4416, "estimated_quality": 0.6588, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.46, "max_latency_ms": 31547, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4416, "quality": 0.6588, "quality_target_met": false, "utility": 43.286444}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.46,\"max_latency_ms\":31547,\"minimum_quality\":0.7198},\"task\":{\"ambiguity\":0.662,\"category\":\"coding\",\"complexity\":0.5909,\"consequence\":0.5797,\"evidence_need\":0.5362,\"family\":\"migration_planning\",\"input_tokens\":4840,\"reversibility\":0.5129,\"risk\":0.4568,\"summary\":\"Plan a zero-downtime migration of schema evolution with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.176,\"id\":\"worker_a\",\"p95_latency_ms\":49862,\"pass_rate\":0.8416,\"provider_group\":\"provider_x\",\"quality\":0.8282,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.699,\"id\":\"worker_b\",\"p95_latency_ms\":56442,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.156,\"id\":\"worker_c\",\"p95_latency_ms\":19736,\"pass_rate\":0.5957,\"provider_group\":\"provider_y\",\"quality\":0.763,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.156,\"estimated_latency_ms\":19736,\"estimated_pass_rate\":0.4416,\"estimated_quality\":0.6588,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.46,\"max_latency_ms\":31547,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.662, "category": "coding", "complexity": 0.5909, "consequence": 0.5797, "evidence_need": 0.5362, "family": "migration_planning", "input_tokens": 4840, "request_id": "synthetic-migration_planning-000394", "reversibility": 0.5129, "risk": 0.4568, "summary": "Plan a zero-downtime migration of schema evolution with rollback checkpoints."}, "workers": [{"credits_per_call": 3.176, "id": "worker_a", "p95_latency_ms": 49862, "pass_rate": 0.8416, "provider_group": "provider_x", "quality": 0.8282, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.699, "id": "worker_b", "p95_latency_ms": 56442, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.156, "id": "worker_c", "p95_latency_ms": 19736, "pass_rate": 0.5957, "provider_group": "provider_y", "quality": 0.763, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.309, "max_latency_ms": 169270, "minimum_quality": 0.8458}, "family": "migration_planning", "id": "synthetic-migration_planning-001027", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.912, "estimated_latency_ms": 126954, "estimated_pass_rate": 0.7102, "estimated_quality": 0.9053, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.309, "max_latency_ms": 169270, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7102, "quality": 0.9053, "quality_target_met": true, "utility": 69.997391}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.309,\"max_latency_ms\":169270,\"minimum_quality\":0.8458},\"task\":{\"ambiguity\":0.3022,\"category\":\"coding\",\"complexity\":0.9389,\"consequence\":0.401,\"evidence_need\":0.4905,\"family\":\"migration_planning\",\"input_tokens\":7920,\"reversibility\":0.6226,\"risk\":0.4351,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.989,\"id\":\"worker_a\",\"p95_latency_ms\":19785,\"pass_rate\":0.509,\"provider_group\":\"provider_y\",\"quality\":0.7406,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.8,\"id\":\"worker_b\",\"p95_latency_ms\":55002,\"pass_rate\":0.8043,\"provider_group\":\"provider_x\",\"quality\":0.8622,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.123,\"id\":\"worker_c\",\"p95_latency_ms\":52167,\"pass_rate\":0.9309,\"provider_group\":\"provider_z\",\"quality\":0.9901,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.912,\"estimated_latency_ms\":126954,\"estimated_pass_rate\":0.7102,\"estimated_quality\":0.9053,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.309,\"max_latency_ms\":169270,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3022, "category": "coding", "complexity": 0.9389, "consequence": 0.401, "evidence_need": 0.4905, "family": "migration_planning", "input_tokens": 7920, "request_id": "synthetic-migration_planning-001027", "reversibility": 0.6226, "risk": 0.4351, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 0.989, "id": "worker_a", "p95_latency_ms": 19785, "pass_rate": 0.509, "provider_group": "provider_y", "quality": 0.7406, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.8, "id": "worker_b", "p95_latency_ms": 55002, "pass_rate": 0.8043, "provider_group": "provider_x", "quality": 0.8622, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.123, "id": "worker_c", "p95_latency_ms": 52167, "pass_rate": 0.9309, "provider_group": "provider_z", "quality": 0.9901, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.568, "max_latency_ms": 27380, "minimum_quality": 0.648}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000135", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.085, "estimated_latency_ms": 25085, "estimated_pass_rate": 0.3008, "estimated_quality": 0.6137, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.568, "max_latency_ms": 27380, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3008, "quality": 0.6137, "quality_target_met": false, "utility": 20.749135}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.568,\"max_latency_ms\":27380,\"minimum_quality\":0.648},\"task\":{\"ambiguity\":0.5537,\"category\":\"review\",\"complexity\":0.8051,\"consequence\":0.7161,\"evidence_need\":0.9413,\"family\":\"policy_compliance\",\"input_tokens\":2280,\"reversibility\":0.4821,\"risk\":0.8466,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.805,\"id\":\"worker_a\",\"p95_latency_ms\":48271,\"pass_rate\":0.7652,\"provider_group\":\"provider_x\",\"quality\":0.8888,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.485,\"id\":\"worker_b\",\"p95_latency_ms\":56851,\"pass_rate\":0.9868,\"provider_group\":\"provider_z\",\"quality\":0.9744,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.085,\"id\":\"worker_c\",\"p95_latency_ms\":25085,\"pass_rate\":0.5025,\"provider_group\":\"provider_y\",\"quality\":0.7774,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.085,\"estimated_latency_ms\":25085,\"estimated_pass_rate\":0.3008,\"estimated_quality\":0.6137,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.568,\"max_latency_ms\":27380,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5537, "category": "review", "complexity": 0.8051, "consequence": 0.7161, "evidence_need": 0.9413, "family": "policy_compliance", "input_tokens": 2280, "request_id": "synthetic-policy_compliance-000135", "reversibility": 0.4821, "risk": 0.8466, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.805, "id": "worker_a", "p95_latency_ms": 48271, "pass_rate": 0.7652, "provider_group": "provider_x", "quality": 0.8888, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.485, "id": "worker_b", "p95_latency_ms": 56851, "pass_rate": 0.9868, "provider_group": "provider_z", "quality": 0.9744, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.085, "id": "worker_c", "p95_latency_ms": 25085, "pass_rate": 0.5025, "provider_group": "provider_y", "quality": 0.7774, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.797, "max_latency_ms": 32597, "minimum_quality": 0.578}, "family": "migration_planning", "id": "synthetic-migration_planning-000607", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.03, "estimated_latency_ms": 18105, "estimated_pass_rate": 0.3478, "estimated_quality": 0.6291, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.797, "max_latency_ms": 32597, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3478, "quality": 0.6291, "quality_target_met": true, "utility": 40.491971}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.797,\"max_latency_ms\":32597,\"minimum_quality\":0.578},\"task\":{\"ambiguity\":0.7137,\"category\":\"coding\",\"complexity\":0.587,\"consequence\":0.5253,\"evidence_need\":0.4231,\"family\":\"migration_planning\",\"input_tokens\":2640,\"reversibility\":0.6306,\"risk\":0.5634,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.795,\"id\":\"worker_a\",\"p95_latency_ms\":71035,\"pass_rate\":0.7388,\"provider_group\":\"provider_x\",\"quality\":0.8689,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.969,\"id\":\"worker_b\",\"p95_latency_ms\":66782,\"pass_rate\":0.9115,\"provider_group\":\"provider_z\",\"quality\":0.9718,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.03,\"id\":\"worker_c\",\"p95_latency_ms\":18105,\"pass_rate\":0.514,\"provider_group\":\"provider_y\",\"quality\":0.7438,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.03,\"estimated_latency_ms\":18105,\"estimated_pass_rate\":0.3478,\"estimated_quality\":0.6291,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.797,\"max_latency_ms\":32597,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7137, "category": "coding", "complexity": 0.587, "consequence": 0.5253, "evidence_need": 0.4231, "family": "migration_planning", "input_tokens": 2640, "request_id": "synthetic-migration_planning-000607", "reversibility": 0.6306, "risk": 0.5634, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.795, "id": "worker_a", "p95_latency_ms": 71035, "pass_rate": 0.7388, "provider_group": "provider_x", "quality": 0.8689, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 14.969, "id": "worker_b", "p95_latency_ms": 66782, "pass_rate": 0.9115, "provider_group": "provider_z", "quality": 0.9718, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.03, "id": "worker_c", "p95_latency_ms": 18105, "pass_rate": 0.514, "provider_group": "provider_y", "quality": 0.7438, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.105, "max_latency_ms": 177625, "minimum_quality": 0.8975}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000284", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.199, "estimated_latency_ms": 145225, "estimated_pass_rate": 0.7892, "estimated_quality": 0.9134, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.105, "max_latency_ms": 177625, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7892, "quality": 0.9134, "quality_target_met": true, "utility": 73.909038}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.105,\"max_latency_ms\":177625,\"minimum_quality\":0.8975},\"task\":{\"ambiguity\":0.6594,\"category\":\"analysis\",\"complexity\":0.7219,\"consequence\":0.6302,\"evidence_need\":0.5311,\"family\":\"performance_tuning\",\"input_tokens\":6800,\"reversibility\":0.8461,\"risk\":0.5523,\"summary\":\"Review a proposed optimization for tenant-scoped caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.79,\"id\":\"worker_a\",\"p95_latency_ms\":47738,\"pass_rate\":0.9166,\"provider_group\":\"provider_x\",\"quality\":0.9622,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.321,\"id\":\"worker_b\",\"p95_latency_ms\":76880,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.088,\"id\":\"worker_c\",\"p95_latency_ms\":20607,\"pass_rate\":0.5248,\"provider_group\":\"provider_z\",\"quality\":0.6228,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.199,\"estimated_latency_ms\":145225,\"estimated_pass_rate\":0.7892,\"estimated_quality\":0.9134,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.105,\"max_latency_ms\":177625,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6594, "category": "analysis", "complexity": 0.7219, "consequence": 0.6302, "evidence_need": 0.5311, "family": "performance_tuning", "input_tokens": 6800, "request_id": "synthetic-performance_tuning-000284", "reversibility": 0.8461, "risk": 0.5523, "summary": "Review a proposed optimization for tenant-scoped caching for hidden regressions."}, "workers": [{"credits_per_call": 3.79, "id": "worker_a", "p95_latency_ms": 47738, "pass_rate": 0.9166, "provider_group": "provider_x", "quality": 0.9622, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.321, "id": "worker_b", "p95_latency_ms": 76880, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.088, "id": "worker_c", "p95_latency_ms": 20607, "pass_rate": 0.5248, "provider_group": "provider_z", "quality": 0.6228, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.48, "max_latency_ms": 21522, "minimum_quality": 0.5275}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000852", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.967, "estimated_latency_ms": 17761, "estimated_pass_rate": 0.4173, "estimated_quality": 0.5667, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.48, "max_latency_ms": 21522, "policy": "lean", "reason_codes": ["independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4173, "quality": 0.5667, "quality_target_met": true, "utility": 24.513102}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.48,\"max_latency_ms\":21522,\"minimum_quality\":0.5275},\"task\":{\"ambiguity\":0.6784,\"category\":\"review\",\"complexity\":0.6822,\"consequence\":0.6477,\"evidence_need\":0.8151,\"family\":\"policy_compliance\",\"input_tokens\":920,\"reversibility\":0.6652,\"risk\":0.7177,\"summary\":\"Identify evidence required before approving artifact storage for production.\"},\"workers\":[{\"credits_per_call\":12.76,\"id\":\"worker_a\",\"p95_latency_ms\":84190,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.967,\"id\":\"worker_b\",\"p95_latency_ms\":17761,\"pass_rate\":0.6041,\"provider_group\":\"provider_z\",\"quality\":0.7073,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.973,\"id\":\"worker_c\",\"p95_latency_ms\":57067,\"pass_rate\":0.8868,\"provider_group\":\"provider_x\",\"quality\":0.7939,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.967,\"estimated_latency_ms\":17761,\"estimated_pass_rate\":0.4173,\"estimated_quality\":0.5667,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.48,\"max_latency_ms\":21522,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6784, "category": "review", "complexity": 0.6822, "consequence": 0.6477, "evidence_need": 0.8151, "family": "policy_compliance", "input_tokens": 920, "request_id": "synthetic-policy_compliance-000852", "reversibility": 0.6652, "risk": 0.7177, "summary": "Identify evidence required before approving artifact storage for production."}, "workers": [{"credits_per_call": 12.76, "id": "worker_a", "p95_latency_ms": 84190, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.967, "id": "worker_b", "p95_latency_ms": 17761, "pass_rate": 0.6041, "provider_group": "provider_z", "quality": 0.7073, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.973, "id": "worker_c", "p95_latency_ms": 57067, "pass_rate": 0.8868, "provider_group": "provider_x", "quality": 0.7939, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.426, "max_latency_ms": 210838, "minimum_quality": 0.8292}, "family": "migration_planning", "id": "synthetic-migration_planning-000376", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.192, "estimated_latency_ms": 162092, "estimated_pass_rate": 0.7634, "estimated_quality": 0.9047, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.426, "max_latency_ms": 210838, "policy": "strict", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7634, "quality": 0.9047, "quality_target_met": true, "utility": 72.823904}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.426,\"max_latency_ms\":210838,\"minimum_quality\":0.8292},\"task\":{\"ambiguity\":0.5459,\"category\":\"coding\",\"complexity\":0.6035,\"consequence\":0.4281,\"evidence_need\":0.64,\"family\":\"migration_planning\",\"input_tokens\":2960,\"reversibility\":0.3432,\"risk\":0.7426,\"summary\":\"Plan a zero-downtime migration of usage metering with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.926,\"id\":\"worker_a\",\"p95_latency_ms\":83692,\"pass_rate\":0.9692,\"provider_group\":\"provider_x\",\"quality\":0.9506,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.079,\"id\":\"worker_b\",\"p95_latency_ms\":23112,\"pass_rate\":0.554,\"provider_group\":\"provider_y\",\"quality\":0.7872,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.187,\"id\":\"worker_c\",\"p95_latency_ms\":55288,\"pass_rate\":0.8442,\"provider_group\":\"provider_z\",\"quality\":0.7972,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.192,\"estimated_latency_ms\":162092,\"estimated_pass_rate\":0.7634,\"estimated_quality\":0.9047,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.426,\"max_latency_ms\":210838,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5459, "category": "coding", "complexity": 0.6035, "consequence": 0.4281, "evidence_need": 0.64, "family": "migration_planning", "input_tokens": 2960, "request_id": "synthetic-migration_planning-000376", "reversibility": 0.3432, "risk": 0.7426, "summary": "Plan a zero-downtime migration of usage metering with rollback checkpoints."}, "workers": [{"credits_per_call": 14.926, "id": "worker_a", "p95_latency_ms": 83692, "pass_rate": 0.9692, "provider_group": "provider_x", "quality": 0.9506, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.079, "id": "worker_b", "p95_latency_ms": 23112, "pass_rate": 0.554, "provider_group": "provider_y", "quality": 0.7872, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.187, "id": "worker_c", "p95_latency_ms": 55288, "pass_rate": 0.8442, "provider_group": "provider_z", "quality": 0.7972, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.864, "max_latency_ms": 192217, "minimum_quality": 0.7138}, "family": "migration_planning", "id": "synthetic-migration_planning-000628", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.664, "estimated_latency_ms": 92198, "estimated_pass_rate": 0.7734, "estimated_quality": 0.8549, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.864, "max_latency_ms": 192217, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7734, "quality": 0.8549, "quality_target_met": true, "utility": 65.256834}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.864,\"max_latency_ms\":192217,\"minimum_quality\":0.7138},\"task\":{\"ambiguity\":0.4253,\"category\":\"coding\",\"complexity\":0.7133,\"consequence\":0.716,\"evidence_need\":0.7864,\"family\":\"migration_planning\",\"input_tokens\":4400,\"reversibility\":0.3626,\"risk\":0.3895,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.868,\"id\":\"worker_a\",\"p95_latency_ms\":73745,\"pass_rate\":0.8613,\"provider_group\":\"provider_z\",\"quality\":0.9113,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.796,\"id\":\"worker_b\",\"p95_latency_ms\":18453,\"pass_rate\":0.5766,\"provider_group\":\"provider_y\",\"quality\":0.7569,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.115,\"id\":\"worker_c\",\"p95_latency_ms\":78644,\"pass_rate\":0.9792,\"provider_group\":\"provider_x\",\"quality\":0.9781,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.664,\"estimated_latency_ms\":92198,\"estimated_pass_rate\":0.7734,\"estimated_quality\":0.8549,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.864,\"max_latency_ms\":192217,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4253, "category": "coding", "complexity": 0.7133, "consequence": 0.716, "evidence_need": 0.7864, "family": "migration_planning", "input_tokens": 4400, "request_id": "synthetic-migration_planning-000628", "reversibility": 0.3626, "risk": 0.3895, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 2.868, "id": "worker_a", "p95_latency_ms": 73745, "pass_rate": 0.8613, "provider_group": "provider_z", "quality": 0.9113, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.796, "id": "worker_b", "p95_latency_ms": 18453, "pass_rate": 0.5766, "provider_group": "provider_y", "quality": 0.7569, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.115, "id": "worker_c", "p95_latency_ms": 78644, "pass_rate": 0.9792, "provider_group": "provider_x", "quality": 0.9781, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.913, "max_latency_ms": 44906, "minimum_quality": 0.728}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000351", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.792, "estimated_latency_ms": 32589, "estimated_pass_rate": 0.3171, "estimated_quality": 0.5332, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.913, "max_latency_ms": 44906, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3171, "quality": 0.5332, "quality_target_met": false, "utility": 7.865247}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.913,\"max_latency_ms\":44906,\"minimum_quality\":0.728},\"task\":{\"ambiguity\":0.7148,\"category\":\"review\",\"complexity\":0.8487,\"consequence\":0.933,\"evidence_need\":0.8171,\"family\":\"policy_compliance\",\"input_tokens\":6520,\"reversibility\":0.4493,\"risk\":0.7982,\"summary\":\"Review whether private credentials satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.792,\"id\":\"worker_a\",\"p95_latency_ms\":32589,\"pass_rate\":0.532,\"provider_group\":\"provider_x\",\"quality\":0.7121,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.883,\"id\":\"worker_b\",\"p95_latency_ms\":71644,\"pass_rate\":0.9651,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.917,\"id\":\"worker_c\",\"p95_latency_ms\":72426,\"pass_rate\":0.8406,\"provider_group\":\"provider_y\",\"quality\":0.8466,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.792,\"estimated_latency_ms\":32589,\"estimated_pass_rate\":0.3171,\"estimated_quality\":0.5332,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.913,\"max_latency_ms\":44906,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7148, "category": "review", "complexity": 0.8487, "consequence": 0.933, "evidence_need": 0.8171, "family": "policy_compliance", "input_tokens": 6520, "request_id": "synthetic-policy_compliance-000351", "reversibility": 0.4493, "risk": 0.7982, "summary": "Review whether private credentials satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.792, "id": "worker_a", "p95_latency_ms": 32589, "pass_rate": 0.532, "provider_group": "provider_x", "quality": 0.7121, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.883, "id": "worker_b", "p95_latency_ms": 71644, "pass_rate": 0.9651, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.917, "id": "worker_c", "p95_latency_ms": 72426, "pass_rate": 0.8406, "provider_group": "provider_y", "quality": 0.8466, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.828, "max_latency_ms": 37778, "minimum_quality": 0.7283}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000171", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.143, "estimated_latency_ms": 32621, "estimated_pass_rate": 0.4328, "estimated_quality": 0.6172, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.828, "max_latency_ms": 37778, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4328, "quality": 0.6172, "quality_target_met": false, "utility": 21.225483}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.828,\"max_latency_ms\":37778,\"minimum_quality\":0.7283},\"task\":{\"ambiguity\":0.4407,\"category\":\"review\",\"complexity\":0.7278,\"consequence\":0.705,\"evidence_need\":0.8489,\"family\":\"policy_compliance\",\"input_tokens\":4240,\"reversibility\":0.4171,\"risk\":0.8487,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.181,\"id\":\"worker_a\",\"p95_latency_ms\":80571,\"pass_rate\":0.85,\"provider_group\":\"provider_y\",\"quality\":0.8536,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.143,\"id\":\"worker_b\",\"p95_latency_ms\":32621,\"pass_rate\":0.6174,\"provider_group\":\"provider_z\",\"quality\":0.7595,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.525,\"id\":\"worker_c\",\"p95_latency_ms\":64324,\"pass_rate\":0.9645,\"provider_group\":\"provider_x\",\"quality\":0.9102,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.143,\"estimated_latency_ms\":32621,\"estimated_pass_rate\":0.4328,\"estimated_quality\":0.6172,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.828,\"max_latency_ms\":37778,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4407, "category": "review", "complexity": 0.7278, "consequence": 0.705, "evidence_need": 0.8489, "family": "policy_compliance", "input_tokens": 4240, "request_id": "synthetic-policy_compliance-000171", "reversibility": 0.4171, "risk": 0.8487, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.181, "id": "worker_a", "p95_latency_ms": 80571, "pass_rate": 0.85, "provider_group": "provider_y", "quality": 0.8536, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.143, "id": "worker_b", "p95_latency_ms": 32621, "pass_rate": 0.6174, "provider_group": "provider_z", "quality": 0.7595, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.525, "id": "worker_c", "p95_latency_ms": 64324, "pass_rate": 0.9645, "provider_group": "provider_x", "quality": 0.9102, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.741, "max_latency_ms": 26495, "minimum_quality": 0.5276}, "family": "migration_planning", "id": "synthetic-migration_planning-000859", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.101, "estimated_latency_ms": 23486, "estimated_pass_rate": 0.4388, "estimated_quality": 0.584, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.741, "max_latency_ms": 26495, "policy": "lean", "reason_codes": ["latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4388, "quality": 0.584, "quality_target_met": true, "utility": 40.251314}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.741,\"max_latency_ms\":26495,\"minimum_quality\":0.5276},\"task\":{\"ambiguity\":0.4001,\"category\":\"coding\",\"complexity\":0.6325,\"consequence\":0.4092,\"evidence_need\":0.5513,\"family\":\"migration_planning\",\"input_tokens\":2880,\"reversibility\":0.5013,\"risk\":0.4147,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.223,\"id\":\"worker_a\",\"p95_latency_ms\":67820,\"pass_rate\":0.7586,\"provider_group\":\"provider_x\",\"quality\":0.8878,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.101,\"id\":\"worker_b\",\"p95_latency_ms\":23486,\"pass_rate\":0.5744,\"provider_group\":\"provider_z\",\"quality\":0.6751,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.168,\"id\":\"worker_c\",\"p95_latency_ms\":88863,\"pass_rate\":0.9025,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.101,\"estimated_latency_ms\":23486,\"estimated_pass_rate\":0.4388,\"estimated_quality\":0.584,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.741,\"max_latency_ms\":26495,\"policy\":\"lean\",\"reason_codes\":[\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4001, "category": "coding", "complexity": 0.6325, "consequence": 0.4092, "evidence_need": 0.5513, "family": "migration_planning", "input_tokens": 2880, "request_id": "synthetic-migration_planning-000859", "reversibility": 0.5013, "risk": 0.4147, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 3.223, "id": "worker_a", "p95_latency_ms": 67820, "pass_rate": 0.7586, "provider_group": "provider_x", "quality": 0.8878, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.101, "id": "worker_b", "p95_latency_ms": 23486, "pass_rate": 0.5744, "provider_group": "provider_z", "quality": 0.6751, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.168, "id": "worker_c", "p95_latency_ms": 88863, "pass_rate": 0.9025, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.944, "max_latency_ms": 215414, "minimum_quality": 0.9204}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000114", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.439, "estimated_latency_ms": 176387, "estimated_pass_rate": 0.7206, "estimated_quality": 0.9085, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.944, "max_latency_ms": 215414, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7206, "quality": 0.9085, "quality_target_met": false, "utility": 69.802323}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.944,\"max_latency_ms\":215414,\"minimum_quality\":0.9204},\"task\":{\"ambiguity\":0.7656,\"category\":\"review\",\"complexity\":0.5865,\"consequence\":0.5787,\"evidence_need\":0.5774,\"family\":\"policy_compliance\",\"input_tokens\":7040,\"reversibility\":0.4251,\"risk\":0.518,\"summary\":\"Review whether schema evolution satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.05,\"id\":\"worker_a\",\"p95_latency_ms\":29583,\"pass_rate\":0.4918,\"provider_group\":\"provider_y\",\"quality\":0.7861,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.292,\"id\":\"worker_b\",\"p95_latency_ms\":65031,\"pass_rate\":0.7694,\"provider_group\":\"provider_z\",\"quality\":0.819,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.097,\"id\":\"worker_c\",\"p95_latency_ms\":81773,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9657,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.439,\"estimated_latency_ms\":176387,\"estimated_pass_rate\":0.7206,\"estimated_quality\":0.9085,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.944,\"max_latency_ms\":215414,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7656, "category": "review", "complexity": 0.5865, "consequence": 0.5787, "evidence_need": 0.5774, "family": "policy_compliance", "input_tokens": 7040, "request_id": "synthetic-policy_compliance-000114", "reversibility": 0.4251, "risk": 0.518, "summary": "Review whether schema evolution satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.05, "id": "worker_a", "p95_latency_ms": 29583, "pass_rate": 0.4918, "provider_group": "provider_y", "quality": 0.7861, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.292, "id": "worker_b", "p95_latency_ms": 65031, "pass_rate": 0.7694, "provider_group": "provider_z", "quality": 0.819, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.097, "id": "worker_c", "p95_latency_ms": 81773, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9657, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.003, "max_latency_ms": 196298, "minimum_quality": 0.8193}, "family": "migration_planning", "id": "synthetic-migration_planning-000934", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.603, "estimated_latency_ms": 154240, "estimated_pass_rate": 0.7556, "estimated_quality": 0.9252, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.003, "max_latency_ms": 196298, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7556, "quality": 0.9252, "quality_target_met": true, "utility": 70.489329}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.003,\"max_latency_ms\":196298,\"minimum_quality\":0.8193},\"task\":{\"ambiguity\":0.6423,\"category\":\"coding\",\"complexity\":0.6379,\"consequence\":0.6074,\"evidence_need\":0.551,\"family\":\"migration_planning\",\"input_tokens\":4480,\"reversibility\":0.2702,\"risk\":0.3965,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":0.825,\"id\":\"worker_a\",\"p95_latency_ms\":32152,\"pass_rate\":0.5728,\"provider_group\":\"provider_x\",\"quality\":0.7851,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.957,\"id\":\"worker_b\",\"p95_latency_ms\":44848,\"pass_rate\":0.7944,\"provider_group\":\"provider_y\",\"quality\":0.8329,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.821,\"id\":\"worker_c\",\"p95_latency_ms\":77240,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.603,\"estimated_latency_ms\":154240,\"estimated_pass_rate\":0.7556,\"estimated_quality\":0.9252,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.003,\"max_latency_ms\":196298,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6423, "category": "coding", "complexity": 0.6379, "consequence": 0.6074, "evidence_need": 0.551, "family": "migration_planning", "input_tokens": 4480, "request_id": "synthetic-migration_planning-000934", "reversibility": 0.2702, "risk": 0.3965, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 0.825, "id": "worker_a", "p95_latency_ms": 32152, "pass_rate": 0.5728, "provider_group": "provider_x", "quality": 0.7851, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.957, "id": "worker_b", "p95_latency_ms": 44848, "pass_rate": 0.7944, "provider_group": "provider_y", "quality": 0.8329, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.821, "id": "worker_c", "p95_latency_ms": 77240, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.997, "max_latency_ms": 30074, "minimum_quality": 0.5689}, "family": "migration_planning", "id": "synthetic-migration_planning-000790", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.899, "estimated_latency_ms": 21508, "estimated_pass_rate": 0.4089, "estimated_quality": 0.5546, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.997, "max_latency_ms": 30074, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4089, "quality": 0.5546, "quality_target_met": false, "utility": 36.819812}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.997,\"max_latency_ms\":30074,\"minimum_quality\":0.5689},\"task\":{\"ambiguity\":0.6954,\"category\":\"coding\",\"complexity\":0.6979,\"consequence\":0.7425,\"evidence_need\":0.6355,\"family\":\"migration_planning\",\"input_tokens\":7680,\"reversibility\":0.5863,\"risk\":0.542,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":15.111,\"id\":\"worker_a\",\"p95_latency_ms\":88292,\"pass_rate\":0.9586,\"provider_group\":\"provider_x\",\"quality\":0.9756,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.669,\"id\":\"worker_b\",\"p95_latency_ms\":82052,\"pass_rate\":0.8821,\"provider_group\":\"provider_z\",\"quality\":0.91,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.899,\"id\":\"worker_c\",\"p95_latency_ms\":21508,\"pass_rate\":0.5845,\"provider_group\":\"provider_y\",\"quality\":0.6865,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.899,\"estimated_latency_ms\":21508,\"estimated_pass_rate\":0.4089,\"estimated_quality\":0.5546,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.997,\"max_latency_ms\":30074,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6954, "category": "coding", "complexity": 0.6979, "consequence": 0.7425, "evidence_need": 0.6355, "family": "migration_planning", "input_tokens": 7680, "request_id": "synthetic-migration_planning-000790", "reversibility": 0.5863, "risk": 0.542, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 15.111, "id": "worker_a", "p95_latency_ms": 88292, "pass_rate": 0.9586, "provider_group": "provider_x", "quality": 0.9756, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.669, "id": "worker_b", "p95_latency_ms": 82052, "pass_rate": 0.8821, "provider_group": "provider_z", "quality": 0.91, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.899, "id": "worker_c", "p95_latency_ms": 21508, "pass_rate": 0.5845, "provider_group": "provider_y", "quality": 0.6865, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 22.946, "max_latency_ms": 107078, "minimum_quality": 0.828}, "family": "migration_planning", "id": "synthetic-migration_planning-000196", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.416, "estimated_latency_ms": 84584, "estimated_pass_rate": 0.7443, "estimated_quality": 0.8961, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 22.946, "max_latency_ms": 107078, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7443, "quality": 0.8961, "quality_target_met": true, "utility": 66.111948}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":22.946,\"max_latency_ms\":107078,\"minimum_quality\":0.828},\"task\":{\"ambiguity\":0.6619,\"category\":\"coding\",\"complexity\":0.7495,\"consequence\":0.4538,\"evidence_need\":0.8035,\"family\":\"migration_planning\",\"input_tokens\":6160,\"reversibility\":0.4689,\"risk\":0.6431,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.425,\"id\":\"worker_a\",\"p95_latency_ms\":59715,\"pass_rate\":0.7841,\"provider_group\":\"provider_x\",\"quality\":0.8047,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.337,\"id\":\"worker_b\",\"p95_latency_ms\":66280,\"pass_rate\":0.9909,\"provider_group\":\"provider_z\",\"quality\":0.9645,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.079,\"id\":\"worker_c\",\"p95_latency_ms\":18304,\"pass_rate\":0.6151,\"provider_group\":\"provider_y\",\"quality\":0.7806,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.416,\"estimated_latency_ms\":84584,\"estimated_pass_rate\":0.7443,\"estimated_quality\":0.8961,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":22.946,\"max_latency_ms\":107078,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6619, "category": "coding", "complexity": 0.7495, "consequence": 0.4538, "evidence_need": 0.8035, "family": "migration_planning", "input_tokens": 6160, "request_id": "synthetic-migration_planning-000196", "reversibility": 0.4689, "risk": 0.6431, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 3.425, "id": "worker_a", "p95_latency_ms": 59715, "pass_rate": 0.7841, "provider_group": "provider_x", "quality": 0.8047, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.337, "id": "worker_b", "p95_latency_ms": 66280, "pass_rate": 0.9909, "provider_group": "provider_z", "quality": 0.9645, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.079, "id": "worker_c", "p95_latency_ms": 18304, "pass_rate": 0.6151, "provider_group": "provider_y", "quality": 0.7806, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.513, "max_latency_ms": 190122, "minimum_quality": 0.8989}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000761", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.919, "estimated_latency_ms": 159475, "estimated_pass_rate": 0.7822, "estimated_quality": 0.9014, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.513, "max_latency_ms": 190122, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7822, "quality": 0.9014, "quality_target_met": true, "utility": 72.563078}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.513,\"max_latency_ms\":190122,\"minimum_quality\":0.8989},\"task\":{\"ambiguity\":0.6474,\"category\":\"analysis\",\"complexity\":0.807,\"consequence\":0.5037,\"evidence_need\":0.4296,\"family\":\"performance_tuning\",\"input_tokens\":4920,\"reversibility\":0.5594,\"risk\":0.5014,\"summary\":\"Choose between throughput and tail-latency changes for observability.\"},\"workers\":[{\"credits_per_call\":12.04,\"id\":\"worker_a\",\"p95_latency_ms\":80453,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.792,\"id\":\"worker_b\",\"p95_latency_ms\":17622,\"pass_rate\":0.5525,\"provider_group\":\"provider_x\",\"quality\":0.6133,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.087,\"id\":\"worker_c\",\"p95_latency_ms\":61400,\"pass_rate\":0.8516,\"provider_group\":\"provider_z\",\"quality\":0.9873,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.919,\"estimated_latency_ms\":159475,\"estimated_pass_rate\":0.7822,\"estimated_quality\":0.9014,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.513,\"max_latency_ms\":190122,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6474, "category": "analysis", "complexity": 0.807, "consequence": 0.5037, "evidence_need": 0.4296, "family": "performance_tuning", "input_tokens": 4920, "request_id": "synthetic-performance_tuning-000761", "reversibility": 0.5594, "risk": 0.5014, "summary": "Choose between throughput and tail-latency changes for observability."}, "workers": [{"credits_per_call": 12.04, "id": "worker_a", "p95_latency_ms": 80453, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.792, "id": "worker_b", "p95_latency_ms": 17622, "pass_rate": 0.5525, "provider_group": "provider_x", "quality": 0.6133, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.087, "id": "worker_c", "p95_latency_ms": 61400, "pass_rate": 0.8516, "provider_group": "provider_z", "quality": 0.9873, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.337, "max_latency_ms": 161802, "minimum_quality": 0.7618}, "family": "migration_planning", "id": "synthetic-migration_planning-000826", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.498, "estimated_latency_ms": 128920, "estimated_pass_rate": 0.8842, "estimated_quality": 0.9361, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.337, "max_latency_ms": 161802, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8842, "quality": 0.9361, "quality_target_met": true, "utility": 70.436267}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.337,\"max_latency_ms\":161802,\"minimum_quality\":0.7618},\"task\":{\"ambiguity\":0.3738,\"category\":\"coding\",\"complexity\":0.6939,\"consequence\":0.602,\"evidence_need\":0.6043,\"family\":\"migration_planning\",\"input_tokens\":6320,\"reversibility\":0.6431,\"risk\":0.5458,\"summary\":\"Review a staged migration for database failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.601,\"id\":\"worker_a\",\"p95_latency_ms\":77943,\"pass_rate\":0.8736,\"provider_group\":\"provider_x\",\"quality\":0.8122,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.897,\"id\":\"worker_b\",\"p95_latency_ms\":50977,\"pass_rate\":0.964,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.891,\"id\":\"worker_c\",\"p95_latency_ms\":19559,\"pass_rate\":0.4625,\"provider_group\":\"provider_z\",\"quality\":0.7292,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.498,\"estimated_latency_ms\":128920,\"estimated_pass_rate\":0.8842,\"estimated_quality\":0.9361,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.337,\"max_latency_ms\":161802,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3738, "category": "coding", "complexity": 0.6939, "consequence": 0.602, "evidence_need": 0.6043, "family": "migration_planning", "input_tokens": 6320, "request_id": "synthetic-migration_planning-000826", "reversibility": 0.6431, "risk": 0.5458, "summary": "Review a staged migration for database failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.601, "id": "worker_a", "p95_latency_ms": 77943, "pass_rate": 0.8736, "provider_group": "provider_x", "quality": 0.8122, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.897, "id": "worker_b", "p95_latency_ms": 50977, "pass_rate": 0.964, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.891, "id": "worker_c", "p95_latency_ms": 19559, "pass_rate": 0.4625, "provider_group": "provider_z", "quality": 0.7292, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 0.923, "max_latency_ms": 47330, "minimum_quality": 0.7714}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000612", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.826, "estimated_latency_ms": 31598, "estimated_pass_rate": 0.3092, "estimated_quality": 0.6046, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 0.923, "max_latency_ms": 47330, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3092, "quality": 0.6046, "quality_target_met": false, "utility": 9.328387}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":0.923,\"max_latency_ms\":47330,\"minimum_quality\":0.7714},\"task\":{\"ambiguity\":0.5251,\"category\":\"review\",\"complexity\":0.8484,\"consequence\":0.9218,\"evidence_need\":0.7622,\"family\":\"policy_compliance\",\"input_tokens\":5040,\"reversibility\":0.5224,\"risk\":0.8241,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.893,\"id\":\"worker_a\",\"p95_latency_ms\":68348,\"pass_rate\":0.8417,\"provider_group\":\"provider_y\",\"quality\":0.8443,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.826,\"id\":\"worker_b\",\"p95_latency_ms\":31598,\"pass_rate\":0.5119,\"provider_group\":\"provider_x\",\"quality\":0.7726,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.639,\"id\":\"worker_c\",\"p95_latency_ms\":85657,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.826,\"estimated_latency_ms\":31598,\"estimated_pass_rate\":0.3092,\"estimated_quality\":0.6046,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":0.923,\"max_latency_ms\":47330,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5251, "category": "review", "complexity": 0.8484, "consequence": 0.9218, "evidence_need": 0.7622, "family": "policy_compliance", "input_tokens": 5040, "request_id": "synthetic-policy_compliance-000612", "reversibility": 0.5224, "risk": 0.8241, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.893, "id": "worker_a", "p95_latency_ms": 68348, "pass_rate": 0.8417, "provider_group": "provider_y", "quality": 0.8443, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.826, "id": "worker_b", "p95_latency_ms": 31598, "pass_rate": 0.5119, "provider_group": "provider_x", "quality": 0.7726, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.639, "id": "worker_c", "p95_latency_ms": 85657, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.544, "max_latency_ms": 216894, "minimum_quality": 0.8203}, "family": "migration_planning", "id": "synthetic-migration_planning-000682", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.912, "estimated_latency_ms": 137457, "estimated_pass_rate": 0.8561, "estimated_quality": 0.945, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.544, "max_latency_ms": 216894, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8561, "quality": 0.945, "quality_target_met": true, "utility": 70.458686}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.544,\"max_latency_ms\":216894,\"minimum_quality\":0.8203},\"task\":{\"ambiguity\":0.4888,\"category\":\"coding\",\"complexity\":0.6218,\"consequence\":0.448,\"evidence_need\":0.6508,\"family\":\"migration_planning\",\"input_tokens\":3840,\"reversibility\":0.2603,\"risk\":0.6501,\"summary\":\"Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.18,\"id\":\"worker_a\",\"p95_latency_ms\":31189,\"pass_rate\":0.4662,\"provider_group\":\"provider_y\",\"quality\":0.7207,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.743,\"id\":\"worker_b\",\"p95_latency_ms\":81930,\"pass_rate\":0.8827,\"provider_group\":\"provider_z\",\"quality\":0.8872,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.169,\"id\":\"worker_c\",\"p95_latency_ms\":55527,\"pass_rate\":0.8972,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.912,\"estimated_latency_ms\":137457,\"estimated_pass_rate\":0.8561,\"estimated_quality\":0.945,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.544,\"max_latency_ms\":216894,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4888, "category": "coding", "complexity": 0.6218, "consequence": 0.448, "evidence_need": 0.6508, "family": "migration_planning", "input_tokens": 3840, "request_id": "synthetic-migration_planning-000682", "reversibility": 0.2603, "risk": 0.6501, "summary": "Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints."}, "workers": [{"credits_per_call": 1.18, "id": "worker_a", "p95_latency_ms": 31189, "pass_rate": 0.4662, "provider_group": "provider_y", "quality": 0.7207, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.743, "id": "worker_b", "p95_latency_ms": 81930, "pass_rate": 0.8827, "provider_group": "provider_z", "quality": 0.8872, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 15.169, "id": "worker_c", "p95_latency_ms": 55527, "pass_rate": 0.8972, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.056, "max_latency_ms": 201203, "minimum_quality": 0.8226}, "family": "migration_planning", "id": "synthetic-migration_planning-000670", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.655, "estimated_latency_ms": 168815, "estimated_pass_rate": 0.7288, "estimated_quality": 0.8901, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.056, "max_latency_ms": 201203, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7288, "quality": 0.8901, "quality_target_met": true, "utility": 73.515399}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.056,\"max_latency_ms\":201203,\"minimum_quality\":0.8226},\"task\":{\"ambiguity\":0.3785,\"category\":\"coding\",\"complexity\":0.8828,\"consequence\":0.7956,\"evidence_need\":0.5043,\"family\":\"migration_planning\",\"input_tokens\":4680,\"reversibility\":0.4283,\"risk\":0.6413,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.114,\"id\":\"worker_a\",\"p95_latency_ms\":19762,\"pass_rate\":0.5258,\"provider_group\":\"provider_x\",\"quality\":0.7444,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.069,\"id\":\"worker_b\",\"p95_latency_ms\":80690,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9703,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.472,\"id\":\"worker_c\",\"p95_latency_ms\":68363,\"pass_rate\":0.7238,\"provider_group\":\"provider_y\",\"quality\":0.8083,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.655,\"estimated_latency_ms\":168815,\"estimated_pass_rate\":0.7288,\"estimated_quality\":0.8901,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.056,\"max_latency_ms\":201203,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3785, "category": "coding", "complexity": 0.8828, "consequence": 0.7956, "evidence_need": 0.5043, "family": "migration_planning", "input_tokens": 4680, "request_id": "synthetic-migration_planning-000670", "reversibility": 0.4283, "risk": 0.6413, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 1.114, "id": "worker_a", "p95_latency_ms": 19762, "pass_rate": 0.5258, "provider_group": "provider_x", "quality": 0.7444, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.069, "id": "worker_b", "p95_latency_ms": 80690, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9703, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.472, "id": "worker_c", "p95_latency_ms": 68363, "pass_rate": 0.7238, "provider_group": "provider_y", "quality": 0.8083, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.554, "max_latency_ms": 183990, "minimum_quality": 0.8695}, "family": "migration_planning", "id": "synthetic-migration_planning-001036", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.889, "estimated_latency_ms": 159447, "estimated_pass_rate": 0.7173, "estimated_quality": 0.8843, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.554, "max_latency_ms": 183990, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7173, "quality": 0.8843, "quality_target_met": true, "utility": 70.168527}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.554,\"max_latency_ms\":183990,\"minimum_quality\":0.8695},\"task\":{\"ambiguity\":0.4832,\"category\":\"coding\",\"complexity\":0.9179,\"consequence\":0.6981,\"evidence_need\":0.4504,\"family\":\"migration_planning\",\"input_tokens\":1720,\"reversibility\":0.4686,\"risk\":0.4265,\"summary\":\"Review a staged migration for artifact storage and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.876,\"id\":\"worker_a\",\"p95_latency_ms\":21296,\"pass_rate\":0.5162,\"provider_group\":\"provider_z\",\"quality\":0.729,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.235,\"id\":\"worker_b\",\"p95_latency_ms\":49145,\"pass_rate\":0.7575,\"provider_group\":\"provider_y\",\"quality\":0.8207,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.778,\"id\":\"worker_c\",\"p95_latency_ms\":89006,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9587,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.889,\"estimated_latency_ms\":159447,\"estimated_pass_rate\":0.7173,\"estimated_quality\":0.8843,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.554,\"max_latency_ms\":183990,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4832, "category": "coding", "complexity": 0.9179, "consequence": 0.6981, "evidence_need": 0.4504, "family": "migration_planning", "input_tokens": 1720, "request_id": "synthetic-migration_planning-001036", "reversibility": 0.4686, "risk": 0.4265, "summary": "Review a staged migration for artifact storage and identify irreversible steps."}, "workers": [{"credits_per_call": 0.876, "id": "worker_a", "p95_latency_ms": 21296, "pass_rate": 0.5162, "provider_group": "provider_z", "quality": 0.729, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.235, "id": "worker_b", "p95_latency_ms": 49145, "pass_rate": 0.7575, "provider_group": "provider_y", "quality": 0.8207, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.778, "id": "worker_c", "p95_latency_ms": 89006, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9587, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 24.449, "max_latency_ms": 244611, "minimum_quality": 0.8746}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000501", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.229, "estimated_latency_ms": 159935, "estimated_pass_rate": 0.8383, "estimated_quality": 0.9285, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 24.449, "max_latency_ms": 244611, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8383, "quality": 0.9285, "quality_target_met": true, "utility": 72.003307}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":24.449,\"max_latency_ms\":244611,\"minimum_quality\":0.8746},\"task\":{\"ambiguity\":0.4174,\"category\":\"review\",\"complexity\":0.8229,\"consequence\":0.7471,\"evidence_need\":0.8741,\"family\":\"policy_compliance\",\"input_tokens\":6840,\"reversibility\":0.6496,\"risk\":0.6034,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.124,\"id\":\"worker_a\",\"p95_latency_ms\":87484,\"pass_rate\":0.9116,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.105,\"id\":\"worker_b\",\"p95_latency_ms\":72451,\"pass_rate\":0.8267,\"provider_group\":\"provider_z\",\"quality\":0.8305,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.839,\"id\":\"worker_c\",\"p95_latency_ms\":20412,\"pass_rate\":0.4767,\"provider_group\":\"provider_x\",\"quality\":0.7409,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.229,\"estimated_latency_ms\":159935,\"estimated_pass_rate\":0.8383,\"estimated_quality\":0.9285,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":24.449,\"max_latency_ms\":244611,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4174, "category": "review", "complexity": 0.8229, "consequence": 0.7471, "evidence_need": 0.8741, "family": "policy_compliance", "input_tokens": 6840, "request_id": "synthetic-policy_compliance-000501", "reversibility": 0.6496, "risk": 0.6034, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 15.124, "id": "worker_a", "p95_latency_ms": 87484, "pass_rate": 0.9116, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.105, "id": "worker_b", "p95_latency_ms": 72451, "pass_rate": 0.8267, "provider_group": "provider_z", "quality": 0.8305, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.839, "id": "worker_c", "p95_latency_ms": 20412, "pass_rate": 0.4767, "provider_group": "provider_x", "quality": 0.7409, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.155, "max_latency_ms": 151869, "minimum_quality": 0.8908}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000945", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 10.896, "estimated_latency_ms": 72763, "estimated_pass_rate": 0.8022, "estimated_quality": 0.8774, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.155, "max_latency_ms": 151869, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8022, "quality": 0.8774, "quality_target_met": false, "utility": 64.083494}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.155,\"max_latency_ms\":151869,\"minimum_quality\":0.8908},\"task\":{\"ambiguity\":0.7646,\"category\":\"review\",\"complexity\":0.7905,\"consequence\":0.9482,\"evidence_need\":0.5912,\"family\":\"policy_compliance\",\"input_tokens\":4040,\"reversibility\":0.319,\"risk\":0.5079,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.905,\"id\":\"worker_a\",\"p95_latency_ms\":21730,\"pass_rate\":0.5514,\"provider_group\":\"provider_y\",\"quality\":0.7989,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.991,\"id\":\"worker_b\",\"p95_latency_ms\":51033,\"pass_rate\":0.9364,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.465,\"id\":\"worker_c\",\"p95_latency_ms\":76275,\"pass_rate\":0.8363,\"provider_group\":\"provider_z\",\"quality\":0.8504,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":10.896,\"estimated_latency_ms\":72763,\"estimated_pass_rate\":0.8022,\"estimated_quality\":0.8774,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.155,\"max_latency_ms\":151869,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7646, "category": "review", "complexity": 0.7905, "consequence": 0.9482, "evidence_need": 0.5912, "family": "policy_compliance", "input_tokens": 4040, "request_id": "synthetic-policy_compliance-000945", "reversibility": 0.319, "risk": 0.5079, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.905, "id": "worker_a", "p95_latency_ms": 21730, "pass_rate": 0.5514, "provider_group": "provider_y", "quality": 0.7989, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.991, "id": "worker_b", "p95_latency_ms": 51033, "pass_rate": 0.9364, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.465, "id": "worker_c", "p95_latency_ms": 76275, "pass_rate": 0.8363, "provider_group": "provider_z", "quality": 0.8504, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.809, "max_latency_ms": 121058, "minimum_quality": 0.7814}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000213", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.47, "estimated_latency_ms": 96141, "estimated_pass_rate": 0.6524, "estimated_quality": 0.8213, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.809, "max_latency_ms": 121058, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6524, "quality": 0.8213, "quality_target_met": true, "utility": 62.073172}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.809,\"max_latency_ms\":121058,\"minimum_quality\":0.7814},\"task\":{\"ambiguity\":0.4326,\"category\":\"review\",\"complexity\":0.7567,\"consequence\":0.5451,\"evidence_need\":0.7936,\"family\":\"policy_compliance\",\"input_tokens\":5240,\"reversibility\":0.6979,\"risk\":0.7434,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":0.919,\"id\":\"worker_a\",\"p95_latency_ms\":18482,\"pass_rate\":0.6102,\"provider_group\":\"provider_y\",\"quality\":0.7965,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.398,\"id\":\"worker_b\",\"p95_latency_ms\":70088,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.551,\"id\":\"worker_c\",\"p95_latency_ms\":77659,\"pass_rate\":0.7804,\"provider_group\":\"provider_x\",\"quality\":0.8164,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.47,\"estimated_latency_ms\":96141,\"estimated_pass_rate\":0.6524,\"estimated_quality\":0.8213,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.809,\"max_latency_ms\":121058,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4326, "category": "review", "complexity": 0.7567, "consequence": 0.5451, "evidence_need": 0.7936, "family": "policy_compliance", "input_tokens": 5240, "request_id": "synthetic-policy_compliance-000213", "reversibility": 0.6979, "risk": 0.7434, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 0.919, "id": "worker_a", "p95_latency_ms": 18482, "pass_rate": 0.6102, "provider_group": "provider_y", "quality": 0.7965, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.398, "id": "worker_b", "p95_latency_ms": 70088, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.551, "id": "worker_c", "p95_latency_ms": 77659, "pass_rate": 0.7804, "provider_group": "provider_x", "quality": 0.8164, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 0.959, "max_latency_ms": 47197, "minimum_quality": 0.6805}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000717", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.886, "estimated_latency_ms": 23638, "estimated_pass_rate": 0.3785, "estimated_quality": 0.69, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 0.959, "max_latency_ms": 47197, "policy": "lean", "reason_codes": ["credit_constrained", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3785, "quality": 0.69, "quality_target_met": true, "utility": 42.764581}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":0.959,\"max_latency_ms\":47197,\"minimum_quality\":0.6805},\"task\":{\"ambiguity\":0.5772,\"category\":\"review\",\"complexity\":0.566,\"consequence\":0.5283,\"evidence_need\":0.7798,\"family\":\"policy_compliance\",\"input_tokens\":6800,\"reversibility\":0.6121,\"risk\":0.5916,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.886,\"id\":\"worker_a\",\"p95_latency_ms\":23638,\"pass_rate\":0.5342,\"provider_group\":\"provider_y\",\"quality\":0.7937,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.032,\"id\":\"worker_b\",\"p95_latency_ms\":86057,\"pass_rate\":0.8878,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.094,\"id\":\"worker_c\",\"p95_latency_ms\":49247,\"pass_rate\":0.7878,\"provider_group\":\"provider_z\",\"quality\":0.8566,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.886,\"estimated_latency_ms\":23638,\"estimated_pass_rate\":0.3785,\"estimated_quality\":0.69,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":0.959,\"max_latency_ms\":47197,\"policy\":\"lean\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5772, "category": "review", "complexity": 0.566, "consequence": 0.5283, "evidence_need": 0.7798, "family": "policy_compliance", "input_tokens": 6800, "request_id": "synthetic-policy_compliance-000717", "reversibility": 0.6121, "risk": 0.5916, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 0.886, "id": "worker_a", "p95_latency_ms": 23638, "pass_rate": 0.5342, "provider_group": "provider_y", "quality": 0.7937, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.032, "id": "worker_b", "p95_latency_ms": 86057, "pass_rate": 0.8878, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.094, "id": "worker_c", "p95_latency_ms": 49247, "pass_rate": 0.7878, "provider_group": "provider_z", "quality": 0.8566, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.769, "max_latency_ms": 205957, "minimum_quality": 0.8369}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001014", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.197, "estimated_latency_ms": 180126, "estimated_pass_rate": 0.751, "estimated_quality": 0.8985, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.769, "max_latency_ms": 205957, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.751, "quality": 0.8985, "quality_target_met": true, "utility": 74.337257}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.769,\"max_latency_ms\":205957,\"minimum_quality\":0.8369},\"task\":{\"ambiguity\":0.6627,\"category\":\"review\",\"complexity\":0.6971,\"consequence\":0.9013,\"evidence_need\":0.6191,\"family\":\"policy_compliance\",\"input_tokens\":7040,\"reversibility\":0.6832,\"risk\":0.5823,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.353,\"id\":\"worker_a\",\"p95_latency_ms\":86413,\"pass_rate\":0.9048,\"provider_group\":\"provider_y\",\"quality\":0.9549,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.034,\"id\":\"worker_b\",\"p95_latency_ms\":63266,\"pass_rate\":0.8296,\"provider_group\":\"provider_x\",\"quality\":0.9218,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.81,\"id\":\"worker_c\",\"p95_latency_ms\":30447,\"pass_rate\":0.5781,\"provider_group\":\"provider_z\",\"quality\":0.7277,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.197,\"estimated_latency_ms\":180126,\"estimated_pass_rate\":0.751,\"estimated_quality\":0.8985,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.769,\"max_latency_ms\":205957,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6627, "category": "review", "complexity": 0.6971, "consequence": 0.9013, "evidence_need": 0.6191, "family": "policy_compliance", "input_tokens": 7040, "request_id": "synthetic-policy_compliance-001014", "reversibility": 0.6832, "risk": 0.5823, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 11.353, "id": "worker_a", "p95_latency_ms": 86413, "pass_rate": 0.9048, "provider_group": "provider_y", "quality": 0.9549, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.034, "id": "worker_b", "p95_latency_ms": 63266, "pass_rate": 0.8296, "provider_group": "provider_x", "quality": 0.9218, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.81, "id": "worker_c", "p95_latency_ms": 30447, "pass_rate": 0.5781, "provider_group": "provider_z", "quality": 0.7277, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.613, "max_latency_ms": 216822, "minimum_quality": 0.8054}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000024", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.34, "estimated_latency_ms": 186660, "estimated_pass_rate": 0.7787, "estimated_quality": 0.9064, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.613, "max_latency_ms": 216822, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7787, "quality": 0.9064, "quality_target_met": true, "utility": 77.003865}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.613,\"max_latency_ms\":216822,\"minimum_quality\":0.8054},\"task\":{\"ambiguity\":0.414,\"category\":\"review\",\"complexity\":0.8118,\"consequence\":0.7586,\"evidence_need\":0.58,\"family\":\"policy_compliance\",\"input_tokens\":2960,\"reversibility\":0.6741,\"risk\":0.8278,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.212,\"id\":\"worker_a\",\"p95_latency_ms\":66405,\"pass_rate\":0.8631,\"provider_group\":\"provider_x\",\"quality\":0.8226,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.134,\"id\":\"worker_b\",\"p95_latency_ms\":32622,\"pass_rate\":0.5414,\"provider_group\":\"provider_z\",\"quality\":0.7686,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.994,\"id\":\"worker_c\",\"p95_latency_ms\":87633,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9671,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.34,\"estimated_latency_ms\":186660,\"estimated_pass_rate\":0.7787,\"estimated_quality\":0.9064,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.613,\"max_latency_ms\":216822,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.414, "category": "review", "complexity": 0.8118, "consequence": 0.7586, "evidence_need": 0.58, "family": "policy_compliance", "input_tokens": 2960, "request_id": "synthetic-policy_compliance-000024", "reversibility": 0.6741, "risk": 0.8278, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 4.212, "id": "worker_a", "p95_latency_ms": 66405, "pass_rate": 0.8631, "provider_group": "provider_x", "quality": 0.8226, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.134, "id": "worker_b", "p95_latency_ms": 32622, "pass_rate": 0.5414, "provider_group": "provider_z", "quality": 0.7686, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.994, "id": "worker_c", "p95_latency_ms": 87633, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9671, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.344, "max_latency_ms": 40011, "minimum_quality": 0.7566}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000489", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.889, "estimated_latency_ms": 32392, "estimated_pass_rate": 0.278, "estimated_quality": 0.5517, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.344, "max_latency_ms": 40011, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.278, "quality": 0.5517, "quality_target_met": false, "utility": 10.702301}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.344,\"max_latency_ms\":40011,\"minimum_quality\":0.7566},\"task\":{\"ambiguity\":0.7967,\"category\":\"review\",\"complexity\":0.8087,\"consequence\":0.7311,\"evidence_need\":0.7044,\"family\":\"policy_compliance\",\"input_tokens\":5360,\"reversibility\":0.304,\"risk\":0.7335,\"summary\":\"Review whether webhook delivery satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.889,\"id\":\"worker_a\",\"p95_latency_ms\":32392,\"pass_rate\":0.4894,\"provider_group\":\"provider_x\",\"quality\":0.7243,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.798,\"id\":\"worker_b\",\"p95_latency_ms\":76773,\"pass_rate\":0.9577,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.471,\"id\":\"worker_c\",\"p95_latency_ms\":47490,\"pass_rate\":0.8429,\"provider_group\":\"provider_z\",\"quality\":0.831,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.889,\"estimated_latency_ms\":32392,\"estimated_pass_rate\":0.278,\"estimated_quality\":0.5517,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.344,\"max_latency_ms\":40011,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7967, "category": "review", "complexity": 0.8087, "consequence": 0.7311, "evidence_need": 0.7044, "family": "policy_compliance", "input_tokens": 5360, "request_id": "synthetic-policy_compliance-000489", "reversibility": 0.304, "risk": 0.7335, "summary": "Review whether webhook delivery satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.889, "id": "worker_a", "p95_latency_ms": 32392, "pass_rate": 0.4894, "provider_group": "provider_x", "quality": 0.7243, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.798, "id": "worker_b", "p95_latency_ms": 76773, "pass_rate": 0.9577, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.471, "id": "worker_c", "p95_latency_ms": 47490, "pass_rate": 0.8429, "provider_group": "provider_z", "quality": 0.831, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.894, "max_latency_ms": 171350, "minimum_quality": 0.8124}, "family": "migration_planning", "id": "synthetic-migration_planning-000061", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.678, "estimated_latency_ms": 124717, "estimated_pass_rate": 0.6848, "estimated_quality": 0.9219, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 19.894, "max_latency_ms": 171350, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6848, "quality": 0.9219, "quality_target_met": true, "utility": 72.792132}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.894,\"max_latency_ms\":171350,\"minimum_quality\":0.8124},\"task\":{\"ambiguity\":0.7171,\"category\":\"coding\",\"complexity\":0.7599,\"consequence\":0.5179,\"evidence_need\":0.7879,\"family\":\"migration_planning\",\"input_tokens\":2560,\"reversibility\":0.455,\"risk\":0.4162,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":13.152,\"id\":\"worker_a\",\"p95_latency_ms\":57160,\"pass_rate\":0.9067,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.713,\"id\":\"worker_b\",\"p95_latency_ms\":45207,\"pass_rate\":0.834,\"provider_group\":\"provider_y\",\"quality\":0.8188,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.813,\"id\":\"worker_c\",\"p95_latency_ms\":22350,\"pass_rate\":0.489,\"provider_group\":\"provider_z\",\"quality\":0.7697,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.678,\"estimated_latency_ms\":124717,\"estimated_pass_rate\":0.6848,\"estimated_quality\":0.9219,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.894,\"max_latency_ms\":171350,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7171, "category": "coding", "complexity": 0.7599, "consequence": 0.5179, "evidence_need": 0.7879, "family": "migration_planning", "input_tokens": 2560, "request_id": "synthetic-migration_planning-000061", "reversibility": 0.455, "risk": 0.4162, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 13.152, "id": "worker_a", "p95_latency_ms": 57160, "pass_rate": 0.9067, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.713, "id": "worker_b", "p95_latency_ms": 45207, "pass_rate": 0.834, "provider_group": "provider_y", "quality": 0.8188, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.813, "id": "worker_c", "p95_latency_ms": 22350, "pass_rate": 0.489, "provider_group": "provider_z", "quality": 0.7697, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.628, "max_latency_ms": 49176, "minimum_quality": 0.7439}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000530", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.793, "estimated_latency_ms": 27198, "estimated_pass_rate": 0.2648, "estimated_quality": 0.5474, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.628, "max_latency_ms": 49176, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2648, "quality": 0.5474, "quality_target_met": false, "utility": 27.175111}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.628,\"max_latency_ms\":49176,\"minimum_quality\":0.7439},\"task\":{\"ambiguity\":0.6495,\"category\":\"analysis\",\"complexity\":0.8132,\"consequence\":0.4851,\"evidence_need\":0.4635,\"family\":\"performance_tuning\",\"input_tokens\":4400,\"reversibility\":0.6466,\"risk\":0.3886,\"summary\":\"Choose between throughput and tail-latency changes for data retention.\"},\"workers\":[{\"credits_per_call\":0.793,\"id\":\"worker_a\",\"p95_latency_ms\":27198,\"pass_rate\":0.4376,\"provider_group\":\"provider_x\",\"quality\":0.6861,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.809,\"id\":\"worker_b\",\"p95_latency_ms\":55055,\"pass_rate\":0.871,\"provider_group\":\"provider_z\",\"quality\":0.9544,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.94,\"id\":\"worker_c\",\"p95_latency_ms\":74231,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.793,\"estimated_latency_ms\":27198,\"estimated_pass_rate\":0.2648,\"estimated_quality\":0.5474,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.628,\"max_latency_ms\":49176,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6495, "category": "analysis", "complexity": 0.8132, "consequence": 0.4851, "evidence_need": 0.4635, "family": "performance_tuning", "input_tokens": 4400, "request_id": "synthetic-performance_tuning-000530", "reversibility": 0.6466, "risk": 0.3886, "summary": "Choose between throughput and tail-latency changes for data retention."}, "workers": [{"credits_per_call": 0.793, "id": "worker_a", "p95_latency_ms": 27198, "pass_rate": 0.4376, "provider_group": "provider_x", "quality": 0.6861, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.809, "id": "worker_b", "p95_latency_ms": 55055, "pass_rate": 0.871, "provider_group": "provider_z", "quality": 0.9544, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.94, "id": "worker_c", "p95_latency_ms": 74231, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.638, "max_latency_ms": 200110, "minimum_quality": 0.7411}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000911", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.848, "estimated_latency_ms": 118400, "estimated_pass_rate": 0.8939, "estimated_quality": 0.9433, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.638, "max_latency_ms": 200110, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8939, "quality": 0.9433, "quality_target_met": true, "utility": 70.650373}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.638,\"max_latency_ms\":200110,\"minimum_quality\":0.7411},\"task\":{\"ambiguity\":0.5259,\"category\":\"analysis\",\"complexity\":0.6678,\"consequence\":0.6768,\"evidence_need\":0.487,\"family\":\"performance_tuning\",\"input_tokens\":680,\"reversibility\":0.5851,\"risk\":0.2553,\"summary\":\"Diagnose latency variance in observability and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":2.893,\"id\":\"worker_a\",\"p95_latency_ms\":61671,\"pass_rate\":0.8996,\"provider_group\":\"provider_y\",\"quality\":0.916,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.955,\"id\":\"worker_b\",\"p95_latency_ms\":56729,\"pass_rate\":0.9689,\"provider_group\":\"provider_x\",\"quality\":0.9923,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.832,\"id\":\"worker_c\",\"p95_latency_ms\":23826,\"pass_rate\":0.5196,\"provider_group\":\"provider_z\",\"quality\":0.638,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.848,\"estimated_latency_ms\":118400,\"estimated_pass_rate\":0.8939,\"estimated_quality\":0.9433,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.638,\"max_latency_ms\":200110,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5259, "category": "analysis", "complexity": 0.6678, "consequence": 0.6768, "evidence_need": 0.487, "family": "performance_tuning", "input_tokens": 680, "request_id": "synthetic-performance_tuning-000911", "reversibility": 0.5851, "risk": 0.2553, "summary": "Diagnose latency variance in observability and rank the next experiments."}, "workers": [{"credits_per_call": 2.893, "id": "worker_a", "p95_latency_ms": 61671, "pass_rate": 0.8996, "provider_group": "provider_y", "quality": 0.916, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.955, "id": "worker_b", "p95_latency_ms": 56729, "pass_rate": 0.9689, "provider_group": "provider_x", "quality": 0.9923, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.832, "id": "worker_c", "p95_latency_ms": 23826, "pass_rate": 0.5196, "provider_group": "provider_z", "quality": 0.638, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.739, "max_latency_ms": 93774, "minimum_quality": 0.8937}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000930", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.852, "estimated_latency_ms": 85381, "estimated_pass_rate": 0.7369, "estimated_quality": 0.908, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.739, "max_latency_ms": 93774, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7369, "quality": 0.908, "quality_target_met": true, "utility": 68.993966}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.739,\"max_latency_ms\":93774,\"minimum_quality\":0.8937},\"task\":{\"ambiguity\":0.5106,\"category\":\"review\",\"complexity\":0.409,\"consequence\":0.803,\"evidence_need\":0.8467,\"family\":\"policy_compliance\",\"input_tokens\":4720,\"reversibility\":0.6028,\"risk\":0.8424,\"summary\":\"Identify evidence required before approving tenant-scoped caching for production.\"},\"workers\":[{\"credits_per_call\":3.585,\"id\":\"worker_a\",\"p95_latency_ms\":65734,\"pass_rate\":0.8658,\"provider_group\":\"provider_z\",\"quality\":0.8637,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.826,\"id\":\"worker_b\",\"p95_latency_ms\":61400,\"pass_rate\":0.8885,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.026,\"id\":\"worker_c\",\"p95_latency_ms\":23981,\"pass_rate\":0.5392,\"provider_group\":\"provider_x\",\"quality\":0.769,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.852,\"estimated_latency_ms\":85381,\"estimated_pass_rate\":0.7369,\"estimated_quality\":0.908,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.739,\"max_latency_ms\":93774,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5106, "category": "review", "complexity": 0.409, "consequence": 0.803, "evidence_need": 0.8467, "family": "policy_compliance", "input_tokens": 4720, "request_id": "synthetic-policy_compliance-000930", "reversibility": 0.6028, "risk": 0.8424, "summary": "Identify evidence required before approving tenant-scoped caching for production."}, "workers": [{"credits_per_call": 3.585, "id": "worker_a", "p95_latency_ms": 65734, "pass_rate": 0.8658, "provider_group": "provider_z", "quality": 0.8637, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 13.826, "id": "worker_b", "p95_latency_ms": 61400, "pass_rate": 0.8885, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.026, "id": "worker_c", "p95_latency_ms": 23981, "pass_rate": 0.5392, "provider_group": "provider_x", "quality": 0.769, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.69, "max_latency_ms": 201679, "minimum_quality": 0.7538}, "family": "migration_planning", "id": "synthetic-migration_planning-001024", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.401, "estimated_latency_ms": 77403, "estimated_pass_rate": 0.693, "estimated_quality": 0.8051, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.69, "max_latency_ms": 201679, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.693, "quality": 0.8051, "quality_target_met": true, "utility": 60.841998}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.69,\"max_latency_ms\":201679,\"minimum_quality\":0.7538},\"task\":{\"ambiguity\":0.6299,\"category\":\"coding\",\"complexity\":0.9129,\"consequence\":0.527,\"evidence_need\":0.6447,\"family\":\"migration_planning\",\"input_tokens\":240,\"reversibility\":0.2714,\"risk\":0.7012,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.559,\"id\":\"worker_a\",\"p95_latency_ms\":49384,\"pass_rate\":0.8028,\"provider_group\":\"provider_y\",\"quality\":0.917,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.842,\"id\":\"worker_b\",\"p95_latency_ms\":28019,\"pass_rate\":0.6097,\"provider_group\":\"provider_z\",\"quality\":0.6952,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.292,\"id\":\"worker_c\",\"p95_latency_ms\":92031,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.401,\"estimated_latency_ms\":77403,\"estimated_pass_rate\":0.693,\"estimated_quality\":0.8051,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.69,\"max_latency_ms\":201679,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6299, "category": "coding", "complexity": 0.9129, "consequence": 0.527, "evidence_need": 0.6447, "family": "migration_planning", "input_tokens": 240, "request_id": "synthetic-migration_planning-001024", "reversibility": 0.2714, "risk": 0.7012, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 3.559, "id": "worker_a", "p95_latency_ms": 49384, "pass_rate": 0.8028, "provider_group": "provider_y", "quality": 0.917, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.842, "id": "worker_b", "p95_latency_ms": 28019, "pass_rate": 0.6097, "provider_group": "provider_z", "quality": 0.6952, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.292, "id": "worker_c", "p95_latency_ms": 92031, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.459, "max_latency_ms": 197539, "minimum_quality": 0.9148}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001226", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.705, "estimated_latency_ms": 139392, "estimated_pass_rate": 0.8009, "estimated_quality": 0.9261, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 24.459, "max_latency_ms": 197539, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8009, "quality": 0.9261, "quality_target_met": true, "utility": 75.284159}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.459,\"max_latency_ms\":197539,\"minimum_quality\":0.9148},\"task\":{\"ambiguity\":0.5032,\"category\":\"analysis\",\"complexity\":0.7065,\"consequence\":0.6166,\"evidence_need\":0.6832,\"family\":\"performance_tuning\",\"input_tokens\":2360,\"reversibility\":0.5784,\"risk\":0.498,\"summary\":\"Choose between throughput and tail-latency changes for usage metering.\"},\"workers\":[{\"credits_per_call\":14.49,\"id\":\"worker_a\",\"p95_latency_ms\":49971,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.908,\"id\":\"worker_b\",\"p95_latency_ms\":28234,\"pass_rate\":0.5351,\"provider_group\":\"provider_z\",\"quality\":0.6162,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.307,\"id\":\"worker_c\",\"p95_latency_ms\":61187,\"pass_rate\":0.9483,\"provider_group\":\"provider_x\",\"quality\":0.9758,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.705,\"estimated_latency_ms\":139392,\"estimated_pass_rate\":0.8009,\"estimated_quality\":0.9261,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.459,\"max_latency_ms\":197539,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5032, "category": "analysis", "complexity": 0.7065, "consequence": 0.6166, "evidence_need": 0.6832, "family": "performance_tuning", "input_tokens": 2360, "request_id": "synthetic-performance_tuning-001226", "reversibility": 0.5784, "risk": 0.498, "summary": "Choose between throughput and tail-latency changes for usage metering."}, "workers": [{"credits_per_call": 14.49, "id": "worker_a", "p95_latency_ms": 49971, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.908, "id": "worker_b", "p95_latency_ms": 28234, "pass_rate": 0.5351, "provider_group": "provider_z", "quality": 0.6162, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.307, "id": "worker_c", "p95_latency_ms": 61187, "pass_rate": 0.9483, "provider_group": "provider_x", "quality": 0.9758, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.645, "max_latency_ms": 60922, "minimum_quality": 0.5373}, "family": "migration_planning", "id": "synthetic-migration_planning-000688", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 3.486, "estimated_latency_ms": 46333, "estimated_pass_rate": 0.5932, "estimated_quality": 0.7582, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.645, "max_latency_ms": 60922, "policy": "lean", "reason_codes": ["complex_task", "credit_constrained", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 2, "pass_rate": 0.5932, "quality": 0.7582, "quality_target_met": true, "utility": 50.534965}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.645,\"max_latency_ms\":60922,\"minimum_quality\":0.5373},\"task\":{\"ambiguity\":0.4397,\"category\":\"coding\",\"complexity\":0.8369,\"consequence\":0.7885,\"evidence_need\":0.5195,\"family\":\"migration_planning\",\"input_tokens\":2720,\"reversibility\":0.5682,\"risk\":0.5575,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.149,\"id\":\"worker_a\",\"p95_latency_ms\":30871,\"pass_rate\":0.5142,\"provider_group\":\"provider_y\",\"quality\":0.7722,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.672,\"id\":\"worker_b\",\"p95_latency_ms\":55011,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9296,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.486,\"id\":\"worker_c\",\"p95_latency_ms\":46333,\"pass_rate\":0.7665,\"provider_group\":\"provider_z\",\"quality\":0.8992,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":3.486,\"estimated_latency_ms\":46333,\"estimated_pass_rate\":0.5932,\"estimated_quality\":0.7582,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.645,\"max_latency_ms\":60922,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4397, "category": "coding", "complexity": 0.8369, "consequence": 0.7885, "evidence_need": 0.5195, "family": "migration_planning", "input_tokens": 2720, "request_id": "synthetic-migration_planning-000688", "reversibility": 0.5682, "risk": 0.5575, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 1.149, "id": "worker_a", "p95_latency_ms": 30871, "pass_rate": 0.5142, "provider_group": "provider_y", "quality": 0.7722, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.672, "id": "worker_b", "p95_latency_ms": 55011, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9296, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.486, "id": "worker_c", "p95_latency_ms": 46333, "pass_rate": 0.7665, "provider_group": "provider_z", "quality": 0.8992, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.143, "max_latency_ms": 234946, "minimum_quality": 0.7849}, "family": "migration_planning", "id": "synthetic-migration_planning-000469", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.543, "estimated_latency_ms": 82507, "estimated_pass_rate": 0.6258, "estimated_quality": 0.8046, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.143, "max_latency_ms": 234946, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6258, "quality": 0.8046, "quality_target_met": true, "utility": 57.493185}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.143,\"max_latency_ms\":234946,\"minimum_quality\":0.7849},\"task\":{\"ambiguity\":0.6747,\"category\":\"coding\",\"complexity\":0.7155,\"consequence\":0.6321,\"evidence_need\":0.4715,\"family\":\"migration_planning\",\"input_tokens\":1080,\"reversibility\":0.5726,\"risk\":0.6064,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":1.087,\"id\":\"worker_a\",\"p95_latency_ms\":18163,\"pass_rate\":0.5024,\"provider_group\":\"provider_z\",\"quality\":0.6958,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.456,\"id\":\"worker_b\",\"p95_latency_ms\":64344,\"pass_rate\":0.7215,\"provider_group\":\"provider_y\",\"quality\":0.9093,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.039,\"id\":\"worker_c\",\"p95_latency_ms\":78000,\"pass_rate\":0.9616,\"provider_group\":\"provider_x\",\"quality\":0.9208,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.543,\"estimated_latency_ms\":82507,\"estimated_pass_rate\":0.6258,\"estimated_quality\":0.8046,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.143,\"max_latency_ms\":234946,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6747, "category": "coding", "complexity": 0.7155, "consequence": 0.6321, "evidence_need": 0.4715, "family": "migration_planning", "input_tokens": 1080, "request_id": "synthetic-migration_planning-000469", "reversibility": 0.5726, "risk": 0.6064, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 1.087, "id": "worker_a", "p95_latency_ms": 18163, "pass_rate": 0.5024, "provider_group": "provider_z", "quality": 0.6958, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.456, "id": "worker_b", "p95_latency_ms": 64344, "pass_rate": 0.7215, "provider_group": "provider_y", "quality": 0.9093, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.039, "id": "worker_c", "p95_latency_ms": 78000, "pass_rate": 0.9616, "provider_group": "provider_x", "quality": 0.9208, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.908, "max_latency_ms": 180160, "minimum_quality": 0.9221}, "family": "migration_planning", "id": "synthetic-migration_planning-000580", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.972, "estimated_latency_ms": 170447, "estimated_pass_rate": 0.776, "estimated_quality": 0.9158, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.908, "max_latency_ms": 180160, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.776, "quality": 0.9158, "quality_target_met": false, "utility": 77.00935}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.908,\"max_latency_ms\":180160,\"minimum_quality\":0.9221},\"task\":{\"ambiguity\":0.3144,\"category\":\"coding\",\"complexity\":0.8298,\"consequence\":0.8274,\"evidence_need\":0.8091,\"family\":\"migration_planning\",\"input_tokens\":6960,\"reversibility\":0.2915,\"risk\":0.635,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":14.808,\"id\":\"worker_a\",\"p95_latency_ms\":64343,\"pass_rate\":0.9607,\"provider_group\":\"provider_z\",\"quality\":0.9723,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.155,\"id\":\"worker_b\",\"p95_latency_ms\":26635,\"pass_rate\":0.6143,\"provider_group\":\"provider_x\",\"quality\":0.7772,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.009,\"id\":\"worker_c\",\"p95_latency_ms\":79469,\"pass_rate\":0.7909,\"provider_group\":\"provider_y\",\"quality\":0.864,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.972,\"estimated_latency_ms\":170447,\"estimated_pass_rate\":0.776,\"estimated_quality\":0.9158,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.908,\"max_latency_ms\":180160,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3144, "category": "coding", "complexity": 0.8298, "consequence": 0.8274, "evidence_need": 0.8091, "family": "migration_planning", "input_tokens": 6960, "request_id": "synthetic-migration_planning-000580", "reversibility": 0.2915, "risk": 0.635, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 14.808, "id": "worker_a", "p95_latency_ms": 64343, "pass_rate": 0.9607, "provider_group": "provider_z", "quality": 0.9723, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.155, "id": "worker_b", "p95_latency_ms": 26635, "pass_rate": 0.6143, "provider_group": "provider_x", "quality": 0.7772, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.009, "id": "worker_c", "p95_latency_ms": 79469, "pass_rate": 0.7909, "provider_group": "provider_y", "quality": 0.864, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.337, "max_latency_ms": 211520, "minimum_quality": 0.8762}, "family": "migration_planning", "id": "synthetic-migration_planning-000043", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.986, "estimated_latency_ms": 173620, "estimated_pass_rate": 0.7459, "estimated_quality": 0.8868, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.337, "max_latency_ms": 211520, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7459, "quality": 0.8868, "quality_target_met": true, "utility": 69.866534}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.337,\"max_latency_ms\":211520,\"minimum_quality\":0.8762},\"task\":{\"ambiguity\":0.5554,\"category\":\"coding\",\"complexity\":0.7037,\"consequence\":0.4695,\"evidence_need\":0.5869,\"family\":\"migration_planning\",\"input_tokens\":6160,\"reversibility\":0.5615,\"risk\":0.5863,\"summary\":\"Plan a zero-downtime migration of rate limiting with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.028,\"id\":\"worker_a\",\"p95_latency_ms\":26900,\"pass_rate\":0.5298,\"provider_group\":\"provider_x\",\"quality\":0.7146,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.507,\"id\":\"worker_b\",\"p95_latency_ms\":64591,\"pass_rate\":0.947,\"provider_group\":\"provider_y\",\"quality\":0.938,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.451,\"id\":\"worker_c\",\"p95_latency_ms\":82129,\"pass_rate\":0.8105,\"provider_group\":\"provider_z\",\"quality\":0.9057,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.986,\"estimated_latency_ms\":173620,\"estimated_pass_rate\":0.7459,\"estimated_quality\":0.8868,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.337,\"max_latency_ms\":211520,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5554, "category": "coding", "complexity": 0.7037, "consequence": 0.4695, "evidence_need": 0.5869, "family": "migration_planning", "input_tokens": 6160, "request_id": "synthetic-migration_planning-000043", "reversibility": 0.5615, "risk": 0.5863, "summary": "Plan a zero-downtime migration of rate limiting with rollback checkpoints."}, "workers": [{"credits_per_call": 1.028, "id": "worker_a", "p95_latency_ms": 26900, "pass_rate": 0.5298, "provider_group": "provider_x", "quality": 0.7146, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.507, "id": "worker_b", "p95_latency_ms": 64591, "pass_rate": 0.947, "provider_group": "provider_y", "quality": 0.938, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.451, "id": "worker_c", "p95_latency_ms": 82129, "pass_rate": 0.8105, "provider_group": "provider_z", "quality": 0.9057, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.581, "max_latency_ms": 50922, "minimum_quality": 0.7176}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000377", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.986, "estimated_latency_ms": 28573, "estimated_pass_rate": 0.3382, "estimated_quality": 0.5086, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.581, "max_latency_ms": 50922, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3382, "quality": 0.5086, "quality_target_met": false, "utility": 23.6432}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.581,\"max_latency_ms\":50922,\"minimum_quality\":0.7176},\"task\":{\"ambiguity\":0.5759,\"category\":\"analysis\",\"complexity\":0.7537,\"consequence\":0.6786,\"evidence_need\":0.3618,\"family\":\"performance_tuning\",\"input_tokens\":2480,\"reversibility\":0.8831,\"risk\":0.3113,\"summary\":\"Diagnose latency variance in credit reservation and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":4.037,\"id\":\"worker_a\",\"p95_latency_ms\":67545,\"pass_rate\":0.837,\"provider_group\":\"provider_x\",\"quality\":0.8835,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.494,\"id\":\"worker_b\",\"p95_latency_ms\":51672,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.986,\"id\":\"worker_c\",\"p95_latency_ms\":28573,\"pass_rate\":0.4924,\"provider_group\":\"provider_y\",\"quality\":0.626,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.986,\"estimated_latency_ms\":28573,\"estimated_pass_rate\":0.3382,\"estimated_quality\":0.5086,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.581,\"max_latency_ms\":50922,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5759, "category": "analysis", "complexity": 0.7537, "consequence": 0.6786, "evidence_need": 0.3618, "family": "performance_tuning", "input_tokens": 2480, "request_id": "synthetic-performance_tuning-000377", "reversibility": 0.8831, "risk": 0.3113, "summary": "Diagnose latency variance in credit reservation and rank the next experiments."}, "workers": [{"credits_per_call": 4.037, "id": "worker_a", "p95_latency_ms": 67545, "pass_rate": 0.837, "provider_group": "provider_x", "quality": 0.8835, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.494, "id": "worker_b", "p95_latency_ms": 51672, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.986, "id": "worker_c", "p95_latency_ms": 28573, "pass_rate": 0.4924, "provider_group": "provider_y", "quality": 0.626, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.103, "max_latency_ms": 24260, "minimum_quality": 0.5292}, "family": "migration_planning", "id": "synthetic-migration_planning-000838", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.074, "estimated_latency_ms": 20559, "estimated_pass_rate": 0.3681, "estimated_quality": 0.596, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.103, "max_latency_ms": 24260, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3681, "quality": 0.596, "quality_target_met": true, "utility": 38.029907}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.103,\"max_latency_ms\":24260,\"minimum_quality\":0.5292},\"task\":{\"ambiguity\":0.7357,\"category\":\"coding\",\"complexity\":0.6624,\"consequence\":0.6027,\"evidence_need\":0.4646,\"family\":\"migration_planning\",\"input_tokens\":6080,\"reversibility\":0.3382,\"risk\":0.3502,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.074,\"id\":\"worker_a\",\"p95_latency_ms\":20559,\"pass_rate\":0.5272,\"provider_group\":\"provider_x\",\"quality\":0.7104,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.901,\"id\":\"worker_b\",\"p95_latency_ms\":57722,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9191,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.155,\"id\":\"worker_c\",\"p95_latency_ms\":67268,\"pass_rate\":0.7895,\"provider_group\":\"provider_z\",\"quality\":0.8972,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.074,\"estimated_latency_ms\":20559,\"estimated_pass_rate\":0.3681,\"estimated_quality\":0.596,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.103,\"max_latency_ms\":24260,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7357, "category": "coding", "complexity": 0.6624, "consequence": 0.6027, "evidence_need": 0.4646, "family": "migration_planning", "input_tokens": 6080, "request_id": "synthetic-migration_planning-000838", "reversibility": 0.3382, "risk": 0.3502, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 1.074, "id": "worker_a", "p95_latency_ms": 20559, "pass_rate": 0.5272, "provider_group": "provider_x", "quality": 0.7104, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.901, "id": "worker_b", "p95_latency_ms": 57722, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9191, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.155, "id": "worker_c", "p95_latency_ms": 67268, "pass_rate": 0.7895, "provider_group": "provider_z", "quality": 0.8972, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.424, "max_latency_ms": 188561, "minimum_quality": 0.9144}, "family": "migration_planning", "id": "synthetic-migration_planning-001207", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.73, "estimated_latency_ms": 140298, "estimated_pass_rate": 0.7062, "estimated_quality": 0.8876, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.424, "max_latency_ms": 188561, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7062, "quality": 0.8876, "quality_target_met": false, "utility": 69.124837}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.424,\"max_latency_ms\":188561,\"minimum_quality\":0.9144},\"task\":{\"ambiguity\":0.745,\"category\":\"coding\",\"complexity\":0.7034,\"consequence\":0.7365,\"evidence_need\":0.7583,\"family\":\"migration_planning\",\"input_tokens\":3440,\"reversibility\":0.4608,\"risk\":0.369,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":0.937,\"id\":\"worker_a\",\"p95_latency_ms\":31985,\"pass_rate\":0.5084,\"provider_group\":\"provider_z\",\"quality\":0.712,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.44,\"id\":\"worker_b\",\"p95_latency_ms\":44650,\"pass_rate\":0.8857,\"provider_group\":\"provider_y\",\"quality\":0.88,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.353,\"id\":\"worker_c\",\"p95_latency_ms\":63663,\"pass_rate\":0.8944,\"provider_group\":\"provider_x\",\"quality\":0.9234,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.73,\"estimated_latency_ms\":140298,\"estimated_pass_rate\":0.7062,\"estimated_quality\":0.8876,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.424,\"max_latency_ms\":188561,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.745, "category": "coding", "complexity": 0.7034, "consequence": 0.7365, "evidence_need": 0.7583, "family": "migration_planning", "input_tokens": 3440, "request_id": "synthetic-migration_planning-001207", "reversibility": 0.4608, "risk": 0.369, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 0.937, "id": "worker_a", "p95_latency_ms": 31985, "pass_rate": 0.5084, "provider_group": "provider_z", "quality": 0.712, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.44, "id": "worker_b", "p95_latency_ms": 44650, "pass_rate": 0.8857, "provider_group": "provider_y", "quality": 0.88, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.353, "id": "worker_c", "p95_latency_ms": 63663, "pass_rate": 0.8944, "provider_group": "provider_x", "quality": 0.9234, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.844, "max_latency_ms": 127335, "minimum_quality": 0.7587}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000741", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.677, "estimated_latency_ms": 106783, "estimated_pass_rate": 0.8436, "estimated_quality": 0.8794, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.844, "max_latency_ms": 127335, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8436, "quality": 0.8794, "quality_target_met": true, "utility": 63.00424}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.844,\"max_latency_ms\":127335,\"minimum_quality\":0.7587},\"task\":{\"ambiguity\":0.5753,\"category\":\"review\",\"complexity\":0.7964,\"consequence\":0.5784,\"evidence_need\":0.6787,\"family\":\"policy_compliance\",\"input_tokens\":2240,\"reversibility\":0.4948,\"risk\":0.6121,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.356,\"id\":\"worker_a\",\"p95_latency_ms\":78499,\"pass_rate\":0.846,\"provider_group\":\"provider_x\",\"quality\":0.8123,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.209,\"id\":\"worker_b\",\"p95_latency_ms\":22870,\"pass_rate\":0.6033,\"provider_group\":\"provider_y\",\"quality\":0.693,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.468,\"id\":\"worker_c\",\"p95_latency_ms\":83913,\"pass_rate\":0.9863,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.677,\"estimated_latency_ms\":106783,\"estimated_pass_rate\":0.8436,\"estimated_quality\":0.8794,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.844,\"max_latency_ms\":127335,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5753, "category": "review", "complexity": 0.7964, "consequence": 0.5784, "evidence_need": 0.6787, "family": "policy_compliance", "input_tokens": 2240, "request_id": "synthetic-policy_compliance-000741", "reversibility": 0.4948, "risk": 0.6121, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.356, "id": "worker_a", "p95_latency_ms": 78499, "pass_rate": 0.846, "provider_group": "provider_x", "quality": 0.8123, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.209, "id": "worker_b", "p95_latency_ms": 22870, "pass_rate": 0.6033, "provider_group": "provider_y", "quality": 0.693, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.468, "id": "worker_c", "p95_latency_ms": 83913, "pass_rate": 0.9863, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.669, "max_latency_ms": 124027, "minimum_quality": 0.7452}, "family": "migration_planning", "id": "synthetic-migration_planning-000631", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.901, "estimated_latency_ms": 109670, "estimated_pass_rate": 0.8898, "estimated_quality": 0.9023, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.669, "max_latency_ms": 124027, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8898, "quality": 0.9023, "quality_target_met": true, "utility": 68.411933}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.669,\"max_latency_ms\":124027,\"minimum_quality\":0.7452},\"task\":{\"ambiguity\":0.6724,\"category\":\"coding\",\"complexity\":0.6799,\"consequence\":0.6093,\"evidence_need\":0.6123,\"family\":\"migration_planning\",\"input_tokens\":3160,\"reversibility\":0.2905,\"risk\":0.682,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":10.621,\"id\":\"worker_a\",\"p95_latency_ms\":50439,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9518,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.28,\"id\":\"worker_b\",\"p95_latency_ms\":59231,\"pass_rate\":0.8685,\"provider_group\":\"provider_x\",\"quality\":0.8195,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.91,\"id\":\"worker_c\",\"p95_latency_ms\":19427,\"pass_rate\":0.4568,\"provider_group\":\"provider_z\",\"quality\":0.7143,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.901,\"estimated_latency_ms\":109670,\"estimated_pass_rate\":0.8898,\"estimated_quality\":0.9023,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.669,\"max_latency_ms\":124027,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6724, "category": "coding", "complexity": 0.6799, "consequence": 0.6093, "evidence_need": 0.6123, "family": "migration_planning", "input_tokens": 3160, "request_id": "synthetic-migration_planning-000631", "reversibility": 0.2905, "risk": 0.682, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 10.621, "id": "worker_a", "p95_latency_ms": 50439, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9518, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.28, "id": "worker_b", "p95_latency_ms": 59231, "pass_rate": 0.8685, "provider_group": "provider_x", "quality": 0.8195, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.91, "id": "worker_c", "p95_latency_ms": 19427, "pass_rate": 0.4568, "provider_group": "provider_z", "quality": 0.7143, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 2.47, "max_latency_ms": 50315, "minimum_quality": 0.7728}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000419", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.121, "estimated_latency_ms": 30853, "estimated_pass_rate": 0.3693, "estimated_quality": 0.5449, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.47, "max_latency_ms": 50315, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3693, "quality": 0.5449, "quality_target_met": false, "utility": 26.769591}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":2.47,\"max_latency_ms\":50315,\"minimum_quality\":0.7728},\"task\":{\"ambiguity\":0.3518,\"category\":\"analysis\",\"complexity\":0.5739,\"consequence\":0.4594,\"evidence_need\":0.3291,\"family\":\"performance_tuning\",\"input_tokens\":7240,\"reversibility\":0.7532,\"risk\":0.4353,\"summary\":\"Choose between throughput and tail-latency changes for provider failover.\"},\"workers\":[{\"credits_per_call\":1.121,\"id\":\"worker_a\",\"p95_latency_ms\":30853,\"pass_rate\":0.4962,\"provider_group\":\"provider_x\",\"quality\":0.6236,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.722,\"id\":\"worker_b\",\"p95_latency_ms\":67106,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.85,\"id\":\"worker_c\",\"p95_latency_ms\":47762,\"pass_rate\":0.931,\"provider_group\":\"provider_z\",\"quality\":0.9098,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.121,\"estimated_latency_ms\":30853,\"estimated_pass_rate\":0.3693,\"estimated_quality\":0.5449,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.47,\"max_latency_ms\":50315,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3518, "category": "analysis", "complexity": 0.5739, "consequence": 0.4594, "evidence_need": 0.3291, "family": "performance_tuning", "input_tokens": 7240, "request_id": "synthetic-performance_tuning-000419", "reversibility": 0.7532, "risk": 0.4353, "summary": "Choose between throughput and tail-latency changes for provider failover."}, "workers": [{"credits_per_call": 1.121, "id": "worker_a", "p95_latency_ms": 30853, "pass_rate": 0.4962, "provider_group": "provider_x", "quality": 0.6236, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.722, "id": "worker_b", "p95_latency_ms": 67106, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.85, "id": "worker_c", "p95_latency_ms": 47762, "pass_rate": 0.931, "provider_group": "provider_z", "quality": 0.9098, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.559, "max_latency_ms": 140521, "minimum_quality": 0.8103}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001094", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.161, "estimated_latency_ms": 112843, "estimated_pass_rate": 0.9122, "estimated_quality": 0.9328, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.559, "max_latency_ms": 140521, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9122, "quality": 0.9328, "quality_target_met": true, "utility": 70.216122}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.559,\"max_latency_ms\":140521,\"minimum_quality\":0.8103},\"task\":{\"ambiguity\":0.3465,\"category\":\"analysis\",\"complexity\":0.8371,\"consequence\":0.6498,\"evidence_need\":0.3364,\"family\":\"performance_tuning\",\"input_tokens\":2520,\"reversibility\":0.8116,\"risk\":0.3395,\"summary\":\"Diagnose latency variance in streaming inference and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.818,\"id\":\"worker_a\",\"p95_latency_ms\":53771,\"pass_rate\":0.9511,\"provider_group\":\"provider_x\",\"quality\":0.9485,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.343,\"id\":\"worker_b\",\"p95_latency_ms\":59072,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.909,\"id\":\"worker_c\",\"p95_latency_ms\":27293,\"pass_rate\":0.5088,\"provider_group\":\"provider_z\",\"quality\":0.6274,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.161,\"estimated_latency_ms\":112843,\"estimated_pass_rate\":0.9122,\"estimated_quality\":0.9328,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.559,\"max_latency_ms\":140521,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3465, "category": "analysis", "complexity": 0.8371, "consequence": 0.6498, "evidence_need": 0.3364, "family": "performance_tuning", "input_tokens": 2520, "request_id": "synthetic-performance_tuning-001094", "reversibility": 0.8116, "risk": 0.3395, "summary": "Diagnose latency variance in streaming inference and rank the next experiments."}, "workers": [{"credits_per_call": 3.818, "id": "worker_a", "p95_latency_ms": 53771, "pass_rate": 0.9511, "provider_group": "provider_x", "quality": 0.9485, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.343, "id": "worker_b", "p95_latency_ms": 59072, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.909, "id": "worker_c", "p95_latency_ms": 27293, "pass_rate": 0.5088, "provider_group": "provider_z", "quality": 0.6274, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.211, "max_latency_ms": 53879, "minimum_quality": 0.4987}, "family": "migration_planning", "id": "synthetic-migration_planning-000538", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.145, "estimated_latency_ms": 32407, "estimated_pass_rate": 0.3683, "estimated_quality": 0.6093, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.211, "max_latency_ms": 53879, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3683, "quality": 0.6093, "quality_target_met": true, "utility": 33.544745}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.211,\"max_latency_ms\":53879,\"minimum_quality\":0.4987},\"task\":{\"ambiguity\":0.3298,\"category\":\"coding\",\"complexity\":0.5823,\"consequence\":0.4423,\"evidence_need\":0.7967,\"family\":\"migration_planning\",\"input_tokens\":5720,\"reversibility\":0.2945,\"risk\":0.578,\"summary\":\"Review a staged migration for provider failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.913,\"id\":\"worker_a\",\"p95_latency_ms\":62564,\"pass_rate\":0.737,\"provider_group\":\"provider_z\",\"quality\":0.9081,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.058,\"id\":\"worker_b\",\"p95_latency_ms\":68689,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9426,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.145,\"id\":\"worker_c\",\"p95_latency_ms\":32407,\"pass_rate\":0.5061,\"provider_group\":\"provider_x\",\"quality\":0.6984,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.145,\"estimated_latency_ms\":32407,\"estimated_pass_rate\":0.3683,\"estimated_quality\":0.6093,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.211,\"max_latency_ms\":53879,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3298, "category": "coding", "complexity": 0.5823, "consequence": 0.4423, "evidence_need": 0.7967, "family": "migration_planning", "input_tokens": 5720, "request_id": "synthetic-migration_planning-000538", "reversibility": 0.2945, "risk": 0.578, "summary": "Review a staged migration for provider failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.913, "id": "worker_a", "p95_latency_ms": 62564, "pass_rate": 0.737, "provider_group": "provider_z", "quality": 0.9081, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.058, "id": "worker_b", "p95_latency_ms": 68689, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9426, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.145, "id": "worker_c", "p95_latency_ms": 32407, "pass_rate": 0.5061, "provider_group": "provider_x", "quality": 0.6984, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.943, "max_latency_ms": 203506, "minimum_quality": 0.8072}, "family": "migration_planning", "id": "synthetic-migration_planning-001045", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.328, "estimated_latency_ms": 180390, "estimated_pass_rate": 0.7334, "estimated_quality": 0.9053, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.943, "max_latency_ms": 203506, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7334, "quality": 0.9053, "quality_target_met": true, "utility": 72.940993}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.943,\"max_latency_ms\":203506,\"minimum_quality\":0.8072},\"task\":{\"ambiguity\":0.3756,\"category\":\"coding\",\"complexity\":0.867,\"consequence\":0.7316,\"evidence_need\":0.5464,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.4717,\"risk\":0.5701,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.666,\"id\":\"worker_a\",\"p95_latency_ms\":79505,\"pass_rate\":0.7186,\"provider_group\":\"provider_x\",\"quality\":0.8111,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.65,\"id\":\"worker_b\",\"p95_latency_ms\":81525,\"pass_rate\":0.9936,\"provider_group\":\"provider_y\",\"quality\":0.9778,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.012,\"id\":\"worker_c\",\"p95_latency_ms\":19360,\"pass_rate\":0.5453,\"provider_group\":\"provider_z\",\"quality\":0.7902,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.328,\"estimated_latency_ms\":180390,\"estimated_pass_rate\":0.7334,\"estimated_quality\":0.9053,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.943,\"max_latency_ms\":203506,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3756, "category": "coding", "complexity": 0.867, "consequence": 0.7316, "evidence_need": 0.5464, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-001045", "reversibility": 0.4717, "risk": 0.5701, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 3.666, "id": "worker_a", "p95_latency_ms": 79505, "pass_rate": 0.7186, "provider_group": "provider_x", "quality": 0.8111, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 14.65, "id": "worker_b", "p95_latency_ms": 81525, "pass_rate": 0.9936, "provider_group": "provider_y", "quality": 0.9778, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.012, "id": "worker_c", "p95_latency_ms": 19360, "pass_rate": 0.5453, "provider_group": "provider_z", "quality": 0.7902, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 4.775, "max_latency_ms": 157376, "minimum_quality": 0.6819}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000084", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.468, "estimated_latency_ms": 74492, "estimated_pass_rate": 0.6366, "estimated_quality": 0.8308, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.775, "max_latency_ms": 157376, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6366, "quality": 0.8308, "quality_target_met": true, "utility": 62.843088}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":4.775,\"max_latency_ms\":157376,\"minimum_quality\":0.6819},\"task\":{\"ambiguity\":0.4947,\"category\":\"review\",\"complexity\":0.7428,\"consequence\":0.8633,\"evidence_need\":0.9156,\"family\":\"policy_compliance\",\"input_tokens\":760,\"reversibility\":0.6124,\"risk\":0.6895,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":3.371,\"id\":\"worker_a\",\"p95_latency_ms\":55511,\"pass_rate\":0.8031,\"provider_group\":\"provider_z\",\"quality\":0.8596,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.097,\"id\":\"worker_b\",\"p95_latency_ms\":18981,\"pass_rate\":0.5623,\"provider_group\":\"provider_x\",\"quality\":0.7631,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.266,\"id\":\"worker_c\",\"p95_latency_ms\":88839,\"pass_rate\":0.9593,\"provider_group\":\"provider_y\",\"quality\":0.9946,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.468,\"estimated_latency_ms\":74492,\"estimated_pass_rate\":0.6366,\"estimated_quality\":0.8308,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.775,\"max_latency_ms\":157376,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4947, "category": "review", "complexity": 0.7428, "consequence": 0.8633, "evidence_need": 0.9156, "family": "policy_compliance", "input_tokens": 760, "request_id": "synthetic-policy_compliance-000084", "reversibility": 0.6124, "risk": 0.6895, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 3.371, "id": "worker_a", "p95_latency_ms": 55511, "pass_rate": 0.8031, "provider_group": "provider_z", "quality": 0.8596, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.097, "id": "worker_b", "p95_latency_ms": 18981, "pass_rate": 0.5623, "provider_group": "provider_x", "quality": 0.7631, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.266, "id": "worker_c", "p95_latency_ms": 88839, "pass_rate": 0.9593, "provider_group": "provider_y", "quality": 0.9946, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.59, "max_latency_ms": 34741, "minimum_quality": 0.7646}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000519", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.215, "estimated_latency_ms": 27981, "estimated_pass_rate": 0.4056, "estimated_quality": 0.5896, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.59, "max_latency_ms": 34741, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4056, "quality": 0.5896, "quality_target_met": false, "utility": 19.532037}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.59,\"max_latency_ms\":34741,\"minimum_quality\":0.7646},\"task\":{\"ambiguity\":0.5746,\"category\":\"review\",\"complexity\":0.8305,\"consequence\":0.6622,\"evidence_need\":0.637,\"family\":\"policy_compliance\",\"input_tokens\":160,\"reversibility\":0.3475,\"risk\":0.7796,\"summary\":\"Review whether tenant-scoped caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.997,\"id\":\"worker_a\",\"p95_latency_ms\":64039,\"pass_rate\":0.7184,\"provider_group\":\"provider_z\",\"quality\":0.7995,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.215,\"id\":\"worker_b\",\"p95_latency_ms\":27981,\"pass_rate\":0.6064,\"provider_group\":\"provider_y\",\"quality\":0.7545,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.066,\"id\":\"worker_c\",\"p95_latency_ms\":83839,\"pass_rate\":0.9581,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.215,\"estimated_latency_ms\":27981,\"estimated_pass_rate\":0.4056,\"estimated_quality\":0.5896,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.59,\"max_latency_ms\":34741,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5746, "category": "review", "complexity": 0.8305, "consequence": 0.6622, "evidence_need": 0.637, "family": "policy_compliance", "input_tokens": 160, "request_id": "synthetic-policy_compliance-000519", "reversibility": 0.3475, "risk": 0.7796, "summary": "Review whether tenant-scoped caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.997, "id": "worker_a", "p95_latency_ms": 64039, "pass_rate": 0.7184, "provider_group": "provider_z", "quality": 0.7995, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.215, "id": "worker_b", "p95_latency_ms": 27981, "pass_rate": 0.6064, "provider_group": "provider_y", "quality": 0.7545, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.066, "id": "worker_c", "p95_latency_ms": 83839, "pass_rate": 0.9581, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.693, "max_latency_ms": 188899, "minimum_quality": 0.7631}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001026", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.867, "estimated_latency_ms": 82278, "estimated_pass_rate": 0.6233, "estimated_quality": 0.8341, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.693, "max_latency_ms": 188899, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6233, "quality": 0.8341, "quality_target_met": true, "utility": 63.023721}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.693,\"max_latency_ms\":188899,\"minimum_quality\":0.7631},\"task\":{\"ambiguity\":0.3826,\"category\":\"review\",\"complexity\":0.7176,\"consequence\":0.6605,\"evidence_need\":0.8452,\"family\":\"policy_compliance\",\"input_tokens\":7960,\"reversibility\":0.6074,\"risk\":0.5903,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.312,\"id\":\"worker_a\",\"p95_latency_ms\":73406,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9366,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.959,\"id\":\"worker_b\",\"p95_latency_ms\":54127,\"pass_rate\":0.8736,\"provider_group\":\"provider_y\",\"quality\":0.8814,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.908,\"id\":\"worker_c\",\"p95_latency_ms\":28151,\"pass_rate\":0.4957,\"provider_group\":\"provider_z\",\"quality\":0.7246,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.867,\"estimated_latency_ms\":82278,\"estimated_pass_rate\":0.6233,\"estimated_quality\":0.8341,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.693,\"max_latency_ms\":188899,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3826, "category": "review", "complexity": 0.7176, "consequence": 0.6605, "evidence_need": 0.8452, "family": "policy_compliance", "input_tokens": 7960, "request_id": "synthetic-policy_compliance-001026", "reversibility": 0.6074, "risk": 0.5903, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.312, "id": "worker_a", "p95_latency_ms": 73406, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9366, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.959, "id": "worker_b", "p95_latency_ms": 54127, "pass_rate": 0.8736, "provider_group": "provider_y", "quality": 0.8814, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.908, "id": "worker_c", "p95_latency_ms": 28151, "pass_rate": 0.4957, "provider_group": "provider_z", "quality": 0.7246, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.852, "max_latency_ms": 28682, "minimum_quality": 0.6282}, "family": "migration_planning", "id": "synthetic-migration_planning-000052", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.824, "estimated_latency_ms": 20680, "estimated_pass_rate": 0.339, "estimated_quality": 0.66, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.852, "max_latency_ms": 28682, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.339, "quality": 0.66, "quality_target_met": true, "utility": 43.975052}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.852,\"max_latency_ms\":28682,\"minimum_quality\":0.6282},\"task\":{\"ambiguity\":0.6351,\"category\":\"coding\",\"complexity\":0.644,\"consequence\":0.6794,\"evidence_need\":0.4693,\"family\":\"migration_planning\",\"input_tokens\":3200,\"reversibility\":0.4982,\"risk\":0.4428,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":10.567,\"id\":\"worker_a\",\"p95_latency_ms\":61133,\"pass_rate\":0.9427,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.919,\"id\":\"worker_b\",\"p95_latency_ms\":57947,\"pass_rate\":0.8577,\"provider_group\":\"provider_y\",\"quality\":0.845,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.824,\"id\":\"worker_c\",\"p95_latency_ms\":20680,\"pass_rate\":0.4959,\"provider_group\":\"provider_x\",\"quality\":0.771,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.824,\"estimated_latency_ms\":20680,\"estimated_pass_rate\":0.339,\"estimated_quality\":0.66,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.852,\"max_latency_ms\":28682,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6351, "category": "coding", "complexity": 0.644, "consequence": 0.6794, "evidence_need": 0.4693, "family": "migration_planning", "input_tokens": 3200, "request_id": "synthetic-migration_planning-000052", "reversibility": 0.4982, "risk": 0.4428, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 10.567, "id": "worker_a", "p95_latency_ms": 61133, "pass_rate": 0.9427, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.919, "id": "worker_b", "p95_latency_ms": 57947, "pass_rate": 0.8577, "provider_group": "provider_y", "quality": 0.845, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.824, "id": "worker_c", "p95_latency_ms": 20680, "pass_rate": 0.4959, "provider_group": "provider_x", "quality": 0.771, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.658, "max_latency_ms": 158509, "minimum_quality": 0.9166}, "family": "migration_planning", "id": "synthetic-migration_planning-000997", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.714, "estimated_latency_ms": 151718, "estimated_pass_rate": 0.7288, "estimated_quality": 0.8874, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.658, "max_latency_ms": 158509, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7288, "quality": 0.8874, "quality_target_met": false, "utility": 68.90207}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.658,\"max_latency_ms\":158509,\"minimum_quality\":0.9166},\"task\":{\"ambiguity\":0.51,\"category\":\"coding\",\"complexity\":0.7728,\"consequence\":0.759,\"evidence_need\":0.7919,\"family\":\"migration_planning\",\"input_tokens\":7400,\"reversibility\":0.5827,\"risk\":0.433,\"summary\":\"Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.375,\"id\":\"worker_a\",\"p95_latency_ms\":73897,\"pass_rate\":0.9753,\"provider_group\":\"provider_x\",\"quality\":0.9201,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.091,\"id\":\"worker_b\",\"p95_latency_ms\":27691,\"pass_rate\":0.5048,\"provider_group\":\"provider_y\",\"quality\":0.7327,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.248,\"id\":\"worker_c\",\"p95_latency_ms\":50130,\"pass_rate\":0.8138,\"provider_group\":\"provider_z\",\"quality\":0.8663,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.714,\"estimated_latency_ms\":151718,\"estimated_pass_rate\":0.7288,\"estimated_quality\":0.8874,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.658,\"max_latency_ms\":158509,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.51, "category": "coding", "complexity": 0.7728, "consequence": 0.759, "evidence_need": 0.7919, "family": "migration_planning", "input_tokens": 7400, "request_id": "synthetic-migration_planning-000997", "reversibility": 0.5827, "risk": 0.433, "summary": "Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints."}, "workers": [{"credits_per_call": 14.375, "id": "worker_a", "p95_latency_ms": 73897, "pass_rate": 0.9753, "provider_group": "provider_x", "quality": 0.9201, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.091, "id": "worker_b", "p95_latency_ms": 27691, "pass_rate": 0.5048, "provider_group": "provider_y", "quality": 0.7327, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.248, "id": "worker_c", "p95_latency_ms": 50130, "pass_rate": 0.8138, "provider_group": "provider_z", "quality": 0.8663, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.933, "max_latency_ms": 194378, "minimum_quality": 0.8251}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000987", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.77, "estimated_latency_ms": 85745, "estimated_pass_rate": 0.6045, "estimated_quality": 0.8065, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.933, "max_latency_ms": 194378, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6045, "quality": 0.8065, "quality_target_met": false, "utility": 62.109758}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.933,\"max_latency_ms\":194378,\"minimum_quality\":0.8251},\"task\":{\"ambiguity\":0.5973,\"category\":\"review\",\"complexity\":0.6163,\"consequence\":0.7153,\"evidence_need\":0.907,\"family\":\"policy_compliance\",\"input_tokens\":7320,\"reversibility\":0.4693,\"risk\":0.8411,\"summary\":\"Review whether provider failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":11.976,\"id\":\"worker_a\",\"p95_latency_ms\":88099,\"pass_rate\":0.9945,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.759,\"id\":\"worker_b\",\"p95_latency_ms\":55803,\"pass_rate\":0.748,\"provider_group\":\"provider_x\",\"quality\":0.8454,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.011,\"id\":\"worker_c\",\"p95_latency_ms\":29942,\"pass_rate\":0.5159,\"provider_group\":\"provider_z\",\"quality\":0.7319,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.77,\"estimated_latency_ms\":85745,\"estimated_pass_rate\":0.6045,\"estimated_quality\":0.8065,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.933,\"max_latency_ms\":194378,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5973, "category": "review", "complexity": 0.6163, "consequence": 0.7153, "evidence_need": 0.907, "family": "policy_compliance", "input_tokens": 7320, "request_id": "synthetic-policy_compliance-000987", "reversibility": 0.4693, "risk": 0.8411, "summary": "Review whether provider failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 11.976, "id": "worker_a", "p95_latency_ms": 88099, "pass_rate": 0.9945, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.759, "id": "worker_b", "p95_latency_ms": 55803, "pass_rate": 0.748, "provider_group": "provider_x", "quality": 0.8454, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.011, "id": "worker_c", "p95_latency_ms": 29942, "pass_rate": 0.5159, "provider_group": "provider_z", "quality": 0.7319, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.112, "max_latency_ms": 191707, "minimum_quality": 0.9586}, "family": "migration_planning", "id": "synthetic-migration_planning-000856", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.157, "estimated_latency_ms": 149324, "estimated_pass_rate": 0.7294, "estimated_quality": 0.9097, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.112, "max_latency_ms": 191707, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7294, "quality": 0.9097, "quality_target_met": false, "utility": 72.750872}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.112,\"max_latency_ms\":191707,\"minimum_quality\":0.9586},\"task\":{\"ambiguity\":0.735,\"category\":\"coding\",\"complexity\":0.5687,\"consequence\":0.7066,\"evidence_need\":0.7301,\"family\":\"migration_planning\",\"input_tokens\":1920,\"reversibility\":0.4914,\"risk\":0.6848,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":10.756,\"id\":\"worker_a\",\"p95_latency_ms\":78774,\"pass_rate\":0.9204,\"provider_group\":\"provider_z\",\"quality\":0.9703,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.942,\"id\":\"worker_b\",\"p95_latency_ms\":21591,\"pass_rate\":0.4772,\"provider_group\":\"provider_x\",\"quality\":0.7148,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.459,\"id\":\"worker_c\",\"p95_latency_ms\":48959,\"pass_rate\":0.8299,\"provider_group\":\"provider_y\",\"quality\":0.905,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.157,\"estimated_latency_ms\":149324,\"estimated_pass_rate\":0.7294,\"estimated_quality\":0.9097,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.112,\"max_latency_ms\":191707,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.735, "category": "coding", "complexity": 0.5687, "consequence": 0.7066, "evidence_need": 0.7301, "family": "migration_planning", "input_tokens": 1920, "request_id": "synthetic-migration_planning-000856", "reversibility": 0.4914, "risk": 0.6848, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 10.756, "id": "worker_a", "p95_latency_ms": 78774, "pass_rate": 0.9204, "provider_group": "provider_z", "quality": 0.9703, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.942, "id": "worker_b", "p95_latency_ms": 21591, "pass_rate": 0.4772, "provider_group": "provider_x", "quality": 0.7148, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.459, "id": "worker_c", "p95_latency_ms": 48959, "pass_rate": 0.8299, "provider_group": "provider_y", "quality": 0.905, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.815, "max_latency_ms": 57863, "minimum_quality": 0.6616}, "family": "migration_planning", "id": "synthetic-migration_planning-000736", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.141, "estimated_latency_ms": 29609, "estimated_pass_rate": 0.4095, "estimated_quality": 0.5655, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.815, "max_latency_ms": 57863, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4095, "quality": 0.5655, "quality_target_met": false, "utility": 36.047033}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.815,\"max_latency_ms\":57863,\"minimum_quality\":0.6616},\"task\":{\"ambiguity\":0.4602,\"category\":\"coding\",\"complexity\":0.7561,\"consequence\":0.6783,\"evidence_need\":0.8346,\"family\":\"migration_planning\",\"input_tokens\":7480,\"reversibility\":0.3973,\"risk\":0.4612,\"summary\":\"Design compatibility testing for a usage metering schema migration.\"},\"workers\":[{\"credits_per_call\":10.622,\"id\":\"worker_a\",\"p95_latency_ms\":75261,\"pass_rate\":0.969,\"provider_group\":\"provider_x\",\"quality\":0.9468,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.575,\"id\":\"worker_b\",\"p95_latency_ms\":68817,\"pass_rate\":0.7646,\"provider_group\":\"provider_y\",\"quality\":0.9114,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.141,\"id\":\"worker_c\",\"p95_latency_ms\":29609,\"pass_rate\":0.5674,\"provider_group\":\"provider_z\",\"quality\":0.6863,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.141,\"estimated_latency_ms\":29609,\"estimated_pass_rate\":0.4095,\"estimated_quality\":0.5655,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.815,\"max_latency_ms\":57863,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4602, "category": "coding", "complexity": 0.7561, "consequence": 0.6783, "evidence_need": 0.8346, "family": "migration_planning", "input_tokens": 7480, "request_id": "synthetic-migration_planning-000736", "reversibility": 0.3973, "risk": 0.4612, "summary": "Design compatibility testing for a usage metering schema migration."}, "workers": [{"credits_per_call": 10.622, "id": "worker_a", "p95_latency_ms": 75261, "pass_rate": 0.969, "provider_group": "provider_x", "quality": 0.9468, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.575, "id": "worker_b", "p95_latency_ms": 68817, "pass_rate": 0.7646, "provider_group": "provider_y", "quality": 0.9114, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.141, "id": "worker_c", "p95_latency_ms": 29609, "pass_rate": 0.5674, "provider_group": "provider_z", "quality": 0.6863, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.483, "max_latency_ms": 180369, "minimum_quality": 0.9042}, "family": "migration_planning", "id": "synthetic-migration_planning-001222", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.82, "estimated_latency_ms": 145285, "estimated_pass_rate": 0.6987, "estimated_quality": 0.8949, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.483, "max_latency_ms": 180369, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6987, "quality": 0.8949, "quality_target_met": false, "utility": 70.655123}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.483,\"max_latency_ms\":180369,\"minimum_quality\":0.9042},\"task\":{\"ambiguity\":0.673,\"category\":\"coding\",\"complexity\":0.7536,\"consequence\":0.8083,\"evidence_need\":0.5407,\"family\":\"migration_planning\",\"input_tokens\":640,\"reversibility\":0.5184,\"risk\":0.4265,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":11.552,\"id\":\"worker_a\",\"p95_latency_ms\":70404,\"pass_rate\":0.9411,\"provider_group\":\"provider_y\",\"quality\":0.9384,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.057,\"id\":\"worker_b\",\"p95_latency_ms\":47643,\"pass_rate\":0.8838,\"provider_group\":\"provider_z\",\"quality\":0.8097,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.211,\"id\":\"worker_c\",\"p95_latency_ms\":27238,\"pass_rate\":0.467,\"provider_group\":\"provider_x\",\"quality\":0.7832,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.82,\"estimated_latency_ms\":145285,\"estimated_pass_rate\":0.6987,\"estimated_quality\":0.8949,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.483,\"max_latency_ms\":180369,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.673, "category": "coding", "complexity": 0.7536, "consequence": 0.8083, "evidence_need": 0.5407, "family": "migration_planning", "input_tokens": 640, "request_id": "synthetic-migration_planning-001222", "reversibility": 0.5184, "risk": 0.4265, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 11.552, "id": "worker_a", "p95_latency_ms": 70404, "pass_rate": 0.9411, "provider_group": "provider_y", "quality": 0.9384, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.057, "id": "worker_b", "p95_latency_ms": 47643, "pass_rate": 0.8838, "provider_group": "provider_z", "quality": 0.8097, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.211, "id": "worker_c", "p95_latency_ms": 27238, "pass_rate": 0.467, "provider_group": "provider_x", "quality": 0.7832, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.474, "max_latency_ms": 121503, "minimum_quality": 0.8067}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000836", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.351, "estimated_latency_ms": 103070, "estimated_pass_rate": 0.8953, "estimated_quality": 0.9466, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.474, "max_latency_ms": 121503, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8953, "quality": 0.9466, "quality_target_met": true, "utility": 68.704442}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.474,\"max_latency_ms\":121503,\"minimum_quality\":0.8067},\"task\":{\"ambiguity\":0.4395,\"category\":\"analysis\",\"complexity\":0.767,\"consequence\":0.2895,\"evidence_need\":0.7175,\"family\":\"performance_tuning\",\"input_tokens\":1520,\"reversibility\":0.686,\"risk\":0.4337,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":13.1,\"id\":\"worker_a\",\"p95_latency_ms\":54482,\"pass_rate\":0.9712,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.251,\"id\":\"worker_b\",\"p95_latency_ms\":48588,\"pass_rate\":0.8948,\"provider_group\":\"provider_z\",\"quality\":0.9528,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.937,\"id\":\"worker_c\",\"p95_latency_ms\":31074,\"pass_rate\":0.5503,\"provider_group\":\"provider_y\",\"quality\":0.6723,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.351,\"estimated_latency_ms\":103070,\"estimated_pass_rate\":0.8953,\"estimated_quality\":0.9466,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.474,\"max_latency_ms\":121503,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4395, "category": "analysis", "complexity": 0.767, "consequence": 0.2895, "evidence_need": 0.7175, "family": "performance_tuning", "input_tokens": 1520, "request_id": "synthetic-performance_tuning-000836", "reversibility": 0.686, "risk": 0.4337, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 13.1, "id": "worker_a", "p95_latency_ms": 54482, "pass_rate": 0.9712, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.251, "id": "worker_b", "p95_latency_ms": 48588, "pass_rate": 0.8948, "provider_group": "provider_z", "quality": 0.9528, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.937, "id": "worker_c", "p95_latency_ms": 31074, "pass_rate": 0.5503, "provider_group": "provider_y", "quality": 0.6723, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.923, "max_latency_ms": 40880, "minimum_quality": 0.6296}, "family": "migration_planning", "id": "synthetic-migration_planning-000172", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.198, "estimated_latency_ms": 21103, "estimated_pass_rate": 0.3506, "estimated_quality": 0.6384, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.923, "max_latency_ms": 40880, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3506, "quality": 0.6384, "quality_target_met": true, "utility": 42.452648}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.923,\"max_latency_ms\":40880,\"minimum_quality\":0.6296},\"task\":{\"ambiguity\":0.4851,\"category\":\"coding\",\"complexity\":0.5551,\"consequence\":0.5304,\"evidence_need\":0.5408,\"family\":\"migration_planning\",\"input_tokens\":320,\"reversibility\":0.6461,\"risk\":0.673,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.72,\"id\":\"worker_a\",\"p95_latency_ms\":76360,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9324,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.198,\"id\":\"worker_b\",\"p95_latency_ms\":21103,\"pass_rate\":0.5047,\"provider_group\":\"provider_z\",\"quality\":0.7398,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.23,\"id\":\"worker_c\",\"p95_latency_ms\":64809,\"pass_rate\":0.824,\"provider_group\":\"provider_y\",\"quality\":0.9133,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.198,\"estimated_latency_ms\":21103,\"estimated_pass_rate\":0.3506,\"estimated_quality\":0.6384,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.923,\"max_latency_ms\":40880,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4851, "category": "coding", "complexity": 0.5551, "consequence": 0.5304, "evidence_need": 0.5408, "family": "migration_planning", "input_tokens": 320, "request_id": "synthetic-migration_planning-000172", "reversibility": 0.6461, "risk": 0.673, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 11.72, "id": "worker_a", "p95_latency_ms": 76360, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9324, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.198, "id": "worker_b", "p95_latency_ms": 21103, "pass_rate": 0.5047, "provider_group": "provider_z", "quality": 0.7398, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.23, "id": "worker_c", "p95_latency_ms": 64809, "pass_rate": 0.824, "provider_group": "provider_y", "quality": 0.9133, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.199, "max_latency_ms": 24864, "minimum_quality": 0.7575}, "family": "migration_planning", "id": "synthetic-migration_planning-000283", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.099, "estimated_latency_ms": 18978, "estimated_pass_rate": 0.3537, "estimated_quality": 0.5871, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.199, "max_latency_ms": 24864, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3537, "quality": 0.5871, "quality_target_met": false, "utility": 27.527595}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.199,\"max_latency_ms\":24864,\"minimum_quality\":0.7575},\"task\":{\"ambiguity\":0.7397,\"category\":\"coding\",\"complexity\":0.8549,\"consequence\":0.6147,\"evidence_need\":0.6001,\"family\":\"migration_planning\",\"input_tokens\":6880,\"reversibility\":0.4724,\"risk\":0.726,\"summary\":\"Review a staged migration for artifact storage and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.099,\"id\":\"worker_a\",\"p95_latency_ms\":18978,\"pass_rate\":0.5654,\"provider_group\":\"provider_x\",\"quality\":0.7636,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.077,\"id\":\"worker_b\",\"p95_latency_ms\":78027,\"pass_rate\":0.7638,\"provider_group\":\"provider_y\",\"quality\":0.9156,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.951,\"id\":\"worker_c\",\"p95_latency_ms\":73325,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.099,\"estimated_latency_ms\":18978,\"estimated_pass_rate\":0.3537,\"estimated_quality\":0.5871,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.199,\"max_latency_ms\":24864,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7397, "category": "coding", "complexity": 0.8549, "consequence": 0.6147, "evidence_need": 0.6001, "family": "migration_planning", "input_tokens": 6880, "request_id": "synthetic-migration_planning-000283", "reversibility": 0.4724, "risk": 0.726, "summary": "Review a staged migration for artifact storage and identify irreversible steps."}, "workers": [{"credits_per_call": 1.099, "id": "worker_a", "p95_latency_ms": 18978, "pass_rate": 0.5654, "provider_group": "provider_x", "quality": 0.7636, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.077, "id": "worker_b", "p95_latency_ms": 78027, "pass_rate": 0.7638, "provider_group": "provider_y", "quality": 0.9156, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.951, "id": "worker_c", "p95_latency_ms": 73325, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 1.573, "max_latency_ms": 28232, "minimum_quality": 0.4922}, "family": "migration_planning", "id": "synthetic-migration_planning-000397", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.197, "estimated_latency_ms": 19627, "estimated_pass_rate": 0.4645, "estimated_quality": 0.6127, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.573, "max_latency_ms": 28232, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4645, "quality": 0.6127, "quality_target_met": true, "utility": 40.215718}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":1.573,\"max_latency_ms\":28232,\"minimum_quality\":0.4922},\"task\":{\"ambiguity\":0.4325,\"category\":\"coding\",\"complexity\":0.7442,\"consequence\":0.678,\"evidence_need\":0.5557,\"family\":\"migration_planning\",\"input_tokens\":8000,\"reversibility\":0.5853,\"risk\":0.4299,\"summary\":\"Design compatibility testing for a prompt caching schema migration.\"},\"workers\":[{\"credits_per_call\":3.534,\"id\":\"worker_a\",\"p95_latency_ms\":56921,\"pass_rate\":0.8047,\"provider_group\":\"provider_x\",\"quality\":0.8422,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.789,\"id\":\"worker_b\",\"p95_latency_ms\":84222,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.197,\"id\":\"worker_c\",\"p95_latency_ms\":19627,\"pass_rate\":0.6164,\"provider_group\":\"provider_y\",\"quality\":0.7273,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.197,\"estimated_latency_ms\":19627,\"estimated_pass_rate\":0.4645,\"estimated_quality\":0.6127,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.573,\"max_latency_ms\":28232,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4325, "category": "coding", "complexity": 0.7442, "consequence": 0.678, "evidence_need": 0.5557, "family": "migration_planning", "input_tokens": 8000, "request_id": "synthetic-migration_planning-000397", "reversibility": 0.5853, "risk": 0.4299, "summary": "Design compatibility testing for a prompt caching schema migration."}, "workers": [{"credits_per_call": 3.534, "id": "worker_a", "p95_latency_ms": 56921, "pass_rate": 0.8047, "provider_group": "provider_x", "quality": 0.8422, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 13.789, "id": "worker_b", "p95_latency_ms": 84222, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.197, "id": "worker_c", "p95_latency_ms": 19627, "pass_rate": 0.6164, "provider_group": "provider_y", "quality": 0.7273, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.14, "max_latency_ms": 143535, "minimum_quality": 0.8982}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000381", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.92, "estimated_latency_ms": 127543, "estimated_pass_rate": 0.8156, "estimated_quality": 0.9221, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.14, "max_latency_ms": 143535, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8156, "quality": 0.9221, "quality_target_met": true, "utility": 69.662252}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.14,\"max_latency_ms\":143535,\"minimum_quality\":0.8982},\"task\":{\"ambiguity\":0.6235,\"category\":\"review\",\"complexity\":0.7064,\"consequence\":0.6178,\"evidence_need\":0.6886,\"family\":\"policy_compliance\",\"input_tokens\":5120,\"reversibility\":0.6805,\"risk\":0.7713,\"summary\":\"Audit a workflow for data retention against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.035,\"id\":\"worker_a\",\"p95_latency_ms\":59892,\"pass_rate\":0.7682,\"provider_group\":\"provider_z\",\"quality\":0.9157,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.885,\"id\":\"worker_b\",\"p95_latency_ms\":67651,\"pass_rate\":0.9782,\"provider_group\":\"provider_y\",\"quality\":0.9244,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.993,\"id\":\"worker_c\",\"p95_latency_ms\":31259,\"pass_rate\":0.6095,\"provider_group\":\"provider_x\",\"quality\":0.702,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.92,\"estimated_latency_ms\":127543,\"estimated_pass_rate\":0.8156,\"estimated_quality\":0.9221,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.14,\"max_latency_ms\":143535,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6235, "category": "review", "complexity": 0.7064, "consequence": 0.6178, "evidence_need": 0.6886, "family": "policy_compliance", "input_tokens": 5120, "request_id": "synthetic-policy_compliance-000381", "reversibility": 0.6805, "risk": 0.7713, "summary": "Audit a workflow for data retention against explicit access requirements."}, "workers": [{"credits_per_call": 3.035, "id": "worker_a", "p95_latency_ms": 59892, "pass_rate": 0.7682, "provider_group": "provider_z", "quality": 0.9157, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.885, "id": "worker_b", "p95_latency_ms": 67651, "pass_rate": 0.9782, "provider_group": "provider_y", "quality": 0.9244, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.993, "id": "worker_c", "p95_latency_ms": 31259, "pass_rate": 0.6095, "provider_group": "provider_x", "quality": 0.702, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.387, "max_latency_ms": 201408, "minimum_quality": 0.8727}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001236", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.17, "estimated_latency_ms": 118444, "estimated_pass_rate": 0.8723, "estimated_quality": 0.9278, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.387, "max_latency_ms": 201408, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8723, "quality": 0.9278, "quality_target_met": true, "utility": 72.337897}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.387,\"max_latency_ms\":201408,\"minimum_quality\":0.8727},\"task\":{\"ambiguity\":0.4514,\"category\":\"review\",\"complexity\":0.6023,\"consequence\":0.7771,\"evidence_need\":0.7318,\"family\":\"policy_compliance\",\"input_tokens\":5280,\"reversibility\":0.4051,\"risk\":0.5699,\"summary\":\"Review whether database failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.9,\"id\":\"worker_a\",\"p95_latency_ms\":31723,\"pass_rate\":0.4898,\"provider_group\":\"provider_y\",\"quality\":0.7919,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.568,\"id\":\"worker_b\",\"p95_latency_ms\":58032,\"pass_rate\":0.9588,\"provider_group\":\"provider_x\",\"quality\":0.9814,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.602,\"id\":\"worker_c\",\"p95_latency_ms\":60412,\"pass_rate\":0.756,\"provider_group\":\"provider_z\",\"quality\":0.8092,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.17,\"estimated_latency_ms\":118444,\"estimated_pass_rate\":0.8723,\"estimated_quality\":0.9278,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.387,\"max_latency_ms\":201408,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4514, "category": "review", "complexity": 0.6023, "consequence": 0.7771, "evidence_need": 0.7318, "family": "policy_compliance", "input_tokens": 5280, "request_id": "synthetic-policy_compliance-001236", "reversibility": 0.4051, "risk": 0.5699, "summary": "Review whether database failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.9, "id": "worker_a", "p95_latency_ms": 31723, "pass_rate": 0.4898, "provider_group": "provider_y", "quality": 0.7919, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.568, "id": "worker_b", "p95_latency_ms": 58032, "pass_rate": 0.9588, "provider_group": "provider_x", "quality": 0.9814, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.602, "id": "worker_c", "p95_latency_ms": 60412, "pass_rate": 0.756, "provider_group": "provider_z", "quality": 0.8092, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.651, "max_latency_ms": 234036, "minimum_quality": 0.7103}, "family": "migration_planning", "id": "synthetic-migration_planning-000097", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.74, "estimated_latency_ms": 78545, "estimated_pass_rate": 0.5567, "estimated_quality": 0.8007, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.651, "max_latency_ms": 234036, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5567, "quality": 0.8007, "quality_target_met": true, "utility": 58.79}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.651,\"max_latency_ms\":234036,\"minimum_quality\":0.7103},\"task\":{\"ambiguity\":0.5131,\"category\":\"coding\",\"complexity\":0.7682,\"consequence\":0.6891,\"evidence_need\":0.6967,\"family\":\"migration_planning\",\"input_tokens\":5400,\"reversibility\":0.5564,\"risk\":0.608,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":12.584,\"id\":\"worker_a\",\"p95_latency_ms\":76434,\"pass_rate\":0.9446,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.851,\"id\":\"worker_b\",\"p95_latency_ms\":58888,\"pass_rate\":0.7684,\"provider_group\":\"provider_x\",\"quality\":0.8173,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.889,\"id\":\"worker_c\",\"p95_latency_ms\":19657,\"pass_rate\":0.4881,\"provider_group\":\"provider_y\",\"quality\":0.7869,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.74,\"estimated_latency_ms\":78545,\"estimated_pass_rate\":0.5567,\"estimated_quality\":0.8007,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.651,\"max_latency_ms\":234036,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5131, "category": "coding", "complexity": 0.7682, "consequence": 0.6891, "evidence_need": 0.6967, "family": "migration_planning", "input_tokens": 5400, "request_id": "synthetic-migration_planning-000097", "reversibility": 0.5564, "risk": 0.608, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 12.584, "id": "worker_a", "p95_latency_ms": 76434, "pass_rate": 0.9446, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.851, "id": "worker_b", "p95_latency_ms": 58888, "pass_rate": 0.7684, "provider_group": "provider_x", "quality": 0.8173, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.889, "id": "worker_c", "p95_latency_ms": 19657, "pass_rate": 0.4881, "provider_group": "provider_y", "quality": 0.7869, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 6.339, "max_latency_ms": 244104, "minimum_quality": 0.8456}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001146", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.896, "estimated_latency_ms": 104116, "estimated_pass_rate": 0.685, "estimated_quality": 0.8714, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.339, "max_latency_ms": 244104, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.685, "quality": 0.8714, "quality_target_met": true, "utility": 66.178034}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":6.339,\"max_latency_ms\":244104,\"minimum_quality\":0.8456},\"task\":{\"ambiguity\":0.5405,\"category\":\"review\",\"complexity\":0.4309,\"consequence\":0.8014,\"evidence_need\":0.878,\"family\":\"policy_compliance\",\"input_tokens\":4560,\"reversibility\":0.4112,\"risk\":0.7513,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":1.048,\"id\":\"worker_a\",\"p95_latency_ms\":28301,\"pass_rate\":0.5674,\"provider_group\":\"provider_z\",\"quality\":0.7429,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.848,\"id\":\"worker_b\",\"p95_latency_ms\":75815,\"pass_rate\":0.7306,\"provider_group\":\"provider_y\",\"quality\":0.9114,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.874,\"id\":\"worker_c\",\"p95_latency_ms\":66126,\"pass_rate\":0.9487,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.896,\"estimated_latency_ms\":104116,\"estimated_pass_rate\":0.685,\"estimated_quality\":0.8714,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.339,\"max_latency_ms\":244104,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5405, "category": "review", "complexity": 0.4309, "consequence": 0.8014, "evidence_need": 0.878, "family": "policy_compliance", "input_tokens": 4560, "request_id": "synthetic-policy_compliance-001146", "reversibility": 0.4112, "risk": 0.7513, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 1.048, "id": "worker_a", "p95_latency_ms": 28301, "pass_rate": 0.5674, "provider_group": "provider_z", "quality": 0.7429, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.848, "id": "worker_b", "p95_latency_ms": 75815, "pass_rate": 0.7306, "provider_group": "provider_y", "quality": 0.9114, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.874, "id": "worker_c", "p95_latency_ms": 66126, "pass_rate": 0.9487, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.368, "max_latency_ms": 34463, "minimum_quality": 0.4966}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000264", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.104, "estimated_latency_ms": 22192, "estimated_pass_rate": 0.2695, "estimated_quality": 0.614, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.368, "max_latency_ms": 34463, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2695, "quality": 0.614, "quality_target_met": true, "utility": 39.074247}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.368,\"max_latency_ms\":34463,\"minimum_quality\":0.4966},\"task\":{\"ambiguity\":0.5981,\"category\":\"review\",\"complexity\":0.6872,\"consequence\":0.584,\"evidence_need\":0.8694,\"family\":\"policy_compliance\",\"input_tokens\":4840,\"reversibility\":0.5718,\"risk\":0.7484,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.05,\"id\":\"worker_a\",\"p95_latency_ms\":71671,\"pass_rate\":0.8089,\"provider_group\":\"provider_x\",\"quality\":0.8003,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.076,\"id\":\"worker_b\",\"p95_latency_ms\":81658,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.94,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.104,\"id\":\"worker_c\",\"p95_latency_ms\":22192,\"pass_rate\":0.4533,\"provider_group\":\"provider_y\",\"quality\":0.7523,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.104,\"estimated_latency_ms\":22192,\"estimated_pass_rate\":0.2695,\"estimated_quality\":0.614,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.368,\"max_latency_ms\":34463,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5981, "category": "review", "complexity": 0.6872, "consequence": 0.584, "evidence_need": 0.8694, "family": "policy_compliance", "input_tokens": 4840, "request_id": "synthetic-policy_compliance-000264", "reversibility": 0.5718, "risk": 0.7484, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.05, "id": "worker_a", "p95_latency_ms": 71671, "pass_rate": 0.8089, "provider_group": "provider_x", "quality": 0.8003, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.076, "id": "worker_b", "p95_latency_ms": 81658, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.94, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.104, "id": "worker_c", "p95_latency_ms": 22192, "pass_rate": 0.4533, "provider_group": "provider_y", "quality": 0.7523, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.975, "max_latency_ms": 258966, "minimum_quality": 0.8527}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001224", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.077, "estimated_latency_ms": 199512, "estimated_pass_rate": 0.7656, "estimated_quality": 0.8699, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.975, "max_latency_ms": 258966, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7656, "quality": 0.8699, "quality_target_met": true, "utility": 76.576759}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.975,\"max_latency_ms\":258966,\"minimum_quality\":0.8527},\"task\":{\"ambiguity\":0.6784,\"category\":\"review\",\"complexity\":0.789,\"consequence\":0.7414,\"evidence_need\":0.8273,\"family\":\"policy_compliance\",\"input_tokens\":1680,\"reversibility\":0.6814,\"risk\":0.6997,\"summary\":\"Review whether OAuth callbacks satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":9.94,\"id\":\"worker_a\",\"p95_latency_ms\":89289,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9214,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.986,\"id\":\"worker_b\",\"p95_latency_ms\":32334,\"pass_rate\":0.6109,\"provider_group\":\"provider_z\",\"quality\":0.6987,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.151,\"id\":\"worker_c\",\"p95_latency_ms\":77889,\"pass_rate\":0.7536,\"provider_group\":\"provider_x\",\"quality\":0.8052,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.077,\"estimated_latency_ms\":199512,\"estimated_pass_rate\":0.7656,\"estimated_quality\":0.8699,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.975,\"max_latency_ms\":258966,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6784, "category": "review", "complexity": 0.789, "consequence": 0.7414, "evidence_need": 0.8273, "family": "policy_compliance", "input_tokens": 1680, "request_id": "synthetic-policy_compliance-001224", "reversibility": 0.6814, "risk": 0.6997, "summary": "Review whether OAuth callbacks satisfies a documented retention policy."}, "workers": [{"credits_per_call": 9.94, "id": "worker_a", "p95_latency_ms": 89289, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9214, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.986, "id": "worker_b", "p95_latency_ms": 32334, "pass_rate": 0.6109, "provider_group": "provider_z", "quality": 0.6987, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.151, "id": "worker_c", "p95_latency_ms": 77889, "pass_rate": 0.7536, "provider_group": "provider_x", "quality": 0.8052, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.887, "max_latency_ms": 197549, "minimum_quality": 0.683}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000638", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.045, "estimated_latency_ms": 134365, "estimated_pass_rate": 0.8904, "estimated_quality": 0.9285, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.887, "max_latency_ms": 197549, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8904, "quality": 0.9285, "quality_target_met": true, "utility": 69.584436}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.887,\"max_latency_ms\":197549,\"minimum_quality\":0.683},\"task\":{\"ambiguity\":0.6154,\"category\":\"analysis\",\"complexity\":0.7104,\"consequence\":0.6576,\"evidence_need\":0.4363,\"family\":\"performance_tuning\",\"input_tokens\":5320,\"reversibility\":0.7477,\"risk\":0.2368,\"summary\":\"Choose between throughput and tail-latency changes for API-key rotation.\"},\"workers\":[{\"credits_per_call\":11.287,\"id\":\"worker_a\",\"p95_latency_ms\":67368,\"pass_rate\":0.9787,\"provider_group\":\"provider_z\",\"quality\":0.9857,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.758,\"id\":\"worker_b\",\"p95_latency_ms\":66997,\"pass_rate\":0.8275,\"provider_group\":\"provider_y\",\"quality\":0.9552,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.919,\"id\":\"worker_c\",\"p95_latency_ms\":22402,\"pass_rate\":0.5149,\"provider_group\":\"provider_x\",\"quality\":0.7016,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.045,\"estimated_latency_ms\":134365,\"estimated_pass_rate\":0.8904,\"estimated_quality\":0.9285,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.887,\"max_latency_ms\":197549,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6154, "category": "analysis", "complexity": 0.7104, "consequence": 0.6576, "evidence_need": 0.4363, "family": "performance_tuning", "input_tokens": 5320, "request_id": "synthetic-performance_tuning-000638", "reversibility": 0.7477, "risk": 0.2368, "summary": "Choose between throughput and tail-latency changes for API-key rotation."}, "workers": [{"credits_per_call": 11.287, "id": "worker_a", "p95_latency_ms": 67368, "pass_rate": 0.9787, "provider_group": "provider_z", "quality": 0.9857, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.758, "id": "worker_b", "p95_latency_ms": 66997, "pass_rate": 0.8275, "provider_group": "provider_y", "quality": 0.9552, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.919, "id": "worker_c", "p95_latency_ms": 22402, "pass_rate": 0.5149, "provider_group": "provider_x", "quality": 0.7016, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.315, "max_latency_ms": 191781, "minimum_quality": 0.854}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001152", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.824, "estimated_latency_ms": 149711, "estimated_pass_rate": 0.8026, "estimated_quality": 0.9024, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.315, "max_latency_ms": 191781, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8026, "quality": 0.9024, "quality_target_met": true, "utility": 73.633358}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.315,\"max_latency_ms\":191781,\"minimum_quality\":0.854},\"task\":{\"ambiguity\":0.4326,\"category\":\"review\",\"complexity\":0.5195,\"consequence\":0.8151,\"evidence_need\":0.6498,\"family\":\"policy_compliance\",\"input_tokens\":3720,\"reversibility\":0.4466,\"risk\":0.6975,\"summary\":\"Audit a workflow for credit reservation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.841,\"id\":\"worker_a\",\"p95_latency_ms\":21734,\"pass_rate\":0.6174,\"provider_group\":\"provider_y\",\"quality\":0.7798,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.948,\"id\":\"worker_b\",\"p95_latency_ms\":57810,\"pass_rate\":0.9735,\"provider_group\":\"provider_x\",\"quality\":0.9279,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.035,\"id\":\"worker_c\",\"p95_latency_ms\":70167,\"pass_rate\":0.8257,\"provider_group\":\"provider_z\",\"quality\":0.8757,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.824,\"estimated_latency_ms\":149711,\"estimated_pass_rate\":0.8026,\"estimated_quality\":0.9024,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.315,\"max_latency_ms\":191781,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4326, "category": "review", "complexity": 0.5195, "consequence": 0.8151, "evidence_need": 0.6498, "family": "policy_compliance", "input_tokens": 3720, "request_id": "synthetic-policy_compliance-001152", "reversibility": 0.4466, "risk": 0.6975, "summary": "Audit a workflow for credit reservation against explicit access requirements."}, "workers": [{"credits_per_call": 0.841, "id": "worker_a", "p95_latency_ms": 21734, "pass_rate": 0.6174, "provider_group": "provider_y", "quality": 0.7798, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.948, "id": "worker_b", "p95_latency_ms": 57810, "pass_rate": 0.9735, "provider_group": "provider_x", "quality": 0.9279, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.035, "id": "worker_c", "p95_latency_ms": 70167, "pass_rate": 0.8257, "provider_group": "provider_z", "quality": 0.8757, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.498, "max_latency_ms": 44040, "minimum_quality": 0.5045}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000356", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.997, "estimated_latency_ms": 30103, "estimated_pass_rate": 0.3359, "estimated_quality": 0.5511, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.498, "max_latency_ms": 44040, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3359, "quality": 0.5511, "quality_target_met": true, "utility": 36.82284}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.498,\"max_latency_ms\":44040,\"minimum_quality\":0.5045},\"task\":{\"ambiguity\":0.4299,\"category\":\"analysis\",\"complexity\":0.7484,\"consequence\":0.6975,\"evidence_need\":0.4101,\"family\":\"performance_tuning\",\"input_tokens\":1040,\"reversibility\":0.6339,\"risk\":0.4379,\"summary\":\"Review a proposed optimization for credit reservation for hidden regressions.\"},\"workers\":[{\"credits_per_call\":14.201,\"id\":\"worker_a\",\"p95_latency_ms\":71600,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.995,\"id\":\"worker_b\",\"p95_latency_ms\":81505,\"pass_rate\":0.9377,\"provider_group\":\"provider_y\",\"quality\":0.8675,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.997,\"id\":\"worker_c\",\"p95_latency_ms\":30103,\"pass_rate\":0.4887,\"provider_group\":\"provider_x\",\"quality\":0.6668,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.997,\"estimated_latency_ms\":30103,\"estimated_pass_rate\":0.3359,\"estimated_quality\":0.5511,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.498,\"max_latency_ms\":44040,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4299, "category": "analysis", "complexity": 0.7484, "consequence": 0.6975, "evidence_need": 0.4101, "family": "performance_tuning", "input_tokens": 1040, "request_id": "synthetic-performance_tuning-000356", "reversibility": 0.6339, "risk": 0.4379, "summary": "Review a proposed optimization for credit reservation for hidden regressions."}, "workers": [{"credits_per_call": 14.201, "id": "worker_a", "p95_latency_ms": 71600, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.995, "id": "worker_b", "p95_latency_ms": 81505, "pass_rate": 0.9377, "provider_group": "provider_y", "quality": 0.8675, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.997, "id": "worker_c", "p95_latency_ms": 30103, "pass_rate": 0.4887, "provider_group": "provider_x", "quality": 0.6668, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.541, "max_latency_ms": 35734, "minimum_quality": 0.658}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000474", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.898, "estimated_latency_ms": 26954, "estimated_pass_rate": 0.3116, "estimated_quality": 0.6402, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.541, "max_latency_ms": 35734, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3116, "quality": 0.6402, "quality_target_met": false, "utility": 38.6381}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.541,\"max_latency_ms\":35734,\"minimum_quality\":0.658},\"task\":{\"ambiguity\":0.7315,\"category\":\"review\",\"complexity\":0.5416,\"consequence\":0.7169,\"evidence_need\":0.7952,\"family\":\"policy_compliance\",\"input_tokens\":7800,\"reversibility\":0.6188,\"risk\":0.4893,\"summary\":\"Audit a workflow for provider failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.898,\"id\":\"worker_a\",\"p95_latency_ms\":26954,\"pass_rate\":0.468,\"provider_group\":\"provider_z\",\"quality\":0.7425,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.53,\"id\":\"worker_b\",\"p95_latency_ms\":60391,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9481,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.68,\"id\":\"worker_c\",\"p95_latency_ms\":58488,\"pass_rate\":0.8855,\"provider_group\":\"provider_y\",\"quality\":0.871,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.898,\"estimated_latency_ms\":26954,\"estimated_pass_rate\":0.3116,\"estimated_quality\":0.6402,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.541,\"max_latency_ms\":35734,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7315, "category": "review", "complexity": 0.5416, "consequence": 0.7169, "evidence_need": 0.7952, "family": "policy_compliance", "input_tokens": 7800, "request_id": "synthetic-policy_compliance-000474", "reversibility": 0.6188, "risk": 0.4893, "summary": "Audit a workflow for provider failover against explicit access requirements."}, "workers": [{"credits_per_call": 0.898, "id": "worker_a", "p95_latency_ms": 26954, "pass_rate": 0.468, "provider_group": "provider_z", "quality": 0.7425, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.53, "id": "worker_b", "p95_latency_ms": 60391, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9481, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.68, "id": "worker_c", "p95_latency_ms": 58488, "pass_rate": 0.8855, "provider_group": "provider_y", "quality": 0.871, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.238, "max_latency_ms": 194595, "minimum_quality": 0.9508}, "family": "migration_planning", "id": "synthetic-migration_planning-000883", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.008, "estimated_latency_ms": 141310, "estimated_pass_rate": 0.7121, "estimated_quality": 0.8753, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.238, "max_latency_ms": 194595, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7121, "quality": 0.8753, "quality_target_met": false, "utility": 70.285802}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.238,\"max_latency_ms\":194595,\"minimum_quality\":0.9508},\"task\":{\"ambiguity\":0.7241,\"category\":\"coding\",\"complexity\":0.6029,\"consequence\":0.8385,\"evidence_need\":0.5483,\"family\":\"migration_planning\",\"input_tokens\":4440,\"reversibility\":0.2727,\"risk\":0.6766,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":3.754,\"id\":\"worker_a\",\"p95_latency_ms\":44183,\"pass_rate\":0.7558,\"provider_group\":\"provider_y\",\"quality\":0.8234,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.13,\"id\":\"worker_b\",\"p95_latency_ms\":24280,\"pass_rate\":0.4502,\"provider_group\":\"provider_x\",\"quality\":0.6848,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.124,\"id\":\"worker_c\",\"p95_latency_ms\":72847,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9456,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.008,\"estimated_latency_ms\":141310,\"estimated_pass_rate\":0.7121,\"estimated_quality\":0.8753,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.238,\"max_latency_ms\":194595,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7241, "category": "coding", "complexity": 0.6029, "consequence": 0.8385, "evidence_need": 0.5483, "family": "migration_planning", "input_tokens": 4440, "request_id": "synthetic-migration_planning-000883", "reversibility": 0.2727, "risk": 0.6766, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 3.754, "id": "worker_a", "p95_latency_ms": 44183, "pass_rate": 0.7558, "provider_group": "provider_y", "quality": 0.8234, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.13, "id": "worker_b", "p95_latency_ms": 24280, "pass_rate": 0.4502, "provider_group": "provider_x", "quality": 0.6848, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.124, "id": "worker_c", "p95_latency_ms": 72847, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9456, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.488, "max_latency_ms": 133992, "minimum_quality": 0.7558}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001259", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.857, "estimated_latency_ms": 109146, "estimated_pass_rate": 0.8646, "estimated_quality": 0.89, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.488, "max_latency_ms": 133992, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8646, "quality": 0.89, "quality_target_met": true, "utility": 65.392909}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.488,\"max_latency_ms\":133992,\"minimum_quality\":0.7558},\"task\":{\"ambiguity\":0.4155,\"category\":\"analysis\",\"complexity\":0.7648,\"consequence\":0.5649,\"evidence_need\":0.4684,\"family\":\"performance_tuning\",\"input_tokens\":2640,\"reversibility\":0.8783,\"risk\":0.4954,\"summary\":\"Review a proposed optimization for private credentials for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.839,\"id\":\"worker_a\",\"p95_latency_ms\":86688,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.48,\"id\":\"worker_b\",\"p95_latency_ms\":75529,\"pass_rate\":0.8499,\"provider_group\":\"provider_y\",\"quality\":0.9234,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.018,\"id\":\"worker_c\",\"p95_latency_ms\":22458,\"pass_rate\":0.5511,\"provider_group\":\"provider_z\",\"quality\":0.7145,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.857,\"estimated_latency_ms\":109146,\"estimated_pass_rate\":0.8646,\"estimated_quality\":0.89,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.488,\"max_latency_ms\":133992,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4155, "category": "analysis", "complexity": 0.7648, "consequence": 0.5649, "evidence_need": 0.4684, "family": "performance_tuning", "input_tokens": 2640, "request_id": "synthetic-performance_tuning-001259", "reversibility": 0.8783, "risk": 0.4954, "summary": "Review a proposed optimization for private credentials for hidden regressions."}, "workers": [{"credits_per_call": 11.839, "id": "worker_a", "p95_latency_ms": 86688, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.48, "id": "worker_b", "p95_latency_ms": 75529, "pass_rate": 0.8499, "provider_group": "provider_y", "quality": 0.9234, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.018, "id": "worker_c", "p95_latency_ms": 22458, "pass_rate": 0.5511, "provider_group": "provider_z", "quality": 0.7145, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.428, "max_latency_ms": 219388, "minimum_quality": 0.78}, "family": "migration_planning", "id": "synthetic-migration_planning-000028", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.401, "estimated_latency_ms": 177507, "estimated_pass_rate": 0.6821, "estimated_quality": 0.8677, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.428, "max_latency_ms": 219388, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6821, "quality": 0.8677, "quality_target_met": true, "utility": 70.769945}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.428,\"max_latency_ms\":219388,\"minimum_quality\":0.78},\"task\":{\"ambiguity\":0.5224,\"category\":\"coding\",\"complexity\":0.8686,\"consequence\":0.6187,\"evidence_need\":0.541,\"family\":\"migration_planning\",\"input_tokens\":1560,\"reversibility\":0.544,\"risk\":0.5893,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":3.507,\"id\":\"worker_a\",\"p95_latency_ms\":69391,\"pass_rate\":0.7471,\"provider_group\":\"provider_x\",\"quality\":0.8636,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.715,\"id\":\"worker_b\",\"p95_latency_ms\":80740,\"pass_rate\":0.9104,\"provider_group\":\"provider_y\",\"quality\":0.9501,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.179,\"id\":\"worker_c\",\"p95_latency_ms\":27376,\"pass_rate\":0.4695,\"provider_group\":\"provider_z\",\"quality\":0.6754,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.401,\"estimated_latency_ms\":177507,\"estimated_pass_rate\":0.6821,\"estimated_quality\":0.8677,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.428,\"max_latency_ms\":219388,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5224, "category": "coding", "complexity": 0.8686, "consequence": 0.6187, "evidence_need": 0.541, "family": "migration_planning", "input_tokens": 1560, "request_id": "synthetic-migration_planning-000028", "reversibility": 0.544, "risk": 0.5893, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 3.507, "id": "worker_a", "p95_latency_ms": 69391, "pass_rate": 0.7471, "provider_group": "provider_x", "quality": 0.8636, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.715, "id": "worker_b", "p95_latency_ms": 80740, "pass_rate": 0.9104, "provider_group": "provider_y", "quality": 0.9501, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.179, "id": "worker_c", "p95_latency_ms": 27376, "pass_rate": 0.4695, "provider_group": "provider_z", "quality": 0.6754, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.729, "max_latency_ms": 231410, "minimum_quality": 0.7643}, "family": "migration_planning", "id": "synthetic-migration_planning-000646", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.539, "estimated_latency_ms": 140080, "estimated_pass_rate": 0.8554, "estimated_quality": 0.8857, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.729, "max_latency_ms": 231410, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8554, "quality": 0.8857, "quality_target_met": true, "utility": 65.054403}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.729,\"max_latency_ms\":231410,\"minimum_quality\":0.7643},\"task\":{\"ambiguity\":0.7435,\"category\":\"coding\",\"complexity\":0.7974,\"consequence\":0.4292,\"evidence_need\":0.7544,\"family\":\"migration_planning\",\"input_tokens\":3040,\"reversibility\":0.4787,\"risk\":0.3527,\"summary\":\"Plan a zero-downtime migration of background jobs with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.728,\"id\":\"worker_a\",\"p95_latency_ms\":76521,\"pass_rate\":0.9637,\"provider_group\":\"provider_z\",\"quality\":0.9361,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.057,\"id\":\"worker_b\",\"p95_latency_ms\":19360,\"pass_rate\":0.6175,\"provider_group\":\"provider_x\",\"quality\":0.8024,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.811,\"id\":\"worker_c\",\"p95_latency_ms\":63559,\"pass_rate\":0.7698,\"provider_group\":\"provider_y\",\"quality\":0.8766,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.539,\"estimated_latency_ms\":140080,\"estimated_pass_rate\":0.8554,\"estimated_quality\":0.8857,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.729,\"max_latency_ms\":231410,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7435, "category": "coding", "complexity": 0.7974, "consequence": 0.4292, "evidence_need": 0.7544, "family": "migration_planning", "input_tokens": 3040, "request_id": "synthetic-migration_planning-000646", "reversibility": 0.4787, "risk": 0.3527, "summary": "Plan a zero-downtime migration of background jobs with rollback checkpoints."}, "workers": [{"credits_per_call": 14.728, "id": "worker_a", "p95_latency_ms": 76521, "pass_rate": 0.9637, "provider_group": "provider_z", "quality": 0.9361, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.057, "id": "worker_b", "p95_latency_ms": 19360, "pass_rate": 0.6175, "provider_group": "provider_x", "quality": 0.8024, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.811, "id": "worker_c", "p95_latency_ms": 63559, "pass_rate": 0.7698, "provider_group": "provider_y", "quality": 0.8766, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.857, "max_latency_ms": 125930, "minimum_quality": 0.7128}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000294", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.905, "estimated_latency_ms": 74089, "estimated_pass_rate": 0.6806, "estimated_quality": 0.8427, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.857, "max_latency_ms": 125930, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6806, "quality": 0.8427, "quality_target_met": true, "utility": 66.309392}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.857,\"max_latency_ms\":125930,\"minimum_quality\":0.7128},\"task\":{\"ambiguity\":0.3801,\"category\":\"review\",\"complexity\":0.5892,\"consequence\":0.6334,\"evidence_need\":0.6962,\"family\":\"policy_compliance\",\"input_tokens\":5680,\"reversibility\":0.4242,\"risk\":0.8308,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.547,\"id\":\"worker_a\",\"p95_latency_ms\":65558,\"pass_rate\":0.9387,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.066,\"id\":\"worker_b\",\"p95_latency_ms\":25475,\"pass_rate\":0.5509,\"provider_group\":\"provider_y\",\"quality\":0.7604,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.839,\"id\":\"worker_c\",\"p95_latency_ms\":48614,\"pass_rate\":0.8207,\"provider_group\":\"provider_x\",\"quality\":0.9078,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.905,\"estimated_latency_ms\":74089,\"estimated_pass_rate\":0.6806,\"estimated_quality\":0.8427,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.857,\"max_latency_ms\":125930,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3801, "category": "review", "complexity": 0.5892, "consequence": 0.6334, "evidence_need": 0.6962, "family": "policy_compliance", "input_tokens": 5680, "request_id": "synthetic-policy_compliance-000294", "reversibility": 0.4242, "risk": 0.8308, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.547, "id": "worker_a", "p95_latency_ms": 65558, "pass_rate": 0.9387, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.066, "id": "worker_b", "p95_latency_ms": 25475, "pass_rate": 0.5509, "provider_group": "provider_y", "quality": 0.7604, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.839, "id": "worker_c", "p95_latency_ms": 48614, "pass_rate": 0.8207, "provider_group": "provider_x", "quality": 0.9078, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.433, "max_latency_ms": 33846, "minimum_quality": 0.7593}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000515", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.795, "estimated_latency_ms": 20527, "estimated_pass_rate": 0.418, "estimated_quality": 0.6013, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.433, "max_latency_ms": 33846, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.418, "quality": 0.6013, "quality_target_met": false, "utility": 35.24213}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.433,\"max_latency_ms\":33846,\"minimum_quality\":0.7593},\"task\":{\"ambiguity\":0.3644,\"category\":\"analysis\",\"complexity\":0.5767,\"consequence\":0.3363,\"evidence_need\":0.6888,\"family\":\"performance_tuning\",\"input_tokens\":7600,\"reversibility\":0.6657,\"risk\":0.541,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.599,\"id\":\"worker_a\",\"p95_latency_ms\":77468,\"pass_rate\":0.9916,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.795,\"id\":\"worker_b\",\"p95_latency_ms\":20527,\"pass_rate\":0.5548,\"provider_group\":\"provider_z\",\"quality\":0.689,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.271,\"id\":\"worker_c\",\"p95_latency_ms\":50464,\"pass_rate\":0.9064,\"provider_group\":\"provider_y\",\"quality\":0.9704,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.795,\"estimated_latency_ms\":20527,\"estimated_pass_rate\":0.418,\"estimated_quality\":0.6013,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.433,\"max_latency_ms\":33846,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3644, "category": "analysis", "complexity": 0.5767, "consequence": 0.3363, "evidence_need": 0.6888, "family": "performance_tuning", "input_tokens": 7600, "request_id": "synthetic-performance_tuning-000515", "reversibility": 0.6657, "risk": 0.541, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 11.599, "id": "worker_a", "p95_latency_ms": 77468, "pass_rate": 0.9916, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.795, "id": "worker_b", "p95_latency_ms": 20527, "pass_rate": 0.5548, "provider_group": "provider_z", "quality": 0.689, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.271, "id": "worker_c", "p95_latency_ms": 50464, "pass_rate": 0.9064, "provider_group": "provider_y", "quality": 0.9704, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 4.388, "max_latency_ms": 171338, "minimum_quality": 0.7793}, "family": "migration_planning", "id": "synthetic-migration_planning-000034", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.047, "estimated_latency_ms": 73804, "estimated_pass_rate": 0.7351, "estimated_quality": 0.767, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 4.388, "max_latency_ms": 171338, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7351, "quality": 0.767, "quality_target_met": false, "utility": 53.391902}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":4.388,\"max_latency_ms\":171338,\"minimum_quality\":0.7793},\"task\":{\"ambiguity\":0.4773,\"category\":\"coding\",\"complexity\":0.8977,\"consequence\":0.6507,\"evidence_need\":0.4028,\"family\":\"migration_planning\",\"input_tokens\":2000,\"reversibility\":0.4472,\"risk\":0.5072,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.842,\"id\":\"worker_a\",\"p95_latency_ms\":21148,\"pass_rate\":0.4937,\"provider_group\":\"provider_y\",\"quality\":0.6956,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.095,\"id\":\"worker_b\",\"p95_latency_ms\":84033,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9653,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.205,\"id\":\"worker_c\",\"p95_latency_ms\":52656,\"pass_rate\":0.8624,\"provider_group\":\"provider_z\",\"quality\":0.8777,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.047,\"estimated_latency_ms\":73804,\"estimated_pass_rate\":0.7351,\"estimated_quality\":0.767,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":4.388,\"max_latency_ms\":171338,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4773, "category": "coding", "complexity": 0.8977, "consequence": 0.6507, "evidence_need": 0.4028, "family": "migration_planning", "input_tokens": 2000, "request_id": "synthetic-migration_planning-000034", "reversibility": 0.4472, "risk": 0.5072, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 0.842, "id": "worker_a", "p95_latency_ms": 21148, "pass_rate": 0.4937, "provider_group": "provider_y", "quality": 0.6956, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.095, "id": "worker_b", "p95_latency_ms": 84033, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9653, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.205, "id": "worker_c", "p95_latency_ms": 52656, "pass_rate": 0.8624, "provider_group": "provider_z", "quality": 0.8777, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.911, "max_latency_ms": 214875, "minimum_quality": 0.7848}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000411", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.343, "estimated_latency_ms": 164226, "estimated_pass_rate": 0.7388, "estimated_quality": 0.8981, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.911, "max_latency_ms": 214875, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7388, "quality": 0.8981, "quality_target_met": true, "utility": 75.447094}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.911,\"max_latency_ms\":214875,\"minimum_quality\":0.7848},\"task\":{\"ambiguity\":0.6714,\"category\":\"review\",\"complexity\":0.7502,\"consequence\":0.62,\"evidence_need\":0.6704,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.5994,\"risk\":0.6332,\"summary\":\"Review whether rate limiting satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.989,\"id\":\"worker_a\",\"p95_latency_ms\":19544,\"pass_rate\":0.5047,\"provider_group\":\"provider_z\",\"quality\":0.7445,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.933,\"id\":\"worker_b\",\"p95_latency_ms\":67890,\"pass_rate\":0.7804,\"provider_group\":\"provider_x\",\"quality\":0.8642,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.421,\"id\":\"worker_c\",\"p95_latency_ms\":76792,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9561,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.343,\"estimated_latency_ms\":164226,\"estimated_pass_rate\":0.7388,\"estimated_quality\":0.8981,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.911,\"max_latency_ms\":214875,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6714, "category": "review", "complexity": 0.7502, "consequence": 0.62, "evidence_need": 0.6704, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-000411", "reversibility": 0.5994, "risk": 0.6332, "summary": "Review whether rate limiting satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.989, "id": "worker_a", "p95_latency_ms": 19544, "pass_rate": 0.5047, "provider_group": "provider_z", "quality": 0.7445, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.933, "id": "worker_b", "p95_latency_ms": 67890, "pass_rate": 0.7804, "provider_group": "provider_x", "quality": 0.8642, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.421, "id": "worker_c", "p95_latency_ms": 76792, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9561, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.528, "max_latency_ms": 39086, "minimum_quality": 0.6388}, "family": "migration_planning", "id": "synthetic-migration_planning-000712", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.158, "estimated_latency_ms": 25020, "estimated_pass_rate": 0.2884, "estimated_quality": 0.6164, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.528, "max_latency_ms": 39086, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2884, "quality": 0.6164, "quality_target_met": false, "utility": 35.723655}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.528,\"max_latency_ms\":39086,\"minimum_quality\":0.6388},\"task\":{\"ambiguity\":0.7355,\"category\":\"coding\",\"complexity\":0.9306,\"consequence\":0.5152,\"evidence_need\":0.7051,\"family\":\"migration_planning\",\"input_tokens\":6880,\"reversibility\":0.3097,\"risk\":0.466,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":3.48,\"id\":\"worker_a\",\"p95_latency_ms\":81401,\"pass_rate\":0.7353,\"provider_group\":\"provider_y\",\"quality\":0.8375,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.158,\"id\":\"worker_b\",\"p95_latency_ms\":25020,\"pass_rate\":0.4873,\"provider_group\":\"provider_x\",\"quality\":0.7876,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.613,\"id\":\"worker_c\",\"p95_latency_ms\":63719,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9148,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.158,\"estimated_latency_ms\":25020,\"estimated_pass_rate\":0.2884,\"estimated_quality\":0.6164,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.528,\"max_latency_ms\":39086,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7355, "category": "coding", "complexity": 0.9306, "consequence": 0.5152, "evidence_need": 0.7051, "family": "migration_planning", "input_tokens": 6880, "request_id": "synthetic-migration_planning-000712", "reversibility": 0.3097, "risk": 0.466, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 3.48, "id": "worker_a", "p95_latency_ms": 81401, "pass_rate": 0.7353, "provider_group": "provider_y", "quality": 0.8375, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.158, "id": "worker_b", "p95_latency_ms": 25020, "pass_rate": 0.4873, "provider_group": "provider_x", "quality": 0.7876, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.613, "id": "worker_c", "p95_latency_ms": 63719, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9148, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.135, "max_latency_ms": 32008, "minimum_quality": 0.6452}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000882", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.99, "estimated_latency_ms": 25664, "estimated_pass_rate": 0.3743, "estimated_quality": 0.6228, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.135, "max_latency_ms": 32008, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3743, "quality": 0.6228, "quality_target_met": false, "utility": 36.567875}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.135,\"max_latency_ms\":32008,\"minimum_quality\":0.6452},\"task\":{\"ambiguity\":0.4903,\"category\":\"review\",\"complexity\":0.843,\"consequence\":0.5893,\"evidence_need\":0.8545,\"family\":\"policy_compliance\",\"input_tokens\":7080,\"reversibility\":0.543,\"risk\":0.4501,\"summary\":\"Review whether usage metering satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.418,\"id\":\"worker_a\",\"p95_latency_ms\":72764,\"pass_rate\":0.8264,\"provider_group\":\"provider_z\",\"quality\":0.8137,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.99,\"id\":\"worker_b\",\"p95_latency_ms\":25664,\"pass_rate\":0.5434,\"provider_group\":\"provider_y\",\"quality\":0.7606,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.729,\"id\":\"worker_c\",\"p95_latency_ms\":71729,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9779,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.99,\"estimated_latency_ms\":25664,\"estimated_pass_rate\":0.3743,\"estimated_quality\":0.6228,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.135,\"max_latency_ms\":32008,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4903, "category": "review", "complexity": 0.843, "consequence": 0.5893, "evidence_need": 0.8545, "family": "policy_compliance", "input_tokens": 7080, "request_id": "synthetic-policy_compliance-000882", "reversibility": 0.543, "risk": 0.4501, "summary": "Review whether usage metering satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.418, "id": "worker_a", "p95_latency_ms": 72764, "pass_rate": 0.8264, "provider_group": "provider_z", "quality": 0.8137, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.99, "id": "worker_b", "p95_latency_ms": 25664, "pass_rate": 0.5434, "provider_group": "provider_y", "quality": 0.7606, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.729, "id": "worker_c", "p95_latency_ms": 71729, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9779, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.298, "max_latency_ms": 120273, "minimum_quality": 0.8615}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000258", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.995, "estimated_latency_ms": 79053, "estimated_pass_rate": 0.7731, "estimated_quality": 0.8759, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.298, "max_latency_ms": 120273, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7731, "quality": 0.8759, "quality_target_met": true, "utility": 63.89118}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.298,\"max_latency_ms\":120273,\"minimum_quality\":0.8615},\"task\":{\"ambiguity\":0.5754,\"category\":\"review\",\"complexity\":0.6971,\"consequence\":0.7793,\"evidence_need\":0.6648,\"family\":\"policy_compliance\",\"input_tokens\":5560,\"reversibility\":0.6373,\"risk\":0.7697,\"summary\":\"Review whether access control satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.891,\"id\":\"worker_a\",\"p95_latency_ms\":61513,\"pass_rate\":0.7267,\"provider_group\":\"provider_z\",\"quality\":0.8222,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.841,\"id\":\"worker_b\",\"p95_latency_ms\":18291,\"pass_rate\":0.5214,\"provider_group\":\"provider_x\",\"quality\":0.7095,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.154,\"id\":\"worker_c\",\"p95_latency_ms\":60762,\"pass_rate\":0.8984,\"provider_group\":\"provider_y\",\"quality\":0.9911,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.995,\"estimated_latency_ms\":79053,\"estimated_pass_rate\":0.7731,\"estimated_quality\":0.8759,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.298,\"max_latency_ms\":120273,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5754, "category": "review", "complexity": 0.6971, "consequence": 0.7793, "evidence_need": 0.6648, "family": "policy_compliance", "input_tokens": 5560, "request_id": "synthetic-policy_compliance-000258", "reversibility": 0.6373, "risk": 0.7697, "summary": "Review whether access control satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.891, "id": "worker_a", "p95_latency_ms": 61513, "pass_rate": 0.7267, "provider_group": "provider_z", "quality": 0.8222, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.841, "id": "worker_b", "p95_latency_ms": 18291, "pass_rate": 0.5214, "provider_group": "provider_x", "quality": 0.7095, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.154, "id": "worker_c", "p95_latency_ms": 60762, "pass_rate": 0.8984, "provider_group": "provider_y", "quality": 0.9911, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.578, "max_latency_ms": 242876, "minimum_quality": 0.8048}, "family": "migration_planning", "id": "synthetic-migration_planning-000988", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.168, "estimated_latency_ms": 172039, "estimated_pass_rate": 0.7462, "estimated_quality": 0.9102, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 16.578, "max_latency_ms": 242876, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7462, "quality": 0.9102, "quality_target_met": true, "utility": 74.161732}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.578,\"max_latency_ms\":242876,\"minimum_quality\":0.8048},\"task\":{\"ambiguity\":0.578,\"category\":\"coding\",\"complexity\":0.8015,\"consequence\":0.4015,\"evidence_need\":0.7145,\"family\":\"migration_planning\",\"input_tokens\":6880,\"reversibility\":0.6419,\"risk\":0.6364,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":11.123,\"id\":\"worker_a\",\"p95_latency_ms\":70211,\"pass_rate\":0.93,\"provider_group\":\"provider_y\",\"quality\":0.9777,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.05,\"id\":\"worker_b\",\"p95_latency_ms\":72746,\"pass_rate\":0.8187,\"provider_group\":\"provider_x\",\"quality\":0.8175,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.995,\"id\":\"worker_c\",\"p95_latency_ms\":29082,\"pass_rate\":0.591,\"provider_group\":\"provider_z\",\"quality\":0.7739,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.168,\"estimated_latency_ms\":172039,\"estimated_pass_rate\":0.7462,\"estimated_quality\":0.9102,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.578,\"max_latency_ms\":242876,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.578, "category": "coding", "complexity": 0.8015, "consequence": 0.4015, "evidence_need": 0.7145, "family": "migration_planning", "input_tokens": 6880, "request_id": "synthetic-migration_planning-000988", "reversibility": 0.6419, "risk": 0.6364, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 11.123, "id": "worker_a", "p95_latency_ms": 70211, "pass_rate": 0.93, "provider_group": "provider_y", "quality": 0.9777, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.05, "id": "worker_b", "p95_latency_ms": 72746, "pass_rate": 0.8187, "provider_group": "provider_x", "quality": 0.8175, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.995, "id": "worker_c", "p95_latency_ms": 29082, "pass_rate": 0.591, "provider_group": "provider_z", "quality": 0.7739, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.469, "max_latency_ms": 22044, "minimum_quality": 0.7541}, "family": "migration_planning", "id": "synthetic-migration_planning-000793", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.992, "estimated_latency_ms": 18650, "estimated_pass_rate": 0.4153, "estimated_quality": 0.5439, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.469, "max_latency_ms": 22044, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4153, "quality": 0.5439, "quality_target_met": false, "utility": 25.911116}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.469,\"max_latency_ms\":22044,\"minimum_quality\":0.7541},\"task\":{\"ambiguity\":0.6004,\"category\":\"coding\",\"complexity\":0.7694,\"consequence\":0.6369,\"evidence_need\":0.4136,\"family\":\"migration_planning\",\"input_tokens\":520,\"reversibility\":0.614,\"risk\":0.5059,\"summary\":\"Design compatibility testing for a data retention schema migration.\"},\"workers\":[{\"credits_per_call\":0.992,\"id\":\"worker_a\",\"p95_latency_ms\":18650,\"pass_rate\":0.5889,\"provider_group\":\"provider_y\",\"quality\":0.6798,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.037,\"id\":\"worker_b\",\"p95_latency_ms\":73569,\"pass_rate\":0.8876,\"provider_group\":\"provider_x\",\"quality\":0.9002,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.715,\"id\":\"worker_c\",\"p95_latency_ms\":67355,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.992,\"estimated_latency_ms\":18650,\"estimated_pass_rate\":0.4153,\"estimated_quality\":0.5439,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.469,\"max_latency_ms\":22044,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6004, "category": "coding", "complexity": 0.7694, "consequence": 0.6369, "evidence_need": 0.4136, "family": "migration_planning", "input_tokens": 520, "request_id": "synthetic-migration_planning-000793", "reversibility": 0.614, "risk": 0.5059, "summary": "Design compatibility testing for a data retention schema migration."}, "workers": [{"credits_per_call": 0.992, "id": "worker_a", "p95_latency_ms": 18650, "pass_rate": 0.5889, "provider_group": "provider_y", "quality": 0.6798, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.037, "id": "worker_b", "p95_latency_ms": 73569, "pass_rate": 0.8876, "provider_group": "provider_x", "quality": 0.9002, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.715, "id": "worker_c", "p95_latency_ms": 67355, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.475, "max_latency_ms": 195522, "minimum_quality": 0.9241}, "family": "migration_planning", "id": "synthetic-migration_planning-001249", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.597, "estimated_latency_ms": 169299, "estimated_pass_rate": 0.6998, "estimated_quality": 0.9116, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.475, "max_latency_ms": 195522, "policy": "strict", "reason_codes": ["best_feasible_below_target", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6998, "quality": 0.9116, "quality_target_met": false, "utility": 72.048865}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.475,\"max_latency_ms\":195522,\"minimum_quality\":0.9241},\"task\":{\"ambiguity\":0.5663,\"category\":\"coding\",\"complexity\":0.616,\"consequence\":0.7393,\"evidence_need\":0.4725,\"family\":\"migration_planning\",\"input_tokens\":5560,\"reversibility\":0.5503,\"risk\":0.739,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":1.097,\"id\":\"worker_a\",\"p95_latency_ms\":22358,\"pass_rate\":0.4554,\"provider_group\":\"provider_x\",\"quality\":0.7786,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.555,\"id\":\"worker_b\",\"p95_latency_ms\":72757,\"pass_rate\":0.9137,\"provider_group\":\"provider_y\",\"quality\":0.99,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.945,\"id\":\"worker_c\",\"p95_latency_ms\":74184,\"pass_rate\":0.7357,\"provider_group\":\"provider_z\",\"quality\":0.8879,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.597,\"estimated_latency_ms\":169299,\"estimated_pass_rate\":0.6998,\"estimated_quality\":0.9116,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.475,\"max_latency_ms\":195522,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5663, "category": "coding", "complexity": 0.616, "consequence": 0.7393, "evidence_need": 0.4725, "family": "migration_planning", "input_tokens": 5560, "request_id": "synthetic-migration_planning-001249", "reversibility": 0.5503, "risk": 0.739, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 1.097, "id": "worker_a", "p95_latency_ms": 22358, "pass_rate": 0.4554, "provider_group": "provider_x", "quality": 0.7786, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.555, "id": "worker_b", "p95_latency_ms": 72757, "pass_rate": 0.9137, "provider_group": "provider_y", "quality": 0.99, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.945, "id": "worker_c", "p95_latency_ms": 74184, "pass_rate": 0.7357, "provider_group": "provider_z", "quality": 0.8879, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.351, "max_latency_ms": 37065, "minimum_quality": 0.6127}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000273", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.186, "estimated_latency_ms": 20825, "estimated_pass_rate": 0.3752, "estimated_quality": 0.5896, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.351, "max_latency_ms": 37065, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3752, "quality": 0.5896, "quality_target_met": false, "utility": 25.581378}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.351,\"max_latency_ms\":37065,\"minimum_quality\":0.6127},\"task\":{\"ambiguity\":0.3617,\"category\":\"review\",\"complexity\":0.6673,\"consequence\":0.6258,\"evidence_need\":0.6827,\"family\":\"policy_compliance\",\"input_tokens\":5440,\"reversibility\":0.3816,\"risk\":0.8371,\"summary\":\"Identify evidence required before approving deployment rollback for production.\"},\"workers\":[{\"credits_per_call\":3.13,\"id\":\"worker_a\",\"p95_latency_ms\":56109,\"pass_rate\":0.8009,\"provider_group\":\"provider_x\",\"quality\":0.8842,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.186,\"id\":\"worker_b\",\"p95_latency_ms\":20825,\"pass_rate\":0.546,\"provider_group\":\"provider_z\",\"quality\":0.7148,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.25,\"id\":\"worker_c\",\"p95_latency_ms\":91869,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9465,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.186,\"estimated_latency_ms\":20825,\"estimated_pass_rate\":0.3752,\"estimated_quality\":0.5896,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.351,\"max_latency_ms\":37065,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3617, "category": "review", "complexity": 0.6673, "consequence": 0.6258, "evidence_need": 0.6827, "family": "policy_compliance", "input_tokens": 5440, "request_id": "synthetic-policy_compliance-000273", "reversibility": 0.3816, "risk": 0.8371, "summary": "Identify evidence required before approving deployment rollback for production."}, "workers": [{"credits_per_call": 3.13, "id": "worker_a", "p95_latency_ms": 56109, "pass_rate": 0.8009, "provider_group": "provider_x", "quality": 0.8842, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.186, "id": "worker_b", "p95_latency_ms": 20825, "pass_rate": 0.546, "provider_group": "provider_z", "quality": 0.7148, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.25, "id": "worker_c", "p95_latency_ms": 91869, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9465, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 1.884, "max_latency_ms": 35361, "minimum_quality": 0.6912}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000291", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.078, "estimated_latency_ms": 20867, "estimated_pass_rate": 0.3462, "estimated_quality": 0.5843, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.884, "max_latency_ms": 35361, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3462, "quality": 0.5843, "quality_target_met": false, "utility": 33.038582}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":1.884,\"max_latency_ms\":35361,\"minimum_quality\":0.6912},\"task\":{\"ambiguity\":0.7248,\"category\":\"review\",\"complexity\":0.447,\"consequence\":0.5763,\"evidence_need\":0.7706,\"family\":\"policy_compliance\",\"input_tokens\":3200,\"reversibility\":0.6975,\"risk\":0.5896,\"summary\":\"Review whether database failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.563,\"id\":\"worker_a\",\"p95_latency_ms\":74241,\"pass_rate\":0.9675,\"provider_group\":\"provider_y\",\"quality\":0.989,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.078,\"id\":\"worker_b\",\"p95_latency_ms\":20867,\"pass_rate\":0.4994,\"provider_group\":\"provider_x\",\"quality\":0.6763,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.049,\"id\":\"worker_c\",\"p95_latency_ms\":71796,\"pass_rate\":0.7848,\"provider_group\":\"provider_z\",\"quality\":0.8708,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.078,\"estimated_latency_ms\":20867,\"estimated_pass_rate\":0.3462,\"estimated_quality\":0.5843,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.884,\"max_latency_ms\":35361,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7248, "category": "review", "complexity": 0.447, "consequence": 0.5763, "evidence_need": 0.7706, "family": "policy_compliance", "input_tokens": 3200, "request_id": "synthetic-policy_compliance-000291", "reversibility": 0.6975, "risk": 0.5896, "summary": "Review whether database failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.563, "id": "worker_a", "p95_latency_ms": 74241, "pass_rate": 0.9675, "provider_group": "provider_y", "quality": 0.989, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.078, "id": "worker_b", "p95_latency_ms": 20867, "pass_rate": 0.4994, "provider_group": "provider_x", "quality": 0.6763, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.049, "id": "worker_c", "p95_latency_ms": 71796, "pass_rate": 0.7848, "provider_group": "provider_z", "quality": 0.8708, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.664, "max_latency_ms": 238699, "minimum_quality": 0.9445}, "family": "migration_planning", "id": "synthetic-migration_planning-001072", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.744, "estimated_latency_ms": 181063, "estimated_pass_rate": 0.763, "estimated_quality": 0.9055, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.664, "max_latency_ms": 238699, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.763, "quality": 0.9055, "quality_target_met": false, "utility": 72.117155}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.664,\"max_latency_ms\":238699,\"minimum_quality\":0.9445},\"task\":{\"ambiguity\":0.7285,\"category\":\"coding\",\"complexity\":0.7153,\"consequence\":0.6703,\"evidence_need\":0.5717,\"family\":\"migration_planning\",\"input_tokens\":5480,\"reversibility\":0.304,\"risk\":0.5281,\"summary\":\"Review a staged migration for tenant-scoped caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.399,\"id\":\"worker_a\",\"p95_latency_ms\":86182,\"pass_rate\":0.9668,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.342,\"id\":\"worker_b\",\"p95_latency_ms\":70404,\"pass_rate\":0.7533,\"provider_group\":\"provider_y\",\"quality\":0.9115,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.003,\"id\":\"worker_c\",\"p95_latency_ms\":24477,\"pass_rate\":0.6051,\"provider_group\":\"provider_x\",\"quality\":0.6931,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.744,\"estimated_latency_ms\":181063,\"estimated_pass_rate\":0.763,\"estimated_quality\":0.9055,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.664,\"max_latency_ms\":238699,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7285, "category": "coding", "complexity": 0.7153, "consequence": 0.6703, "evidence_need": 0.5717, "family": "migration_planning", "input_tokens": 5480, "request_id": "synthetic-migration_planning-001072", "reversibility": 0.304, "risk": 0.5281, "summary": "Review a staged migration for tenant-scoped caching and identify irreversible steps."}, "workers": [{"credits_per_call": 11.399, "id": "worker_a", "p95_latency_ms": 86182, "pass_rate": 0.9668, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.342, "id": "worker_b", "p95_latency_ms": 70404, "pass_rate": 0.7533, "provider_group": "provider_y", "quality": 0.9115, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.003, "id": "worker_c", "p95_latency_ms": 24477, "pass_rate": 0.6051, "provider_group": "provider_x", "quality": 0.6931, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.131, "max_latency_ms": 171352, "minimum_quality": 0.8549}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000195", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.136, "estimated_latency_ms": 159889, "estimated_pass_rate": 0.7402, "estimated_quality": 0.876, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.131, "max_latency_ms": 171352, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7402, "quality": 0.876, "quality_target_met": true, "utility": 72.264992}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.131,\"max_latency_ms\":171352,\"minimum_quality\":0.8549},\"task\":{\"ambiguity\":0.7172,\"category\":\"review\",\"complexity\":0.8098,\"consequence\":0.7862,\"evidence_need\":0.5637,\"family\":\"policy_compliance\",\"input_tokens\":320,\"reversibility\":0.4079,\"risk\":0.5119,\"summary\":\"Audit a workflow for artifact storage against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.142,\"id\":\"worker_a\",\"p95_latency_ms\":71673,\"pass_rate\":0.903,\"provider_group\":\"provider_x\",\"quality\":0.9215,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.028,\"id\":\"worker_b\",\"p95_latency_ms\":32875,\"pass_rate\":0.5592,\"provider_group\":\"provider_z\",\"quality\":0.7017,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.966,\"id\":\"worker_c\",\"p95_latency_ms\":55341,\"pass_rate\":0.8832,\"provider_group\":\"provider_y\",\"quality\":0.9077,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.136,\"estimated_latency_ms\":159889,\"estimated_pass_rate\":0.7402,\"estimated_quality\":0.876,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.131,\"max_latency_ms\":171352,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7172, "category": "review", "complexity": 0.8098, "consequence": 0.7862, "evidence_need": 0.5637, "family": "policy_compliance", "input_tokens": 320, "request_id": "synthetic-policy_compliance-000195", "reversibility": 0.4079, "risk": 0.5119, "summary": "Audit a workflow for artifact storage against explicit access requirements."}, "workers": [{"credits_per_call": 13.142, "id": "worker_a", "p95_latency_ms": 71673, "pass_rate": 0.903, "provider_group": "provider_x", "quality": 0.9215, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.028, "id": "worker_b", "p95_latency_ms": 32875, "pass_rate": 0.5592, "provider_group": "provider_z", "quality": 0.7017, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.966, "id": "worker_c", "p95_latency_ms": 55341, "pass_rate": 0.8832, "provider_group": "provider_y", "quality": 0.9077, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 14.344, "max_latency_ms": 93298, "minimum_quality": 0.7677}, "family": "migration_planning", "id": "synthetic-migration_planning-000103", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.233, "estimated_latency_ms": 80223, "estimated_pass_rate": 0.6993, "estimated_quality": 0.8735, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.344, "max_latency_ms": 93298, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6993, "quality": 0.8735, "quality_target_met": true, "utility": 61.456426}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":14.344,\"max_latency_ms\":93298,\"minimum_quality\":0.7677},\"task\":{\"ambiguity\":0.5053,\"category\":\"coding\",\"complexity\":0.9375,\"consequence\":0.5461,\"evidence_need\":0.7653,\"family\":\"migration_planning\",\"input_tokens\":680,\"reversibility\":0.6332,\"risk\":0.6024,\"summary\":\"Plan a zero-downtime migration of API-key rotation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":13.109,\"id\":\"worker_a\",\"p95_latency_ms\":50133,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9537,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.124,\"id\":\"worker_b\",\"p95_latency_ms\":30090,\"pass_rate\":0.5453,\"provider_group\":\"provider_x\",\"quality\":0.7595,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.817,\"id\":\"worker_c\",\"p95_latency_ms\":44823,\"pass_rate\":0.7427,\"provider_group\":\"provider_y\",\"quality\":0.8619,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.233,\"estimated_latency_ms\":80223,\"estimated_pass_rate\":0.6993,\"estimated_quality\":0.8735,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.344,\"max_latency_ms\":93298,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5053, "category": "coding", "complexity": 0.9375, "consequence": 0.5461, "evidence_need": 0.7653, "family": "migration_planning", "input_tokens": 680, "request_id": "synthetic-migration_planning-000103", "reversibility": 0.6332, "risk": 0.6024, "summary": "Plan a zero-downtime migration of API-key rotation with rollback checkpoints."}, "workers": [{"credits_per_call": 13.109, "id": "worker_a", "p95_latency_ms": 50133, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9537, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.124, "id": "worker_b", "p95_latency_ms": 30090, "pass_rate": 0.5453, "provider_group": "provider_x", "quality": 0.7595, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.817, "id": "worker_c", "p95_latency_ms": 44823, "pass_rate": 0.7427, "provider_group": "provider_y", "quality": 0.8619, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.629, "max_latency_ms": 202220, "minimum_quality": 0.8185}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000795", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.749, "estimated_latency_ms": 178180, "estimated_pass_rate": 0.7472, "estimated_quality": 0.9194, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.629, "max_latency_ms": 202220, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7472, "quality": 0.9194, "quality_target_met": true, "utility": 73.796432}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.629,\"max_latency_ms\":202220,\"minimum_quality\":0.8185},\"task\":{\"ambiguity\":0.5574,\"category\":\"review\",\"complexity\":0.6877,\"consequence\":0.6495,\"evidence_need\":0.934,\"family\":\"policy_compliance\",\"input_tokens\":4400,\"reversibility\":0.4228,\"risk\":0.5035,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":0.812,\"id\":\"worker_a\",\"p95_latency_ms\":28795,\"pass_rate\":0.6029,\"provider_group\":\"provider_y\",\"quality\":0.7247,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.437,\"id\":\"worker_b\",\"p95_latency_ms\":83655,\"pass_rate\":0.9335,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.5,\"id\":\"worker_c\",\"p95_latency_ms\":65730,\"pass_rate\":0.7683,\"provider_group\":\"provider_z\",\"quality\":0.843,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.749,\"estimated_latency_ms\":178180,\"estimated_pass_rate\":0.7472,\"estimated_quality\":0.9194,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.629,\"max_latency_ms\":202220,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5574, "category": "review", "complexity": 0.6877, "consequence": 0.6495, "evidence_need": 0.934, "family": "policy_compliance", "input_tokens": 4400, "request_id": "synthetic-policy_compliance-000795", "reversibility": 0.4228, "risk": 0.5035, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 0.812, "id": "worker_a", "p95_latency_ms": 28795, "pass_rate": 0.6029, "provider_group": "provider_y", "quality": 0.7247, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.437, "id": "worker_b", "p95_latency_ms": 83655, "pass_rate": 0.9335, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.5, "id": "worker_c", "p95_latency_ms": 65730, "pass_rate": 0.7683, "provider_group": "provider_z", "quality": 0.843, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.082, "max_latency_ms": 36030, "minimum_quality": 0.6505}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000704", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.135, "estimated_latency_ms": 28176, "estimated_pass_rate": 0.4169, "estimated_quality": 0.5656, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.082, "max_latency_ms": 36030, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4169, "quality": 0.5656, "quality_target_met": false, "utility": 33.75499}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.082,\"max_latency_ms\":36030,\"minimum_quality\":0.6505},\"task\":{\"ambiguity\":0.4104,\"category\":\"analysis\",\"complexity\":0.5502,\"consequence\":0.2965,\"evidence_need\":0.6336,\"family\":\"performance_tuning\",\"input_tokens\":4760,\"reversibility\":0.5884,\"risk\":0.4378,\"summary\":\"Review a proposed optimization for model routing for hidden regressions.\"},\"workers\":[{\"credits_per_call\":10.716,\"id\":\"worker_a\",\"p95_latency_ms\":61490,\"pass_rate\":0.973,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.135,\"id\":\"worker_b\",\"p95_latency_ms\":28176,\"pass_rate\":0.5458,\"provider_group\":\"provider_y\",\"quality\":0.6442,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":2.957,\"id\":\"worker_c\",\"p95_latency_ms\":73849,\"pass_rate\":0.9299,\"provider_group\":\"provider_z\",\"quality\":0.9008,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.135,\"estimated_latency_ms\":28176,\"estimated_pass_rate\":0.4169,\"estimated_quality\":0.5656,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.082,\"max_latency_ms\":36030,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4104, "category": "analysis", "complexity": 0.5502, "consequence": 0.2965, "evidence_need": 0.6336, "family": "performance_tuning", "input_tokens": 4760, "request_id": "synthetic-performance_tuning-000704", "reversibility": 0.5884, "risk": 0.4378, "summary": "Review a proposed optimization for model routing for hidden regressions."}, "workers": [{"credits_per_call": 10.716, "id": "worker_a", "p95_latency_ms": 61490, "pass_rate": 0.973, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.135, "id": "worker_b", "p95_latency_ms": 28176, "pass_rate": 0.5458, "provider_group": "provider_y", "quality": 0.6442, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 2.957, "id": "worker_c", "p95_latency_ms": 73849, "pass_rate": 0.9299, "provider_group": "provider_z", "quality": 0.9008, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.364, "max_latency_ms": 179252, "minimum_quality": 0.8347}, "family": "migration_planning", "id": "synthetic-migration_planning-000757", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.757, "estimated_latency_ms": 94450, "estimated_pass_rate": 0.6842, "estimated_quality": 0.9003, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.364, "max_latency_ms": 179252, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6842, "quality": 0.9003, "quality_target_met": true, "utility": 64.129775}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.364,\"max_latency_ms\":179252,\"minimum_quality\":0.8347},\"task\":{\"ambiguity\":0.5381,\"category\":\"coding\",\"complexity\":0.8834,\"consequence\":0.4907,\"evidence_need\":0.8294,\"family\":\"migration_planning\",\"input_tokens\":6960,\"reversibility\":0.3377,\"risk\":0.4405,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":0.859,\"id\":\"worker_a\",\"p95_latency_ms\":26814,\"pass_rate\":0.5404,\"provider_group\":\"provider_z\",\"quality\":0.8043,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.898,\"id\":\"worker_b\",\"p95_latency_ms\":67636,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9928,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.718,\"id\":\"worker_c\",\"p95_latency_ms\":82753,\"pass_rate\":0.8882,\"provider_group\":\"provider_x\",\"quality\":0.8299,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.757,\"estimated_latency_ms\":94450,\"estimated_pass_rate\":0.6842,\"estimated_quality\":0.9003,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.364,\"max_latency_ms\":179252,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5381, "category": "coding", "complexity": 0.8834, "consequence": 0.4907, "evidence_need": 0.8294, "family": "migration_planning", "input_tokens": 6960, "request_id": "synthetic-migration_planning-000757", "reversibility": 0.3377, "risk": 0.4405, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 0.859, "id": "worker_a", "p95_latency_ms": 26814, "pass_rate": 0.5404, "provider_group": "provider_z", "quality": 0.8043, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.898, "id": "worker_b", "p95_latency_ms": 67636, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9928, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.718, "id": "worker_c", "p95_latency_ms": 82753, "pass_rate": 0.8882, "provider_group": "provider_x", "quality": 0.8299, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.538, "max_latency_ms": 23776, "minimum_quality": 0.6139}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000222", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.111, "estimated_latency_ms": 19227, "estimated_pass_rate": 0.2667, "estimated_quality": 0.6062, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.538, "max_latency_ms": 23776, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2667, "quality": 0.6062, "quality_target_met": false, "utility": 20.201267}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.538,\"max_latency_ms\":23776,\"minimum_quality\":0.6139},\"task\":{\"ambiguity\":0.7801,\"category\":\"review\",\"complexity\":0.8006,\"consequence\":0.7965,\"evidence_need\":0.8286,\"family\":\"policy_compliance\",\"input_tokens\":200,\"reversibility\":0.488,\"risk\":0.8433,\"summary\":\"Review whether OAuth callbacks satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.111,\"id\":\"worker_a\",\"p95_latency_ms\":19227,\"pass_rate\":0.4848,\"provider_group\":\"provider_x\",\"quality\":0.7841,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.622,\"id\":\"worker_b\",\"p95_latency_ms\":54706,\"pass_rate\":0.8863,\"provider_group\":\"provider_z\",\"quality\":0.8096,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.193,\"id\":\"worker_c\",\"p95_latency_ms\":75729,\"pass_rate\":0.9778,\"provider_group\":\"provider_y\",\"quality\":0.9589,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.111,\"estimated_latency_ms\":19227,\"estimated_pass_rate\":0.2667,\"estimated_quality\":0.6062,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.538,\"max_latency_ms\":23776,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7801, "category": "review", "complexity": 0.8006, "consequence": 0.7965, "evidence_need": 0.8286, "family": "policy_compliance", "input_tokens": 200, "request_id": "synthetic-policy_compliance-000222", "reversibility": 0.488, "risk": 0.8433, "summary": "Review whether OAuth callbacks satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.111, "id": "worker_a", "p95_latency_ms": 19227, "pass_rate": 0.4848, "provider_group": "provider_x", "quality": 0.7841, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.622, "id": "worker_b", "p95_latency_ms": 54706, "pass_rate": 0.8863, "provider_group": "provider_z", "quality": 0.8096, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 12.193, "id": "worker_c", "p95_latency_ms": 75729, "pass_rate": 0.9778, "provider_group": "provider_y", "quality": 0.9589, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.021, "max_latency_ms": 115926, "minimum_quality": 0.7192}, "family": "migration_planning", "id": "synthetic-migration_planning-000388", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.637, "estimated_latency_ms": 90030, "estimated_pass_rate": 0.8657, "estimated_quality": 0.8865, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.021, "max_latency_ms": 115926, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8657, "quality": 0.8865, "quality_target_met": true, "utility": 62.558172}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.021,\"max_latency_ms\":115926,\"minimum_quality\":0.7192},\"task\":{\"ambiguity\":0.6455,\"category\":\"coding\",\"complexity\":0.6857,\"consequence\":0.4151,\"evidence_need\":0.6119,\"family\":\"migration_planning\",\"input_tokens\":600,\"reversibility\":0.262,\"risk\":0.4762,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":12.837,\"id\":\"worker_a\",\"p95_latency_ms\":62933,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9755,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.8,\"id\":\"worker_b\",\"p95_latency_ms\":27097,\"pass_rate\":0.6182,\"provider_group\":\"provider_y\",\"quality\":0.7484,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.536,\"id\":\"worker_c\",\"p95_latency_ms\":50184,\"pass_rate\":0.8008,\"provider_group\":\"provider_z\",\"quality\":0.8943,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.637,\"estimated_latency_ms\":90030,\"estimated_pass_rate\":0.8657,\"estimated_quality\":0.8865,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.021,\"max_latency_ms\":115926,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6455, "category": "coding", "complexity": 0.6857, "consequence": 0.4151, "evidence_need": 0.6119, "family": "migration_planning", "input_tokens": 600, "request_id": "synthetic-migration_planning-000388", "reversibility": 0.262, "risk": 0.4762, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 12.837, "id": "worker_a", "p95_latency_ms": 62933, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9755, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.8, "id": "worker_b", "p95_latency_ms": 27097, "pass_rate": 0.6182, "provider_group": "provider_y", "quality": 0.7484, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.536, "id": "worker_c", "p95_latency_ms": 50184, "pass_rate": 0.8008, "provider_group": "provider_z", "quality": 0.8943, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.962, "max_latency_ms": 53117, "minimum_quality": 0.5343}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000684", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.932, "estimated_latency_ms": 29906, "estimated_pass_rate": 0.4586, "estimated_quality": 0.6162, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.962, "max_latency_ms": 53117, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4586, "quality": 0.6162, "quality_target_met": true, "utility": 42.893426}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.962,\"max_latency_ms\":53117,\"minimum_quality\":0.5343},\"task\":{\"ambiguity\":0.3547,\"category\":\"review\",\"complexity\":0.4062,\"consequence\":0.5492,\"evidence_need\":0.631,\"family\":\"policy_compliance\",\"input_tokens\":3960,\"reversibility\":0.3685,\"risk\":0.7155,\"summary\":\"Audit a workflow for credit reservation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.932,\"id\":\"worker_a\",\"p95_latency_ms\":29906,\"pass_rate\":0.5894,\"provider_group\":\"provider_y\",\"quality\":0.685,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.082,\"id\":\"worker_b\",\"p95_latency_ms\":77008,\"pass_rate\":0.9949,\"provider_group\":\"provider_z\",\"quality\":0.9164,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.674,\"id\":\"worker_c\",\"p95_latency_ms\":73966,\"pass_rate\":0.7819,\"provider_group\":\"provider_x\",\"quality\":0.7997,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.932,\"estimated_latency_ms\":29906,\"estimated_pass_rate\":0.4586,\"estimated_quality\":0.6162,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.962,\"max_latency_ms\":53117,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3547, "category": "review", "complexity": 0.4062, "consequence": 0.5492, "evidence_need": 0.631, "family": "policy_compliance", "input_tokens": 3960, "request_id": "synthetic-policy_compliance-000684", "reversibility": 0.3685, "risk": 0.7155, "summary": "Audit a workflow for credit reservation against explicit access requirements."}, "workers": [{"credits_per_call": 0.932, "id": "worker_a", "p95_latency_ms": 29906, "pass_rate": 0.5894, "provider_group": "provider_y", "quality": 0.685, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.082, "id": "worker_b", "p95_latency_ms": 77008, "pass_rate": 0.9949, "provider_group": "provider_z", "quality": 0.9164, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.674, "id": "worker_c", "p95_latency_ms": 73966, "pass_rate": 0.7819, "provider_group": "provider_x", "quality": 0.7997, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.653, "max_latency_ms": 208661, "minimum_quality": 0.845}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000267", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.644, "estimated_latency_ms": 172858, "estimated_pass_rate": 0.7523, "estimated_quality": 0.9144, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 24.653, "max_latency_ms": 208661, "policy": "strict", "reason_codes": ["high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7523, "quality": 0.9144, "quality_target_met": true, "utility": 72.853012}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.653,\"max_latency_ms\":208661,\"minimum_quality\":0.845},\"task\":{\"ambiguity\":0.4485,\"category\":\"review\",\"complexity\":0.5571,\"consequence\":0.564,\"evidence_need\":0.6042,\"family\":\"policy_compliance\",\"input_tokens\":5320,\"reversibility\":0.6472,\"risk\":0.7693,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.197,\"id\":\"worker_a\",\"p95_latency_ms\":51989,\"pass_rate\":0.8135,\"provider_group\":\"provider_y\",\"quality\":0.8541,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.882,\"id\":\"worker_b\",\"p95_latency_ms\":29621,\"pass_rate\":0.5708,\"provider_group\":\"provider_x\",\"quality\":0.7423,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.565,\"id\":\"worker_c\",\"p95_latency_ms\":91248,\"pass_rate\":0.8949,\"provider_group\":\"provider_z\",\"quality\":0.99,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.644,\"estimated_latency_ms\":172858,\"estimated_pass_rate\":0.7523,\"estimated_quality\":0.9144,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.653,\"max_latency_ms\":208661,\"policy\":\"strict\",\"reason_codes\":[\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4485, "category": "review", "complexity": 0.5571, "consequence": 0.564, "evidence_need": 0.6042, "family": "policy_compliance", "input_tokens": 5320, "request_id": "synthetic-policy_compliance-000267", "reversibility": 0.6472, "risk": 0.7693, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 4.197, "id": "worker_a", "p95_latency_ms": 51989, "pass_rate": 0.8135, "provider_group": "provider_y", "quality": 0.8541, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.882, "id": "worker_b", "p95_latency_ms": 29621, "pass_rate": 0.5708, "provider_group": "provider_x", "quality": 0.7423, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.565, "id": "worker_c", "p95_latency_ms": 91248, "pass_rate": 0.8949, "provider_group": "provider_z", "quality": 0.99, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.36, "max_latency_ms": 183740, "minimum_quality": 0.7848}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000825", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.265, "estimated_latency_ms": 142842, "estimated_pass_rate": 0.7666, "estimated_quality": 0.9322, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.36, "max_latency_ms": 183740, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7666, "quality": 0.9322, "quality_target_met": true, "utility": 75.606997}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.36,\"max_latency_ms\":183740,\"minimum_quality\":0.7848},\"task\":{\"ambiguity\":0.7253,\"category\":\"review\",\"complexity\":0.4379,\"consequence\":0.5734,\"evidence_need\":0.9257,\"family\":\"policy_compliance\",\"input_tokens\":4000,\"reversibility\":0.5327,\"risk\":0.6697,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.175,\"id\":\"worker_a\",\"p95_latency_ms\":26993,\"pass_rate\":0.5385,\"provider_group\":\"provider_y\",\"quality\":0.7454,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.318,\"id\":\"worker_b\",\"p95_latency_ms\":62307,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.772,\"id\":\"worker_c\",\"p95_latency_ms\":53542,\"pass_rate\":0.7477,\"provider_group\":\"provider_x\",\"quality\":0.908,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.265,\"estimated_latency_ms\":142842,\"estimated_pass_rate\":0.7666,\"estimated_quality\":0.9322,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.36,\"max_latency_ms\":183740,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7253, "category": "review", "complexity": 0.4379, "consequence": 0.5734, "evidence_need": 0.9257, "family": "policy_compliance", "input_tokens": 4000, "request_id": "synthetic-policy_compliance-000825", "reversibility": 0.5327, "risk": 0.6697, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 1.175, "id": "worker_a", "p95_latency_ms": 26993, "pass_rate": 0.5385, "provider_group": "provider_y", "quality": 0.7454, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.318, "id": "worker_b", "p95_latency_ms": 62307, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.772, "id": "worker_c", "p95_latency_ms": 53542, "pass_rate": 0.7477, "provider_group": "provider_x", "quality": 0.908, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.87, "max_latency_ms": 214660, "minimum_quality": 0.7688}, "family": "migration_planning", "id": "synthetic-migration_planning-000064", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.534, "estimated_latency_ms": 129541, "estimated_pass_rate": 0.8801, "estimated_quality": 0.8653, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.87, "max_latency_ms": 214660, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8801, "quality": 0.8653, "quality_target_met": true, "utility": 66.417598}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.87,\"max_latency_ms\":214660,\"minimum_quality\":0.7688},\"task\":{\"ambiguity\":0.5133,\"category\":\"coding\",\"complexity\":0.9499,\"consequence\":0.5896,\"evidence_need\":0.5263,\"family\":\"migration_planning\",\"input_tokens\":4000,\"reversibility\":0.5877,\"risk\":0.3833,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":4.061,\"id\":\"worker_a\",\"p95_latency_ms\":52486,\"pass_rate\":0.8465,\"provider_group\":\"provider_y\",\"quality\":0.843,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.192,\"id\":\"worker_b\",\"p95_latency_ms\":25162,\"pass_rate\":0.5532,\"provider_group\":\"provider_x\",\"quality\":0.7741,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.473,\"id\":\"worker_c\",\"p95_latency_ms\":77055,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9247,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.534,\"estimated_latency_ms\":129541,\"estimated_pass_rate\":0.8801,\"estimated_quality\":0.8653,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.87,\"max_latency_ms\":214660,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5133, "category": "coding", "complexity": 0.9499, "consequence": 0.5896, "evidence_need": 0.5263, "family": "migration_planning", "input_tokens": 4000, "request_id": "synthetic-migration_planning-000064", "reversibility": 0.5877, "risk": 0.3833, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 4.061, "id": "worker_a", "p95_latency_ms": 52486, "pass_rate": 0.8465, "provider_group": "provider_y", "quality": 0.843, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.192, "id": "worker_b", "p95_latency_ms": 25162, "pass_rate": 0.5532, "provider_group": "provider_x", "quality": 0.7741, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.473, "id": "worker_c", "p95_latency_ms": 77055, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9247, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.699, "max_latency_ms": 241049, "minimum_quality": 0.8952}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000486", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.65, "estimated_latency_ms": 194889, "estimated_pass_rate": 0.7731, "estimated_quality": 0.9195, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.699, "max_latency_ms": 241049, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7731, "quality": 0.9195, "quality_target_met": true, "utility": 74.419708}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.699,\"max_latency_ms\":241049,\"minimum_quality\":0.8952},\"task\":{\"ambiguity\":0.7567,\"category\":\"review\",\"complexity\":0.6328,\"consequence\":0.6111,\"evidence_need\":0.7842,\"family\":\"policy_compliance\",\"input_tokens\":640,\"reversibility\":0.6935,\"risk\":0.4664,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":0.831,\"id\":\"worker_a\",\"p95_latency_ms\":27089,\"pass_rate\":0.5959,\"provider_group\":\"provider_z\",\"quality\":0.6995,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.168,\"id\":\"worker_b\",\"p95_latency_ms\":82049,\"pass_rate\":0.8598,\"provider_group\":\"provider_x\",\"quality\":0.82,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.651,\"id\":\"worker_c\",\"p95_latency_ms\":85751,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.65,\"estimated_latency_ms\":194889,\"estimated_pass_rate\":0.7731,\"estimated_quality\":0.9195,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.699,\"max_latency_ms\":241049,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7567, "category": "review", "complexity": 0.6328, "consequence": 0.6111, "evidence_need": 0.7842, "family": "policy_compliance", "input_tokens": 640, "request_id": "synthetic-policy_compliance-000486", "reversibility": 0.6935, "risk": 0.4664, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 0.831, "id": "worker_a", "p95_latency_ms": 27089, "pass_rate": 0.5959, "provider_group": "provider_z", "quality": 0.6995, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.168, "id": "worker_b", "p95_latency_ms": 82049, "pass_rate": 0.8598, "provider_group": "provider_x", "quality": 0.82, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.651, "id": "worker_c", "p95_latency_ms": 85751, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.614, "max_latency_ms": 21505, "minimum_quality": 0.6607}, "family": "migration_planning", "id": "synthetic-migration_planning-000079", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.057, "estimated_latency_ms": 19172, "estimated_pass_rate": 0.3209, "estimated_quality": 0.5989, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.614, "max_latency_ms": 21505, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3209, "quality": 0.5989, "quality_target_met": false, "utility": 35.857773}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.614,\"max_latency_ms\":21505,\"minimum_quality\":0.6607},\"task\":{\"ambiguity\":0.4797,\"category\":\"coding\",\"complexity\":0.6023,\"consequence\":0.4708,\"evidence_need\":0.7308,\"family\":\"migration_planning\",\"input_tokens\":7040,\"reversibility\":0.6477,\"risk\":0.53,\"summary\":\"Review a staged migration for streaming inference and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.088,\"id\":\"worker_a\",\"p95_latency_ms\":89592,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9301,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.057,\"id\":\"worker_b\",\"p95_latency_ms\":19172,\"pass_rate\":0.4684,\"provider_group\":\"provider_x\",\"quality\":0.6982,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.362,\"id\":\"worker_c\",\"p95_latency_ms\":55726,\"pass_rate\":0.7976,\"provider_group\":\"provider_z\",\"quality\":0.8625,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.057,\"estimated_latency_ms\":19172,\"estimated_pass_rate\":0.3209,\"estimated_quality\":0.5989,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.614,\"max_latency_ms\":21505,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4797, "category": "coding", "complexity": 0.6023, "consequence": 0.4708, "evidence_need": 0.7308, "family": "migration_planning", "input_tokens": 7040, "request_id": "synthetic-migration_planning-000079", "reversibility": 0.6477, "risk": 0.53, "summary": "Review a staged migration for streaming inference and identify irreversible steps."}, "workers": [{"credits_per_call": 11.088, "id": "worker_a", "p95_latency_ms": 89592, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9301, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.057, "id": "worker_b", "p95_latency_ms": 19172, "pass_rate": 0.4684, "provider_group": "provider_x", "quality": 0.6982, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.362, "id": "worker_c", "p95_latency_ms": 55726, "pass_rate": 0.7976, "provider_group": "provider_z", "quality": 0.8625, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.015, "max_latency_ms": 251417, "minimum_quality": 0.9591}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001266", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.212, "estimated_latency_ms": 195186, "estimated_pass_rate": 0.7512, "estimated_quality": 0.9234, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 20.015, "max_latency_ms": 251417, "policy": "strict", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7512, "quality": 0.9234, "quality_target_met": false, "utility": 78.457972}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.015,\"max_latency_ms\":251417,\"minimum_quality\":0.9591},\"task\":{\"ambiguity\":0.6959,\"category\":\"review\",\"complexity\":0.7145,\"consequence\":0.6998,\"evidence_need\":0.7978,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.6749,\"risk\":0.7855,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.22,\"id\":\"worker_a\",\"p95_latency_ms\":30168,\"pass_rate\":0.5628,\"provider_group\":\"provider_x\",\"quality\":0.7394,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.878,\"id\":\"worker_b\",\"p95_latency_ms\":92017,\"pass_rate\":0.8882,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.114,\"id\":\"worker_c\",\"p95_latency_ms\":73001,\"pass_rate\":0.8751,\"provider_group\":\"provider_z\",\"quality\":0.8678,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.212,\"estimated_latency_ms\":195186,\"estimated_pass_rate\":0.7512,\"estimated_quality\":0.9234,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.015,\"max_latency_ms\":251417,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6959, "category": "review", "complexity": 0.7145, "consequence": 0.6998, "evidence_need": 0.7978, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-001266", "reversibility": 0.6749, "risk": 0.7855, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.22, "id": "worker_a", "p95_latency_ms": 30168, "pass_rate": 0.5628, "provider_group": "provider_x", "quality": 0.7394, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.878, "id": "worker_b", "p95_latency_ms": 92017, "pass_rate": 0.8882, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.114, "id": "worker_c", "p95_latency_ms": 73001, "pass_rate": 0.8751, "provider_group": "provider_z", "quality": 0.8678, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.517, "max_latency_ms": 202194, "minimum_quality": 0.8578}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001065", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.329, "estimated_latency_ms": 173065, "estimated_pass_rate": 0.7041, "estimated_quality": 0.8971, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.517, "max_latency_ms": 202194, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7041, "quality": 0.8971, "quality_target_met": true, "utility": 71.380283}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.517,\"max_latency_ms\":202194,\"minimum_quality\":0.8578},\"task\":{\"ambiguity\":0.511,\"category\":\"review\",\"complexity\":0.6703,\"consequence\":0.7034,\"evidence_need\":0.8626,\"family\":\"policy_compliance\",\"input_tokens\":7000,\"reversibility\":0.5957,\"risk\":0.4539,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.15,\"id\":\"worker_a\",\"p95_latency_ms\":79583,\"pass_rate\":0.8856,\"provider_group\":\"provider_y\",\"quality\":0.9181,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.184,\"id\":\"worker_b\",\"p95_latency_ms\":60556,\"pass_rate\":0.85,\"provider_group\":\"provider_x\",\"quality\":0.8164,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.995,\"id\":\"worker_c\",\"p95_latency_ms\":32926,\"pass_rate\":0.5193,\"provider_group\":\"provider_z\",\"quality\":0.7896,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.329,\"estimated_latency_ms\":173065,\"estimated_pass_rate\":0.7041,\"estimated_quality\":0.8971,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.517,\"max_latency_ms\":202194,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.511, "category": "review", "complexity": 0.6703, "consequence": 0.7034, "evidence_need": 0.8626, "family": "policy_compliance", "input_tokens": 7000, "request_id": "synthetic-policy_compliance-001065", "reversibility": 0.5957, "risk": 0.4539, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.15, "id": "worker_a", "p95_latency_ms": 79583, "pass_rate": 0.8856, "provider_group": "provider_y", "quality": 0.9181, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.184, "id": "worker_b", "p95_latency_ms": 60556, "pass_rate": 0.85, "provider_group": "provider_x", "quality": 0.8164, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.995, "id": "worker_c", "p95_latency_ms": 32926, "pass_rate": 0.5193, "provider_group": "provider_z", "quality": 0.7896, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.07, "max_latency_ms": 34632, "minimum_quality": 0.7336}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000603", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.991, "estimated_latency_ms": 20847, "estimated_pass_rate": 0.3899, "estimated_quality": 0.6185, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.07, "max_latency_ms": 34632, "policy": "lean", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3899, "quality": 0.6185, "quality_target_met": false, "utility": 17.728551}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.07,\"max_latency_ms\":34632,\"minimum_quality\":0.7336},\"task\":{\"ambiguity\":0.4277,\"category\":\"review\",\"complexity\":0.4404,\"consequence\":0.8829,\"evidence_need\":0.8759,\"family\":\"policy_compliance\",\"input_tokens\":6200,\"reversibility\":0.3971,\"risk\":0.8476,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.272,\"id\":\"worker_a\",\"p95_latency_ms\":65741,\"pass_rate\":0.7251,\"provider_group\":\"provider_y\",\"quality\":0.826,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.991,\"id\":\"worker_b\",\"p95_latency_ms\":20847,\"pass_rate\":0.5408,\"provider_group\":\"provider_z\",\"quality\":0.7079,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.025,\"id\":\"worker_c\",\"p95_latency_ms\":69175,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.991,\"estimated_latency_ms\":20847,\"estimated_pass_rate\":0.3899,\"estimated_quality\":0.6185,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.07,\"max_latency_ms\":34632,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4277, "category": "review", "complexity": 0.4404, "consequence": 0.8829, "evidence_need": 0.8759, "family": "policy_compliance", "input_tokens": 6200, "request_id": "synthetic-policy_compliance-000603", "reversibility": 0.3971, "risk": 0.8476, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.272, "id": "worker_a", "p95_latency_ms": 65741, "pass_rate": 0.7251, "provider_group": "provider_y", "quality": 0.826, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.991, "id": "worker_b", "p95_latency_ms": 20847, "pass_rate": 0.5408, "provider_group": "provider_z", "quality": 0.7079, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.025, "id": "worker_c", "p95_latency_ms": 69175, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 5.567, "max_latency_ms": 145618, "minimum_quality": 0.7022}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000813", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.744, "estimated_latency_ms": 88356, "estimated_pass_rate": 0.6333, "estimated_quality": 0.787, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.567, "max_latency_ms": 145618, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6333, "quality": 0.787, "quality_target_met": true, "utility": 59.203105}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":5.567,\"max_latency_ms\":145618,\"minimum_quality\":0.7022},\"task\":{\"ambiguity\":0.5561,\"category\":\"review\",\"complexity\":0.5152,\"consequence\":0.6721,\"evidence_need\":0.7406,\"family\":\"policy_compliance\",\"input_tokens\":6280,\"reversibility\":0.4387,\"risk\":0.8477,\"summary\":\"Review whether data retention satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.721,\"id\":\"worker_a\",\"p95_latency_ms\":67928,\"pass_rate\":0.8049,\"provider_group\":\"provider_x\",\"quality\":0.8236,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.543,\"id\":\"worker_b\",\"p95_latency_ms\":49418,\"pass_rate\":0.9146,\"provider_group\":\"provider_z\",\"quality\":0.9537,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.023,\"id\":\"worker_c\",\"p95_latency_ms\":20428,\"pass_rate\":0.5274,\"provider_group\":\"provider_y\",\"quality\":0.6835,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.744,\"estimated_latency_ms\":88356,\"estimated_pass_rate\":0.6333,\"estimated_quality\":0.787,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.567,\"max_latency_ms\":145618,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5561, "category": "review", "complexity": 0.5152, "consequence": 0.6721, "evidence_need": 0.7406, "family": "policy_compliance", "input_tokens": 6280, "request_id": "synthetic-policy_compliance-000813", "reversibility": 0.4387, "risk": 0.8477, "summary": "Review whether data retention satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.721, "id": "worker_a", "p95_latency_ms": 67928, "pass_rate": 0.8049, "provider_group": "provider_x", "quality": 0.8236, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.543, "id": "worker_b", "p95_latency_ms": 49418, "pass_rate": 0.9146, "provider_group": "provider_z", "quality": 0.9537, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.023, "id": "worker_c", "p95_latency_ms": 20428, "pass_rate": 0.5274, "provider_group": "provider_y", "quality": 0.6835, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.21, "max_latency_ms": 37220, "minimum_quality": 0.7348}, "family": "migration_planning", "id": "synthetic-migration_planning-000667", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.816, "estimated_latency_ms": 28253, "estimated_pass_rate": 0.4462, "estimated_quality": 0.6087, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.21, "max_latency_ms": 37220, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4462, "quality": 0.6087, "quality_target_met": false, "utility": 27.069274}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.21,\"max_latency_ms\":37220,\"minimum_quality\":0.7348},\"task\":{\"ambiguity\":0.5273,\"category\":\"coding\",\"complexity\":0.6484,\"consequence\":0.5891,\"evidence_need\":0.5648,\"family\":\"migration_planning\",\"input_tokens\":2920,\"reversibility\":0.2625,\"risk\":0.4985,\"summary\":\"Review a staged migration for rate limiting and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.385,\"id\":\"worker_a\",\"p95_latency_ms\":71427,\"pass_rate\":0.7225,\"provider_group\":\"provider_x\",\"quality\":0.8457,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.114,\"id\":\"worker_b\",\"p95_latency_ms\":74210,\"pass_rate\":0.9075,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.816,\"id\":\"worker_c\",\"p95_latency_ms\":28253,\"pass_rate\":0.6,\"provider_group\":\"provider_y\",\"quality\":0.7173,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.816,\"estimated_latency_ms\":28253,\"estimated_pass_rate\":0.4462,\"estimated_quality\":0.6087,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.21,\"max_latency_ms\":37220,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5273, "category": "coding", "complexity": 0.6484, "consequence": 0.5891, "evidence_need": 0.5648, "family": "migration_planning", "input_tokens": 2920, "request_id": "synthetic-migration_planning-000667", "reversibility": 0.2625, "risk": 0.4985, "summary": "Review a staged migration for rate limiting and identify irreversible steps."}, "workers": [{"credits_per_call": 3.385, "id": "worker_a", "p95_latency_ms": 71427, "pass_rate": 0.7225, "provider_group": "provider_x", "quality": 0.8457, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.114, "id": "worker_b", "p95_latency_ms": 74210, "pass_rate": 0.9075, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.816, "id": "worker_c", "p95_latency_ms": 28253, "pass_rate": 0.6, "provider_group": "provider_y", "quality": 0.7173, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.975, "max_latency_ms": 147206, "minimum_quality": 0.7094}, "family": "migration_planning", "id": "synthetic-migration_planning-000082", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.479, "estimated_latency_ms": 105926, "estimated_pass_rate": 0.7215, "estimated_quality": 0.8612, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.975, "max_latency_ms": 147206, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7215, "quality": 0.8612, "quality_target_met": true, "utility": 65.613096}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.975,\"max_latency_ms\":147206,\"minimum_quality\":0.7094},\"task\":{\"ambiguity\":0.6782,\"category\":\"coding\",\"complexity\":0.8894,\"consequence\":0.8486,\"evidence_need\":0.5559,\"family\":\"migration_planning\",\"input_tokens\":6400,\"reversibility\":0.2969,\"risk\":0.6496,\"summary\":\"Design compatibility testing for a artifact storage schema migration.\"},\"workers\":[{\"credits_per_call\":4.003,\"id\":\"worker_a\",\"p95_latency_ms\":71794,\"pass_rate\":0.8168,\"provider_group\":\"provider_y\",\"quality\":0.8775,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.19,\"id\":\"worker_b\",\"p95_latency_ms\":18514,\"pass_rate\":0.5533,\"provider_group\":\"provider_z\",\"quality\":0.7801,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.289,\"id\":\"worker_c\",\"p95_latency_ms\":87412,\"pass_rate\":0.9906,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.479,\"estimated_latency_ms\":105926,\"estimated_pass_rate\":0.7215,\"estimated_quality\":0.8612,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.975,\"max_latency_ms\":147206,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6782, "category": "coding", "complexity": 0.8894, "consequence": 0.8486, "evidence_need": 0.5559, "family": "migration_planning", "input_tokens": 6400, "request_id": "synthetic-migration_planning-000082", "reversibility": 0.2969, "risk": 0.6496, "summary": "Design compatibility testing for a artifact storage schema migration."}, "workers": [{"credits_per_call": 4.003, "id": "worker_a", "p95_latency_ms": 71794, "pass_rate": 0.8168, "provider_group": "provider_y", "quality": 0.8775, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.19, "id": "worker_b", "p95_latency_ms": 18514, "pass_rate": 0.5533, "provider_group": "provider_z", "quality": 0.7801, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.289, "id": "worker_c", "p95_latency_ms": 87412, "pass_rate": 0.9906, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.793, "max_latency_ms": 103377, "minimum_quality": 0.7447}, "family": "migration_planning", "id": "synthetic-migration_planning-001159", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.044, "estimated_latency_ms": 87368, "estimated_pass_rate": 0.5927, "estimated_quality": 0.8134, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.793, "max_latency_ms": 103377, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5927, "quality": 0.8134, "quality_target_met": true, "utility": 58.450352}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.793,\"max_latency_ms\":103377,\"minimum_quality\":0.7447},\"task\":{\"ambiguity\":0.4626,\"category\":\"coding\",\"complexity\":0.9441,\"consequence\":0.6082,\"evidence_need\":0.786,\"family\":\"migration_planning\",\"input_tokens\":6440,\"reversibility\":0.4275,\"risk\":0.5619,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.787,\"id\":\"worker_a\",\"p95_latency_ms\":57976,\"pass_rate\":0.8874,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.107,\"id\":\"worker_b\",\"p95_latency_ms\":58116,\"pass_rate\":0.7347,\"provider_group\":\"provider_z\",\"quality\":0.9121,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.937,\"id\":\"worker_c\",\"p95_latency_ms\":29252,\"pass_rate\":0.5426,\"provider_group\":\"provider_x\",\"quality\":0.8021,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.044,\"estimated_latency_ms\":87368,\"estimated_pass_rate\":0.5927,\"estimated_quality\":0.8134,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.793,\"max_latency_ms\":103377,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4626, "category": "coding", "complexity": 0.9441, "consequence": 0.6082, "evidence_need": 0.786, "family": "migration_planning", "input_tokens": 6440, "request_id": "synthetic-migration_planning-001159", "reversibility": 0.4275, "risk": 0.5619, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 11.787, "id": "worker_a", "p95_latency_ms": 57976, "pass_rate": 0.8874, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.107, "id": "worker_b", "p95_latency_ms": 58116, "pass_rate": 0.7347, "provider_group": "provider_z", "quality": 0.9121, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.937, "id": "worker_c", "p95_latency_ms": 29252, "pass_rate": 0.5426, "provider_group": "provider_x", "quality": 0.8021, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.082, "max_latency_ms": 78295, "minimum_quality": 0.7821}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001136", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.091, "estimated_latency_ms": 70805, "estimated_pass_rate": 0.7977, "estimated_quality": 0.7608, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.082, "max_latency_ms": 78295, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7977, "quality": 0.7608, "quality_target_met": false, "utility": 57.457171}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.082,\"max_latency_ms\":78295,\"minimum_quality\":0.7821},\"task\":{\"ambiguity\":0.548,\"category\":\"analysis\",\"complexity\":0.7634,\"consequence\":0.5616,\"evidence_need\":0.5894,\"family\":\"performance_tuning\",\"input_tokens\":4920,\"reversibility\":0.5682,\"risk\":0.5158,\"summary\":\"Diagnose latency variance in provider failover and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":13.338,\"id\":\"worker_a\",\"p95_latency_ms\":92544,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.221,\"id\":\"worker_b\",\"p95_latency_ms\":52702,\"pass_rate\":0.9419,\"provider_group\":\"provider_y\",\"quality\":0.8603,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.87,\"id\":\"worker_c\",\"p95_latency_ms\":18103,\"pass_rate\":0.4174,\"provider_group\":\"provider_x\",\"quality\":0.6781,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.091,\"estimated_latency_ms\":70805,\"estimated_pass_rate\":0.7977,\"estimated_quality\":0.7608,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.082,\"max_latency_ms\":78295,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.548, "category": "analysis", "complexity": 0.7634, "consequence": 0.5616, "evidence_need": 0.5894, "family": "performance_tuning", "input_tokens": 4920, "request_id": "synthetic-performance_tuning-001136", "reversibility": 0.5682, "risk": 0.5158, "summary": "Diagnose latency variance in provider failover and rank the next experiments."}, "workers": [{"credits_per_call": 13.338, "id": "worker_a", "p95_latency_ms": 92544, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.221, "id": "worker_b", "p95_latency_ms": 52702, "pass_rate": 0.9419, "provider_group": "provider_y", "quality": 0.8603, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.87, "id": "worker_c", "p95_latency_ms": 18103, "pass_rate": 0.4174, "provider_group": "provider_x", "quality": 0.6781, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.834, "max_latency_ms": 188444, "minimum_quality": 0.8403}, "family": "migration_planning", "id": "synthetic-migration_planning-001210", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.251, "estimated_latency_ms": 139047, "estimated_pass_rate": 0.7836, "estimated_quality": 0.8701, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 26.834, "max_latency_ms": 188444, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7836, "quality": 0.8701, "quality_target_met": true, "utility": 71.880968}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.834,\"max_latency_ms\":188444,\"minimum_quality\":0.8403},\"task\":{\"ambiguity\":0.659,\"category\":\"coding\",\"complexity\":0.6513,\"consequence\":0.5616,\"evidence_need\":0.4468,\"family\":\"migration_planning\",\"input_tokens\":5520,\"reversibility\":0.4315,\"risk\":0.6551,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":1.199,\"id\":\"worker_a\",\"p95_latency_ms\":19259,\"pass_rate\":0.5822,\"provider_group\":\"provider_z\",\"quality\":0.6752,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.164,\"id\":\"worker_b\",\"p95_latency_ms\":56420,\"pass_rate\":0.836,\"provider_group\":\"provider_x\",\"quality\":0.8606,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.888,\"id\":\"worker_c\",\"p95_latency_ms\":63368,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9277,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.251,\"estimated_latency_ms\":139047,\"estimated_pass_rate\":0.7836,\"estimated_quality\":0.8701,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.834,\"max_latency_ms\":188444,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.659, "category": "coding", "complexity": 0.6513, "consequence": 0.5616, "evidence_need": 0.4468, "family": "migration_planning", "input_tokens": 5520, "request_id": "synthetic-migration_planning-001210", "reversibility": 0.4315, "risk": 0.6551, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 1.199, "id": "worker_a", "p95_latency_ms": 19259, "pass_rate": 0.5822, "provider_group": "provider_z", "quality": 0.6752, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.164, "id": "worker_b", "p95_latency_ms": 56420, "pass_rate": 0.836, "provider_group": "provider_x", "quality": 0.8606, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 14.888, "id": "worker_c", "p95_latency_ms": 63368, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9277, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.351, "max_latency_ms": 177037, "minimum_quality": 0.8212}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001062", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.411, "estimated_latency_ms": 104641, "estimated_pass_rate": 0.6296, "estimated_quality": 0.8181, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.351, "max_latency_ms": 177037, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6296, "quality": 0.8181, "quality_target_met": false, "utility": 63.487306}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.351,\"max_latency_ms\":177037,\"minimum_quality\":0.8212},\"task\":{\"ambiguity\":0.3745,\"category\":\"review\",\"complexity\":0.5583,\"consequence\":0.6595,\"evidence_need\":0.7537,\"family\":\"policy_compliance\",\"input_tokens\":1880,\"reversibility\":0.5654,\"risk\":0.7211,\"summary\":\"Review whether observability satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.234,\"id\":\"worker_a\",\"p95_latency_ms\":77101,\"pass_rate\":0.7732,\"provider_group\":\"provider_x\",\"quality\":0.8143,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.537,\"id\":\"worker_b\",\"p95_latency_ms\":85635,\"pass_rate\":0.961,\"provider_group\":\"provider_z\",\"quality\":0.9384,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.177,\"id\":\"worker_c\",\"p95_latency_ms\":27540,\"pass_rate\":0.5532,\"provider_group\":\"provider_y\",\"quality\":0.7608,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.411,\"estimated_latency_ms\":104641,\"estimated_pass_rate\":0.6296,\"estimated_quality\":0.8181,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.351,\"max_latency_ms\":177037,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3745, "category": "review", "complexity": 0.5583, "consequence": 0.6595, "evidence_need": 0.7537, "family": "policy_compliance", "input_tokens": 1880, "request_id": "synthetic-policy_compliance-001062", "reversibility": 0.5654, "risk": 0.7211, "summary": "Review whether observability satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.234, "id": "worker_a", "p95_latency_ms": 77101, "pass_rate": 0.7732, "provider_group": "provider_x", "quality": 0.8143, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.537, "id": "worker_b", "p95_latency_ms": 85635, "pass_rate": 0.961, "provider_group": "provider_z", "quality": 0.9384, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.177, "id": "worker_c", "p95_latency_ms": 27540, "pass_rate": 0.5532, "provider_group": "provider_y", "quality": 0.7608, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.444, "max_latency_ms": 26082, "minimum_quality": 0.4986}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000155", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.123, "estimated_latency_ms": 22396, "estimated_pass_rate": 0.2734, "estimated_quality": 0.5109, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.444, "max_latency_ms": 26082, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2734, "quality": 0.5109, "quality_target_met": true, "utility": 30.883507}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.444,\"max_latency_ms\":26082,\"minimum_quality\":0.4986},\"task\":{\"ambiguity\":0.5346,\"category\":\"analysis\",\"complexity\":0.737,\"consequence\":0.5371,\"evidence_need\":0.6911,\"family\":\"performance_tuning\",\"input_tokens\":7680,\"reversibility\":0.708,\"risk\":0.3163,\"summary\":\"Review a proposed optimization for provider failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.123,\"id\":\"worker_a\",\"p95_latency_ms\":22396,\"pass_rate\":0.423,\"provider_group\":\"provider_z\",\"quality\":0.6228,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.07,\"id\":\"worker_b\",\"p95_latency_ms\":80594,\"pass_rate\":0.8347,\"provider_group\":\"provider_x\",\"quality\":0.9219,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.14,\"id\":\"worker_c\",\"p95_latency_ms\":51696,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.123,\"estimated_latency_ms\":22396,\"estimated_pass_rate\":0.2734,\"estimated_quality\":0.5109,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.444,\"max_latency_ms\":26082,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5346, "category": "analysis", "complexity": 0.737, "consequence": 0.5371, "evidence_need": 0.6911, "family": "performance_tuning", "input_tokens": 7680, "request_id": "synthetic-performance_tuning-000155", "reversibility": 0.708, "risk": 0.3163, "summary": "Review a proposed optimization for provider failover for hidden regressions."}, "workers": [{"credits_per_call": 1.123, "id": "worker_a", "p95_latency_ms": 22396, "pass_rate": 0.423, "provider_group": "provider_z", "quality": 0.6228, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.07, "id": "worker_b", "p95_latency_ms": 80594, "pass_rate": 0.8347, "provider_group": "provider_x", "quality": 0.9219, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.14, "id": "worker_c", "p95_latency_ms": 51696, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.514, "max_latency_ms": 143454, "minimum_quality": 0.8624}, "family": "migration_planning", "id": "synthetic-migration_planning-000679", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.941, "estimated_latency_ms": 88804, "estimated_pass_rate": 0.5852, "estimated_quality": 0.8441, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.514, "max_latency_ms": 143454, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5852, "quality": 0.8441, "quality_target_met": false, "utility": 60.996718}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.514,\"max_latency_ms\":143454,\"minimum_quality\":0.8624},\"task\":{\"ambiguity\":0.4654,\"category\":\"coding\",\"complexity\":0.6077,\"consequence\":0.7743,\"evidence_need\":0.8026,\"family\":\"migration_planning\",\"input_tokens\":7040,\"reversibility\":0.3152,\"risk\":0.4131,\"summary\":\"Plan a zero-downtime migration of schema evolution with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.468,\"id\":\"worker_a\",\"p95_latency_ms\":75383,\"pass_rate\":0.981,\"provider_group\":\"provider_y\",\"quality\":0.9824,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.184,\"id\":\"worker_b\",\"p95_latency_ms\":25487,\"pass_rate\":0.4683,\"provider_group\":\"provider_x\",\"quality\":0.7586,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.757,\"id\":\"worker_c\",\"p95_latency_ms\":63317,\"pass_rate\":0.847,\"provider_group\":\"provider_z\",\"quality\":0.8982,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.941,\"estimated_latency_ms\":88804,\"estimated_pass_rate\":0.5852,\"estimated_quality\":0.8441,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.514,\"max_latency_ms\":143454,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4654, "category": "coding", "complexity": 0.6077, "consequence": 0.7743, "evidence_need": 0.8026, "family": "migration_planning", "input_tokens": 7040, "request_id": "synthetic-migration_planning-000679", "reversibility": 0.3152, "risk": 0.4131, "summary": "Plan a zero-downtime migration of schema evolution with rollback checkpoints."}, "workers": [{"credits_per_call": 14.468, "id": "worker_a", "p95_latency_ms": 75383, "pass_rate": 0.981, "provider_group": "provider_y", "quality": 0.9824, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.184, "id": "worker_b", "p95_latency_ms": 25487, "pass_rate": 0.4683, "provider_group": "provider_x", "quality": 0.7586, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.757, "id": "worker_c", "p95_latency_ms": 63317, "pass_rate": 0.847, "provider_group": "provider_z", "quality": 0.8982, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.739, "max_latency_ms": 95782, "minimum_quality": 0.7435}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000450", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.766, "estimated_latency_ms": 88281, "estimated_pass_rate": 0.8663, "estimated_quality": 0.8846, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.739, "max_latency_ms": 95782, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 5, "pass_rate": 0.8663, "quality": 0.8846, "quality_target_met": true, "utility": 66.040692}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.739,\"max_latency_ms\":95782,\"minimum_quality\":0.7435},\"task\":{\"ambiguity\":0.4158,\"category\":\"review\",\"complexity\":0.6906,\"consequence\":0.9242,\"evidence_need\":0.5635,\"family\":\"policy_compliance\",\"input_tokens\":5920,\"reversibility\":0.6333,\"risk\":0.6203,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.378,\"id\":\"worker_a\",\"p95_latency_ms\":75810,\"pass_rate\":0.8266,\"provider_group\":\"provider_z\",\"quality\":0.8053,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.915,\"id\":\"worker_b\",\"p95_latency_ms\":63445,\"pass_rate\":0.9901,\"provider_group\":\"provider_y\",\"quality\":0.9726,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.851,\"id\":\"worker_c\",\"p95_latency_ms\":24836,\"pass_rate\":0.5967,\"provider_group\":\"provider_x\",\"quality\":0.7173,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.766,\"estimated_latency_ms\":88281,\"estimated_pass_rate\":0.8663,\"estimated_quality\":0.8846,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.739,\"max_latency_ms\":95782,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4158, "category": "review", "complexity": 0.6906, "consequence": 0.9242, "evidence_need": 0.5635, "family": "policy_compliance", "input_tokens": 5920, "request_id": "synthetic-policy_compliance-000450", "reversibility": 0.6333, "risk": 0.6203, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 3.378, "id": "worker_a", "p95_latency_ms": 75810, "pass_rate": 0.8266, "provider_group": "provider_z", "quality": 0.8053, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.915, "id": "worker_b", "p95_latency_ms": 63445, "pass_rate": 0.9901, "provider_group": "provider_y", "quality": 0.9726, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.851, "id": "worker_c", "p95_latency_ms": 24836, "pass_rate": 0.5967, "provider_group": "provider_x", "quality": 0.7173, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.181, "max_latency_ms": 213902, "minimum_quality": 0.9104}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000597", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.565, "estimated_latency_ms": 171246, "estimated_pass_rate": 0.7608, "estimated_quality": 0.9122, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.181, "max_latency_ms": 213902, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7608, "quality": 0.9122, "quality_target_met": true, "utility": 73.89283}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.181,\"max_latency_ms\":213902,\"minimum_quality\":0.9104},\"task\":{\"ambiguity\":0.5753,\"category\":\"review\",\"complexity\":0.6475,\"consequence\":0.6368,\"evidence_need\":0.8216,\"family\":\"policy_compliance\",\"input_tokens\":2720,\"reversibility\":0.6481,\"risk\":0.5757,\"summary\":\"Identify evidence required before approving usage metering for production.\"},\"workers\":[{\"credits_per_call\":1.199,\"id\":\"worker_a\",\"p95_latency_ms\":27356,\"pass_rate\":0.5456,\"provider_group\":\"provider_z\",\"quality\":0.8032,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.834,\"id\":\"worker_b\",\"p95_latency_ms\":56157,\"pass_rate\":0.7582,\"provider_group\":\"provider_y\",\"quality\":0.9054,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.532,\"id\":\"worker_c\",\"p95_latency_ms\":87733,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9438,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.565,\"estimated_latency_ms\":171246,\"estimated_pass_rate\":0.7608,\"estimated_quality\":0.9122,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.181,\"max_latency_ms\":213902,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5753, "category": "review", "complexity": 0.6475, "consequence": 0.6368, "evidence_need": 0.8216, "family": "policy_compliance", "input_tokens": 2720, "request_id": "synthetic-policy_compliance-000597", "reversibility": 0.6481, "risk": 0.5757, "summary": "Identify evidence required before approving usage metering for production."}, "workers": [{"credits_per_call": 1.199, "id": "worker_a", "p95_latency_ms": 27356, "pass_rate": 0.5456, "provider_group": "provider_z", "quality": 0.8032, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.834, "id": "worker_b", "p95_latency_ms": 56157, "pass_rate": 0.7582, "provider_group": "provider_y", "quality": 0.9054, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.532, "id": "worker_c", "p95_latency_ms": 87733, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9438, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.279, "max_latency_ms": 164277, "minimum_quality": 0.8701}, "family": "migration_planning", "id": "synthetic-migration_planning-001246", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.884, "estimated_latency_ms": 99911, "estimated_pass_rate": 0.7034, "estimated_quality": 0.7929, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.279, "max_latency_ms": 164277, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7034, "quality": 0.7929, "quality_target_met": false, "utility": 55.277576}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.279,\"max_latency_ms\":164277,\"minimum_quality\":0.8701},\"task\":{\"ambiguity\":0.5138,\"category\":\"coding\",\"complexity\":0.9221,\"consequence\":0.7997,\"evidence_need\":0.4446,\"family\":\"migration_planning\",\"input_tokens\":800,\"reversibility\":0.2865,\"risk\":0.5903,\"summary\":\"Review a staged migration for model routing and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":4.03,\"id\":\"worker_a\",\"p95_latency_ms\":48934,\"pass_rate\":0.845,\"provider_group\":\"provider_z\",\"quality\":0.8388,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.742,\"id\":\"worker_b\",\"p95_latency_ms\":77699,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9104,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.142,\"id\":\"worker_c\",\"p95_latency_ms\":22212,\"pass_rate\":0.5885,\"provider_group\":\"provider_y\",\"quality\":0.7084,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.884,\"estimated_latency_ms\":99911,\"estimated_pass_rate\":0.7034,\"estimated_quality\":0.7929,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.279,\"max_latency_ms\":164277,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5138, "category": "coding", "complexity": 0.9221, "consequence": 0.7997, "evidence_need": 0.4446, "family": "migration_planning", "input_tokens": 800, "request_id": "synthetic-migration_planning-001246", "reversibility": 0.2865, "risk": 0.5903, "summary": "Review a staged migration for model routing and identify irreversible steps."}, "workers": [{"credits_per_call": 4.03, "id": "worker_a", "p95_latency_ms": 48934, "pass_rate": 0.845, "provider_group": "provider_z", "quality": 0.8388, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.742, "id": "worker_b", "p95_latency_ms": 77699, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9104, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.142, "id": "worker_c", "p95_latency_ms": 22212, "pass_rate": 0.5885, "provider_group": "provider_y", "quality": 0.7084, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.012, "max_latency_ms": 45418, "minimum_quality": 0.7335}, "family": "migration_planning", "id": "synthetic-migration_planning-000640", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.83, "estimated_latency_ms": 26117, "estimated_pass_rate": 0.3762, "estimated_quality": 0.6658, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.012, "max_latency_ms": 45418, "policy": "lean", "reason_codes": ["best_feasible_below_target", "credit_constrained", "high_consequence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3762, "quality": 0.6658, "quality_target_met": false, "utility": 38.76003}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.012,\"max_latency_ms\":45418,\"minimum_quality\":0.7335},\"task\":{\"ambiguity\":0.6875,\"category\":\"coding\",\"complexity\":0.6668,\"consequence\":0.8069,\"evidence_need\":0.4452,\"family\":\"migration_planning\",\"input_tokens\":1480,\"reversibility\":0.5467,\"risk\":0.3504,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":15.412,\"id\":\"worker_a\",\"p95_latency_ms\":78095,\"pass_rate\":0.9797,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.266,\"id\":\"worker_b\",\"p95_latency_ms\":51282,\"pass_rate\":0.8595,\"provider_group\":\"provider_z\",\"quality\":0.7996,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.83,\"id\":\"worker_c\",\"p95_latency_ms\":26117,\"pass_rate\":0.5322,\"provider_group\":\"provider_y\",\"quality\":0.7778,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.83,\"estimated_latency_ms\":26117,\"estimated_pass_rate\":0.3762,\"estimated_quality\":0.6658,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.012,\"max_latency_ms\":45418,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6875, "category": "coding", "complexity": 0.6668, "consequence": 0.8069, "evidence_need": 0.4452, "family": "migration_planning", "input_tokens": 1480, "request_id": "synthetic-migration_planning-000640", "reversibility": 0.5467, "risk": 0.3504, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 15.412, "id": "worker_a", "p95_latency_ms": 78095, "pass_rate": 0.9797, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.266, "id": "worker_b", "p95_latency_ms": 51282, "pass_rate": 0.8595, "provider_group": "provider_z", "quality": 0.7996, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.83, "id": "worker_c", "p95_latency_ms": 26117, "pass_rate": 0.5322, "provider_group": "provider_y", "quality": 0.7778, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 20.214, "max_latency_ms": 158436, "minimum_quality": 0.9}, "family": "migration_planning", "id": "synthetic-migration_planning-001102", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.85, "estimated_latency_ms": 106846, "estimated_pass_rate": 0.8925, "estimated_quality": 0.9217, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 20.214, "max_latency_ms": 158436, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8925, "quality": 0.9217, "quality_target_met": true, "utility": 70.987866}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":20.214,\"max_latency_ms\":158436,\"minimum_quality\":0.9},\"task\":{\"ambiguity\":0.6653,\"category\":\"coding\",\"complexity\":0.704,\"consequence\":0.7212,\"evidence_need\":0.4639,\"family\":\"migration_planning\",\"input_tokens\":4760,\"reversibility\":0.4359,\"risk\":0.5084,\"summary\":\"Design compatibility testing for a prompt caching schema migration.\"},\"workers\":[{\"credits_per_call\":12.547,\"id\":\"worker_a\",\"p95_latency_ms\":53088,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9924,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.303,\"id\":\"worker_b\",\"p95_latency_ms\":53758,\"pass_rate\":0.8837,\"provider_group\":\"provider_x\",\"quality\":0.8415,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.847,\"id\":\"worker_c\",\"p95_latency_ms\":33076,\"pass_rate\":0.5211,\"provider_group\":\"provider_z\",\"quality\":0.7809,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.85,\"estimated_latency_ms\":106846,\"estimated_pass_rate\":0.8925,\"estimated_quality\":0.9217,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":20.214,\"max_latency_ms\":158436,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6653, "category": "coding", "complexity": 0.704, "consequence": 0.7212, "evidence_need": 0.4639, "family": "migration_planning", "input_tokens": 4760, "request_id": "synthetic-migration_planning-001102", "reversibility": 0.4359, "risk": 0.5084, "summary": "Design compatibility testing for a prompt caching schema migration."}, "workers": [{"credits_per_call": 12.547, "id": "worker_a", "p95_latency_ms": 53088, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9924, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.303, "id": "worker_b", "p95_latency_ms": 53758, "pass_rate": 0.8837, "provider_group": "provider_x", "quality": 0.8415, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.847, "id": "worker_c", "p95_latency_ms": 33076, "pass_rate": 0.5211, "provider_group": "provider_z", "quality": 0.7809, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.745, "max_latency_ms": 235241, "minimum_quality": 0.8365}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000681", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.176, "estimated_latency_ms": 165312, "estimated_pass_rate": 0.7365, "estimated_quality": 0.8861, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.745, "max_latency_ms": 235241, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7365, "quality": 0.8861, "quality_target_met": true, "utility": 75.027824}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.745,\"max_latency_ms\":235241,\"minimum_quality\":0.8365},\"task\":{\"ambiguity\":0.3846,\"category\":\"review\",\"complexity\":0.7924,\"consequence\":0.7072,\"evidence_need\":0.745,\"family\":\"policy_compliance\",\"input_tokens\":680,\"reversibility\":0.6956,\"risk\":0.7377,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.048,\"id\":\"worker_a\",\"p95_latency_ms\":76902,\"pass_rate\":0.7191,\"provider_group\":\"provider_x\",\"quality\":0.8892,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.131,\"id\":\"worker_b\",\"p95_latency_ms\":18260,\"pass_rate\":0.506,\"provider_group\":\"provider_z\",\"quality\":0.7181,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.997,\"id\":\"worker_c\",\"p95_latency_ms\":70150,\"pass_rate\":0.9764,\"provider_group\":\"provider_y\",\"quality\":0.9445,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.176,\"estimated_latency_ms\":165312,\"estimated_pass_rate\":0.7365,\"estimated_quality\":0.8861,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.745,\"max_latency_ms\":235241,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3846, "category": "review", "complexity": 0.7924, "consequence": 0.7072, "evidence_need": 0.745, "family": "policy_compliance", "input_tokens": 680, "request_id": "synthetic-policy_compliance-000681", "reversibility": 0.6956, "risk": 0.7377, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 3.048, "id": "worker_a", "p95_latency_ms": 76902, "pass_rate": 0.7191, "provider_group": "provider_x", "quality": 0.8892, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.131, "id": "worker_b", "p95_latency_ms": 18260, "pass_rate": 0.506, "provider_group": "provider_z", "quality": 0.7181, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.997, "id": "worker_c", "p95_latency_ms": 70150, "pass_rate": 0.9764, "provider_group": "provider_y", "quality": 0.9445, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.664, "max_latency_ms": 36890, "minimum_quality": 0.5455}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000152", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.939, "estimated_latency_ms": 32121, "estimated_pass_rate": 0.2879, "estimated_quality": 0.6486, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.664, "max_latency_ms": 36890, "policy": "lean", "reason_codes": ["independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2879, "quality": 0.6486, "quality_target_met": true, "utility": 38.966523}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.664,\"max_latency_ms\":36890,\"minimum_quality\":0.5455},\"task\":{\"ambiguity\":0.4972,\"category\":\"analysis\",\"complexity\":0.5364,\"consequence\":0.6251,\"evidence_need\":0.6918,\"family\":\"performance_tuning\",\"input_tokens\":6000,\"reversibility\":0.5626,\"risk\":0.321,\"summary\":\"Review a proposed optimization for OAuth callbacks for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.7,\"id\":\"worker_a\",\"p95_latency_ms\":85554,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9924,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.88,\"id\":\"worker_b\",\"p95_latency_ms\":76178,\"pass_rate\":0.8623,\"provider_group\":\"provider_y\",\"quality\":0.9157,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.939,\"id\":\"worker_c\",\"p95_latency_ms\":32121,\"pass_rate\":0.4123,\"provider_group\":\"provider_x\",\"quality\":0.7221,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.939,\"estimated_latency_ms\":32121,\"estimated_pass_rate\":0.2879,\"estimated_quality\":0.6486,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.664,\"max_latency_ms\":36890,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4972, "category": "analysis", "complexity": 0.5364, "consequence": 0.6251, "evidence_need": 0.6918, "family": "performance_tuning", "input_tokens": 6000, "request_id": "synthetic-performance_tuning-000152", "reversibility": 0.5626, "risk": 0.321, "summary": "Review a proposed optimization for OAuth callbacks for hidden regressions."}, "workers": [{"credits_per_call": 11.7, "id": "worker_a", "p95_latency_ms": 85554, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9924, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.88, "id": "worker_b", "p95_latency_ms": 76178, "pass_rate": 0.8623, "provider_group": "provider_y", "quality": 0.9157, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.939, "id": "worker_c", "p95_latency_ms": 32121, "pass_rate": 0.4123, "provider_group": "provider_x", "quality": 0.7221, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.445, "max_latency_ms": 257602, "minimum_quality": 0.7822}, "family": "migration_planning", "id": "synthetic-migration_planning-000517", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.892, "estimated_latency_ms": 192223, "estimated_pass_rate": 0.7315, "estimated_quality": 0.8775, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 26.445, "max_latency_ms": 257602, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7315, "quality": 0.8775, "quality_target_met": true, "utility": 72.552537}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.445,\"max_latency_ms\":257602,\"minimum_quality\":0.7822},\"task\":{\"ambiguity\":0.3596,\"category\":\"coding\",\"complexity\":0.9273,\"consequence\":0.4893,\"evidence_need\":0.6332,\"family\":\"migration_planning\",\"input_tokens\":6720,\"reversibility\":0.4763,\"risk\":0.5969,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.657,\"id\":\"worker_a\",\"p95_latency_ms\":91942,\"pass_rate\":0.8937,\"provider_group\":\"provider_x\",\"quality\":0.9272,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.267,\"id\":\"worker_b\",\"p95_latency_ms\":79566,\"pass_rate\":0.8599,\"provider_group\":\"provider_y\",\"quality\":0.8757,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.968,\"id\":\"worker_c\",\"p95_latency_ms\":20715,\"pass_rate\":0.5426,\"provider_group\":\"provider_z\",\"quality\":0.7096,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.892,\"estimated_latency_ms\":192223,\"estimated_pass_rate\":0.7315,\"estimated_quality\":0.8775,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.445,\"max_latency_ms\":257602,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3596, "category": "coding", "complexity": 0.9273, "consequence": 0.4893, "evidence_need": 0.6332, "family": "migration_planning", "input_tokens": 6720, "request_id": "synthetic-migration_planning-000517", "reversibility": 0.4763, "risk": 0.5969, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 13.657, "id": "worker_a", "p95_latency_ms": 91942, "pass_rate": 0.8937, "provider_group": "provider_x", "quality": 0.9272, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.267, "id": "worker_b", "p95_latency_ms": 79566, "pass_rate": 0.8599, "provider_group": "provider_y", "quality": 0.8757, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 0.968, "id": "worker_c", "p95_latency_ms": 20715, "pass_rate": 0.5426, "provider_group": "provider_z", "quality": 0.7096, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.009, "max_latency_ms": 35489, "minimum_quality": 0.6762}, "family": "migration_planning", "id": "synthetic-migration_planning-000304", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.129, "estimated_latency_ms": 20944, "estimated_pass_rate": 0.3491, "estimated_quality": 0.693, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.009, "max_latency_ms": 35489, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3491, "quality": 0.693, "quality_target_met": true, "utility": 44.405806}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.009,\"max_latency_ms\":35489,\"minimum_quality\":0.6762},\"task\":{\"ambiguity\":0.3078,\"category\":\"coding\",\"complexity\":0.6085,\"consequence\":0.5815,\"evidence_need\":0.7141,\"family\":\"migration_planning\",\"input_tokens\":2760,\"reversibility\":0.6091,\"risk\":0.4728,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":3.9,\"id\":\"worker_a\",\"p95_latency_ms\":74302,\"pass_rate\":0.784,\"provider_group\":\"provider_z\",\"quality\":0.921,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.195,\"id\":\"worker_b\",\"p95_latency_ms\":58887,\"pass_rate\":0.9597,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.129,\"id\":\"worker_c\",\"p95_latency_ms\":20944,\"pass_rate\":0.4797,\"provider_group\":\"provider_x\",\"quality\":0.7777,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.129,\"estimated_latency_ms\":20944,\"estimated_pass_rate\":0.3491,\"estimated_quality\":0.693,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.009,\"max_latency_ms\":35489,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3078, "category": "coding", "complexity": 0.6085, "consequence": 0.5815, "evidence_need": 0.7141, "family": "migration_planning", "input_tokens": 2760, "request_id": "synthetic-migration_planning-000304", "reversibility": 0.6091, "risk": 0.4728, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 3.9, "id": "worker_a", "p95_latency_ms": 74302, "pass_rate": 0.784, "provider_group": "provider_z", "quality": 0.921, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.195, "id": "worker_b", "p95_latency_ms": 58887, "pass_rate": 0.9597, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.129, "id": "worker_c", "p95_latency_ms": 20944, "pass_rate": 0.4797, "provider_group": "provider_x", "quality": 0.7777, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.673, "max_latency_ms": 168490, "minimum_quality": 0.8835}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000224", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.059, "estimated_latency_ms": 155365, "estimated_pass_rate": 0.7277, "estimated_quality": 0.9155, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 19.673, "max_latency_ms": 168490, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7277, "quality": 0.9155, "quality_target_met": true, "utility": 67.969309}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.673,\"max_latency_ms\":168490,\"minimum_quality\":0.8835},\"task\":{\"ambiguity\":0.4752,\"category\":\"analysis\",\"complexity\":0.866,\"consequence\":0.4471,\"evidence_need\":0.5299,\"family\":\"performance_tuning\",\"input_tokens\":6440,\"reversibility\":0.7858,\"risk\":0.2504,\"summary\":\"Diagnose latency variance in schema evolution and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.07,\"id\":\"worker_a\",\"p95_latency_ms\":25099,\"pass_rate\":0.4793,\"provider_group\":\"provider_z\",\"quality\":0.6967,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.708,\"id\":\"worker_b\",\"p95_latency_ms\":84042,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.281,\"id\":\"worker_c\",\"p95_latency_ms\":46224,\"pass_rate\":0.8566,\"provider_group\":\"provider_y\",\"quality\":0.9291,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.059,\"estimated_latency_ms\":155365,\"estimated_pass_rate\":0.7277,\"estimated_quality\":0.9155,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.673,\"max_latency_ms\":168490,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4752, "category": "analysis", "complexity": 0.866, "consequence": 0.4471, "evidence_need": 0.5299, "family": "performance_tuning", "input_tokens": 6440, "request_id": "synthetic-performance_tuning-000224", "reversibility": 0.7858, "risk": 0.2504, "summary": "Diagnose latency variance in schema evolution and rank the next experiments."}, "workers": [{"credits_per_call": 1.07, "id": "worker_a", "p95_latency_ms": 25099, "pass_rate": 0.4793, "provider_group": "provider_z", "quality": 0.6967, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.708, "id": "worker_b", "p95_latency_ms": 84042, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.281, "id": "worker_c", "p95_latency_ms": 46224, "pass_rate": 0.8566, "provider_group": "provider_y", "quality": 0.9291, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.846, "max_latency_ms": 36803, "minimum_quality": 0.7658}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000858", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.041, "estimated_latency_ms": 20791, "estimated_pass_rate": 0.3344, "estimated_quality": 0.6531, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.846, "max_latency_ms": 36803, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3344, "quality": 0.6531, "quality_target_met": false, "utility": 36.960376}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.846,\"max_latency_ms\":36803,\"minimum_quality\":0.7658},\"task\":{\"ambiguity\":0.6732,\"category\":\"review\",\"complexity\":0.6235,\"consequence\":0.5399,\"evidence_need\":0.6737,\"family\":\"policy_compliance\",\"input_tokens\":3280,\"reversibility\":0.4255,\"risk\":0.618,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.087,\"id\":\"worker_a\",\"p95_latency_ms\":56117,\"pass_rate\":0.979,\"provider_group\":\"provider_y\",\"quality\":0.9168,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.041,\"id\":\"worker_b\",\"p95_latency_ms\":20791,\"pass_rate\":0.5061,\"provider_group\":\"provider_x\",\"quality\":0.7756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.301,\"id\":\"worker_c\",\"p95_latency_ms\":81522,\"pass_rate\":0.8037,\"provider_group\":\"provider_z\",\"quality\":0.8906,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.041,\"estimated_latency_ms\":20791,\"estimated_pass_rate\":0.3344,\"estimated_quality\":0.6531,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.846,\"max_latency_ms\":36803,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6732, "category": "review", "complexity": 0.6235, "consequence": 0.5399, "evidence_need": 0.6737, "family": "policy_compliance", "input_tokens": 3280, "request_id": "synthetic-policy_compliance-000858", "reversibility": 0.4255, "risk": 0.618, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 15.087, "id": "worker_a", "p95_latency_ms": 56117, "pass_rate": 0.979, "provider_group": "provider_y", "quality": 0.9168, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.041, "id": "worker_b", "p95_latency_ms": 20791, "pass_rate": 0.5061, "provider_group": "provider_x", "quality": 0.7756, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.301, "id": "worker_c", "p95_latency_ms": 81522, "pass_rate": 0.8037, "provider_group": "provider_z", "quality": 0.8906, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.482, "max_latency_ms": 227383, "minimum_quality": 0.8099}, "family": "migration_planning", "id": "synthetic-migration_planning-000112", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.763, "estimated_latency_ms": 157049, "estimated_pass_rate": 0.7386, "estimated_quality": 0.9264, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.482, "max_latency_ms": 227383, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7386, "quality": 0.9264, "quality_target_met": true, "utility": 75.344523}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.482,\"max_latency_ms\":227383,\"minimum_quality\":0.8099},\"task\":{\"ambiguity\":0.6801,\"category\":\"coding\",\"complexity\":0.7087,\"consequence\":0.7069,\"evidence_need\":0.822,\"family\":\"migration_planning\",\"input_tokens\":960,\"reversibility\":0.314,\"risk\":0.4065,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.941,\"id\":\"worker_a\",\"p95_latency_ms\":21583,\"pass_rate\":0.606,\"provider_group\":\"provider_y\",\"quality\":0.7876,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.19,\"id\":\"worker_b\",\"p95_latency_ms\":63719,\"pass_rate\":0.8165,\"provider_group\":\"provider_x\",\"quality\":0.8596,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.632,\"id\":\"worker_c\",\"p95_latency_ms\":71747,\"pass_rate\":0.9063,\"provider_group\":\"provider_z\",\"quality\":0.9883,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.763,\"estimated_latency_ms\":157049,\"estimated_pass_rate\":0.7386,\"estimated_quality\":0.9264,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.482,\"max_latency_ms\":227383,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6801, "category": "coding", "complexity": 0.7087, "consequence": 0.7069, "evidence_need": 0.822, "family": "migration_planning", "input_tokens": 960, "request_id": "synthetic-migration_planning-000112", "reversibility": 0.314, "risk": 0.4065, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 0.941, "id": "worker_a", "p95_latency_ms": 21583, "pass_rate": 0.606, "provider_group": "provider_y", "quality": 0.7876, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.19, "id": "worker_b", "p95_latency_ms": 63719, "pass_rate": 0.8165, "provider_group": "provider_x", "quality": 0.8596, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.632, "id": "worker_c", "p95_latency_ms": 71747, "pass_rate": 0.9063, "provider_group": "provider_z", "quality": 0.9883, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.547, "max_latency_ms": 193964, "minimum_quality": 0.8545}, "family": "migration_planning", "id": "synthetic-migration_planning-000373", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.415, "estimated_latency_ms": 135640, "estimated_pass_rate": 0.6853, "estimated_quality": 0.8754, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 21.547, "max_latency_ms": 193964, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6853, "quality": 0.8754, "quality_target_met": true, "utility": 67.324468}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.547,\"max_latency_ms\":193964,\"minimum_quality\":0.8545},\"task\":{\"ambiguity\":0.3639,\"category\":\"coding\",\"complexity\":0.778,\"consequence\":0.4519,\"evidence_need\":0.6198,\"family\":\"migration_planning\",\"input_tokens\":6120,\"reversibility\":0.3054,\"risk\":0.4824,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.799,\"id\":\"worker_a\",\"p95_latency_ms\":24413,\"pass_rate\":0.4842,\"provider_group\":\"provider_x\",\"quality\":0.747,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.264,\"id\":\"worker_b\",\"p95_latency_ms\":51115,\"pass_rate\":0.7617,\"provider_group\":\"provider_y\",\"quality\":0.8639,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.352,\"id\":\"worker_c\",\"p95_latency_ms\":60112,\"pass_rate\":0.8897,\"provider_group\":\"provider_z\",\"quality\":0.9185,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.415,\"estimated_latency_ms\":135640,\"estimated_pass_rate\":0.6853,\"estimated_quality\":0.8754,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.547,\"max_latency_ms\":193964,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3639, "category": "coding", "complexity": 0.778, "consequence": 0.4519, "evidence_need": 0.6198, "family": "migration_planning", "input_tokens": 6120, "request_id": "synthetic-migration_planning-000373", "reversibility": 0.3054, "risk": 0.4824, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 0.799, "id": "worker_a", "p95_latency_ms": 24413, "pass_rate": 0.4842, "provider_group": "provider_x", "quality": 0.747, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.264, "id": "worker_b", "p95_latency_ms": 51115, "pass_rate": 0.7617, "provider_group": "provider_y", "quality": 0.8639, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 13.352, "id": "worker_c", "p95_latency_ms": 60112, "pass_rate": 0.8897, "provider_group": "provider_z", "quality": 0.9185, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.133, "max_latency_ms": 50995, "minimum_quality": 0.7687}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000113", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.092, "estimated_latency_ms": 31127, "estimated_pass_rate": 0.338, "estimated_quality": 0.6405, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.133, "max_latency_ms": 50995, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.338, "quality": 0.6405, "quality_target_met": false, "utility": 35.770146}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.133,\"max_latency_ms\":50995,\"minimum_quality\":0.7687},\"task\":{\"ambiguity\":0.7301,\"category\":\"analysis\",\"complexity\":0.5703,\"consequence\":0.6326,\"evidence_need\":0.6417,\"family\":\"performance_tuning\",\"input_tokens\":6520,\"reversibility\":0.5855,\"risk\":0.2536,\"summary\":\"Review a proposed optimization for observability for hidden regressions.\"},\"workers\":[{\"credits_per_call\":4.19,\"id\":\"worker_a\",\"p95_latency_ms\":66153,\"pass_rate\":0.9047,\"provider_group\":\"provider_z\",\"quality\":0.8905,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.092,\"id\":\"worker_b\",\"p95_latency_ms\":31127,\"pass_rate\":0.4784,\"provider_group\":\"provider_x\",\"quality\":0.7309,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.577,\"id\":\"worker_c\",\"p95_latency_ms\":91231,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.092,\"estimated_latency_ms\":31127,\"estimated_pass_rate\":0.338,\"estimated_quality\":0.6405,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.133,\"max_latency_ms\":50995,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7301, "category": "analysis", "complexity": 0.5703, "consequence": 0.6326, "evidence_need": 0.6417, "family": "performance_tuning", "input_tokens": 6520, "request_id": "synthetic-performance_tuning-000113", "reversibility": 0.5855, "risk": 0.2536, "summary": "Review a proposed optimization for observability for hidden regressions."}, "workers": [{"credits_per_call": 4.19, "id": "worker_a", "p95_latency_ms": 66153, "pass_rate": 0.9047, "provider_group": "provider_z", "quality": 0.8905, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.092, "id": "worker_b", "p95_latency_ms": 31127, "pass_rate": 0.4784, "provider_group": "provider_x", "quality": 0.7309, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.577, "id": "worker_c", "p95_latency_ms": 91231, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.301, "max_latency_ms": 47433, "minimum_quality": 0.6628}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000318", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.027, "estimated_latency_ms": 29291, "estimated_pass_rate": 0.432, "estimated_quality": 0.712, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.301, "max_latency_ms": 47433, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.432, "quality": 0.712, "quality_target_met": true, "utility": 48.17685}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.301,\"max_latency_ms\":47433,\"minimum_quality\":0.6628},\"task\":{\"ambiguity\":0.5172,\"category\":\"review\",\"complexity\":0.4289,\"consequence\":0.5678,\"evidence_need\":0.933,\"family\":\"policy_compliance\",\"input_tokens\":6920,\"reversibility\":0.3673,\"risk\":0.6504,\"summary\":\"Review whether database failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.027,\"id\":\"worker_a\",\"p95_latency_ms\":29291,\"pass_rate\":0.5724,\"provider_group\":\"provider_x\",\"quality\":0.7911,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.811,\"id\":\"worker_b\",\"p95_latency_ms\":88648,\"pass_rate\":0.9238,\"provider_group\":\"provider_y\",\"quality\":0.937,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.988,\"id\":\"worker_c\",\"p95_latency_ms\":66230,\"pass_rate\":0.7981,\"provider_group\":\"provider_z\",\"quality\":0.8633,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.027,\"estimated_latency_ms\":29291,\"estimated_pass_rate\":0.432,\"estimated_quality\":0.712,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.301,\"max_latency_ms\":47433,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5172, "category": "review", "complexity": 0.4289, "consequence": 0.5678, "evidence_need": 0.933, "family": "policy_compliance", "input_tokens": 6920, "request_id": "synthetic-policy_compliance-000318", "reversibility": 0.3673, "risk": 0.6504, "summary": "Review whether database failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.027, "id": "worker_a", "p95_latency_ms": 29291, "pass_rate": 0.5724, "provider_group": "provider_x", "quality": 0.7911, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.811, "id": "worker_b", "p95_latency_ms": 88648, "pass_rate": 0.9238, "provider_group": "provider_y", "quality": 0.937, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.988, "id": "worker_c", "p95_latency_ms": 66230, "pass_rate": 0.7981, "provider_group": "provider_z", "quality": 0.8633, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.275, "max_latency_ms": 223036, "minimum_quality": 0.9457}, "family": "migration_planning", "id": "synthetic-migration_planning-000817", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.507, "estimated_latency_ms": 159902, "estimated_pass_rate": 0.756, "estimated_quality": 0.8999, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 16.275, "max_latency_ms": 223036, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.756, "quality": 0.8999, "quality_target_met": false, "utility": 66.882988}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.275,\"max_latency_ms\":223036,\"minimum_quality\":0.9457},\"task\":{\"ambiguity\":0.3377,\"category\":\"coding\",\"complexity\":0.7876,\"consequence\":0.5383,\"evidence_need\":0.4357,\"family\":\"migration_planning\",\"input_tokens\":1600,\"reversibility\":0.4937,\"risk\":0.4833,\"summary\":\"Review a staged migration for tenant-scoped caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":2.87,\"id\":\"worker_a\",\"p95_latency_ms\":55817,\"pass_rate\":0.8154,\"provider_group\":\"provider_y\",\"quality\":0.7935,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.477,\"id\":\"worker_b\",\"p95_latency_ms\":75331,\"pass_rate\":0.9378,\"provider_group\":\"provider_x\",\"quality\":0.9748,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.16,\"id\":\"worker_c\",\"p95_latency_ms\":28754,\"pass_rate\":0.6175,\"provider_group\":\"provider_z\",\"quality\":0.7587,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.507,\"estimated_latency_ms\":159902,\"estimated_pass_rate\":0.756,\"estimated_quality\":0.8999,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.275,\"max_latency_ms\":223036,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3377, "category": "coding", "complexity": 0.7876, "consequence": 0.5383, "evidence_need": 0.4357, "family": "migration_planning", "input_tokens": 1600, "request_id": "synthetic-migration_planning-000817", "reversibility": 0.4937, "risk": 0.4833, "summary": "Review a staged migration for tenant-scoped caching and identify irreversible steps."}, "workers": [{"credits_per_call": 2.87, "id": "worker_a", "p95_latency_ms": 55817, "pass_rate": 0.8154, "provider_group": "provider_y", "quality": 0.7935, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.477, "id": "worker_b", "p95_latency_ms": 75331, "pass_rate": 0.9378, "provider_group": "provider_x", "quality": 0.9748, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.16, "id": "worker_c", "p95_latency_ms": 28754, "pass_rate": 0.6175, "provider_group": "provider_z", "quality": 0.7587, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.749, "max_latency_ms": 230257, "minimum_quality": 0.9132}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000954", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.866, "estimated_latency_ms": 162838, "estimated_pass_rate": 0.7736, "estimated_quality": 0.9227, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.749, "max_latency_ms": 230257, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7736, "quality": 0.9227, "quality_target_met": true, "utility": 75.795478}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.749,\"max_latency_ms\":230257,\"minimum_quality\":0.9132},\"task\":{\"ambiguity\":0.4843,\"category\":\"review\",\"complexity\":0.6931,\"consequence\":0.6845,\"evidence_need\":0.5549,\"family\":\"policy_compliance\",\"input_tokens\":7760,\"reversibility\":0.4225,\"risk\":0.6886,\"summary\":\"Identify evidence required before approving artifact storage for production.\"},\"workers\":[{\"credits_per_call\":2.812,\"id\":\"worker_a\",\"p95_latency_ms\":70165,\"pass_rate\":0.8665,\"provider_group\":\"provider_z\",\"quality\":0.9173,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.18,\"id\":\"worker_b\",\"p95_latency_ms\":20210,\"pass_rate\":0.5453,\"provider_group\":\"provider_y\",\"quality\":0.7876,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.874,\"id\":\"worker_c\",\"p95_latency_ms\":72463,\"pass_rate\":0.9433,\"provider_group\":\"provider_x\",\"quality\":0.9785,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.866,\"estimated_latency_ms\":162838,\"estimated_pass_rate\":0.7736,\"estimated_quality\":0.9227,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.749,\"max_latency_ms\":230257,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4843, "category": "review", "complexity": 0.6931, "consequence": 0.6845, "evidence_need": 0.5549, "family": "policy_compliance", "input_tokens": 7760, "request_id": "synthetic-policy_compliance-000954", "reversibility": 0.4225, "risk": 0.6886, "summary": "Identify evidence required before approving artifact storage for production."}, "workers": [{"credits_per_call": 2.812, "id": "worker_a", "p95_latency_ms": 70165, "pass_rate": 0.8665, "provider_group": "provider_z", "quality": 0.9173, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.18, "id": "worker_b", "p95_latency_ms": 20210, "pass_rate": 0.5453, "provider_group": "provider_y", "quality": 0.7876, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.874, "id": "worker_c", "p95_latency_ms": 72463, "pass_rate": 0.9433, "provider_group": "provider_x", "quality": 0.9785, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.928, "max_latency_ms": 216929, "minimum_quality": 0.7924}, "family": "migration_planning", "id": "synthetic-migration_planning-001030", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.571, "estimated_latency_ms": 186880, "estimated_pass_rate": 0.7715, "estimated_quality": 0.8934, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.928, "max_latency_ms": 216929, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7715, "quality": 0.8934, "quality_target_met": true, "utility": 70.856501}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.928,\"max_latency_ms\":216929,\"minimum_quality\":0.7924},\"task\":{\"ambiguity\":0.517,\"category\":\"coding\",\"complexity\":0.6844,\"consequence\":0.4882,\"evidence_need\":0.7824,\"family\":\"migration_planning\",\"input_tokens\":2880,\"reversibility\":0.3041,\"risk\":0.5251,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.085,\"id\":\"worker_a\",\"p95_latency_ms\":24893,\"pass_rate\":0.5755,\"provider_group\":\"provider_z\",\"quality\":0.7799,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.032,\"id\":\"worker_b\",\"p95_latency_ms\":75954,\"pass_rate\":0.7945,\"provider_group\":\"provider_x\",\"quality\":0.8785,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.454,\"id\":\"worker_c\",\"p95_latency_ms\":86033,\"pass_rate\":0.994,\"provider_group\":\"provider_y\",\"quality\":0.9122,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.571,\"estimated_latency_ms\":186880,\"estimated_pass_rate\":0.7715,\"estimated_quality\":0.8934,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.928,\"max_latency_ms\":216929,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.517, "category": "coding", "complexity": 0.6844, "consequence": 0.4882, "evidence_need": 0.7824, "family": "migration_planning", "input_tokens": 2880, "request_id": "synthetic-migration_planning-001030", "reversibility": 0.3041, "risk": 0.5251, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 1.085, "id": "worker_a", "p95_latency_ms": 24893, "pass_rate": 0.5755, "provider_group": "provider_z", "quality": 0.7799, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.032, "id": "worker_b", "p95_latency_ms": 75954, "pass_rate": 0.7945, "provider_group": "provider_x", "quality": 0.8785, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.454, "id": "worker_c", "p95_latency_ms": 86033, "pass_rate": 0.994, "provider_group": "provider_y", "quality": 0.9122, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.494, "max_latency_ms": 161899, "minimum_quality": 0.7184}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001035", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.107, "estimated_latency_ms": 86255, "estimated_pass_rate": 0.5836, "estimated_quality": 0.8207, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.494, "max_latency_ms": 161899, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5836, "quality": 0.8207, "quality_target_met": true, "utility": 63.102442}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.494,\"max_latency_ms\":161899,\"minimum_quality\":0.7184},\"task\":{\"ambiguity\":0.5031,\"category\":\"review\",\"complexity\":0.6981,\"consequence\":0.789,\"evidence_need\":0.7421,\"family\":\"policy_compliance\",\"input_tokens\":5080,\"reversibility\":0.5675,\"risk\":0.5968,\"summary\":\"Identify evidence required before approving artifact storage for production.\"},\"workers\":[{\"credits_per_call\":14.534,\"id\":\"worker_a\",\"p95_latency_ms\":61703,\"pass_rate\":0.9746,\"provider_group\":\"provider_x\",\"quality\":0.9862,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.266,\"id\":\"worker_b\",\"p95_latency_ms\":63116,\"pass_rate\":0.8152,\"provider_group\":\"provider_z\",\"quality\":0.8868,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.841,\"id\":\"worker_c\",\"p95_latency_ms\":23139,\"pass_rate\":0.464,\"provider_group\":\"provider_y\",\"quality\":0.7544,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.107,\"estimated_latency_ms\":86255,\"estimated_pass_rate\":0.5836,\"estimated_quality\":0.8207,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.494,\"max_latency_ms\":161899,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5031, "category": "review", "complexity": 0.6981, "consequence": 0.789, "evidence_need": 0.7421, "family": "policy_compliance", "input_tokens": 5080, "request_id": "synthetic-policy_compliance-001035", "reversibility": 0.5675, "risk": 0.5968, "summary": "Identify evidence required before approving artifact storage for production."}, "workers": [{"credits_per_call": 14.534, "id": "worker_a", "p95_latency_ms": 61703, "pass_rate": 0.9746, "provider_group": "provider_x", "quality": 0.9862, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.266, "id": "worker_b", "p95_latency_ms": 63116, "pass_rate": 0.8152, "provider_group": "provider_z", "quality": 0.8868, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.841, "id": "worker_c", "p95_latency_ms": 23139, "pass_rate": 0.464, "provider_group": "provider_y", "quality": 0.7544, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.847, "max_latency_ms": 194257, "minimum_quality": 0.8287}, "family": "migration_planning", "id": "synthetic-migration_planning-000241", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.372, "estimated_latency_ms": 156871, "estimated_pass_rate": 0.6988, "estimated_quality": 0.8551, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.847, "max_latency_ms": 194257, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6988, "quality": 0.8551, "quality_target_met": true, "utility": 69.828102}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.847,\"max_latency_ms\":194257,\"minimum_quality\":0.8287},\"task\":{\"ambiguity\":0.6455,\"category\":\"coding\",\"complexity\":0.8671,\"consequence\":0.5894,\"evidence_need\":0.4048,\"family\":\"migration_planning\",\"input_tokens\":1000,\"reversibility\":0.2883,\"risk\":0.6806,\"summary\":\"Review a staged migration for provider failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.559,\"id\":\"worker_a\",\"p95_latency_ms\":50216,\"pass_rate\":0.7854,\"provider_group\":\"provider_z\",\"quality\":0.7982,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.977,\"id\":\"worker_b\",\"p95_latency_ms\":32452,\"pass_rate\":0.4703,\"provider_group\":\"provider_x\",\"quality\":0.7005,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.836,\"id\":\"worker_c\",\"p95_latency_ms\":74203,\"pass_rate\":0.96,\"provider_group\":\"provider_y\",\"quality\":0.9245,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.372,\"estimated_latency_ms\":156871,\"estimated_pass_rate\":0.6988,\"estimated_quality\":0.8551,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.847,\"max_latency_ms\":194257,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6455, "category": "coding", "complexity": 0.8671, "consequence": 0.5894, "evidence_need": 0.4048, "family": "migration_planning", "input_tokens": 1000, "request_id": "synthetic-migration_planning-000241", "reversibility": 0.2883, "risk": 0.6806, "summary": "Review a staged migration for provider failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.559, "id": "worker_a", "p95_latency_ms": 50216, "pass_rate": 0.7854, "provider_group": "provider_z", "quality": 0.7982, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.977, "id": "worker_b", "p95_latency_ms": 32452, "pass_rate": 0.4703, "provider_group": "provider_x", "quality": 0.7005, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.836, "id": "worker_c", "p95_latency_ms": 74203, "pass_rate": 0.96, "provider_group": "provider_y", "quality": 0.9245, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.967, "max_latency_ms": 191627, "minimum_quality": 0.7241}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001271", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.843, "estimated_latency_ms": 132670, "estimated_pass_rate": 0.9099, "estimated_quality": 0.9458, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.967, "max_latency_ms": 191627, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9099, "quality": 0.9458, "quality_target_met": true, "utility": 69.676704}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.967,\"max_latency_ms\":191627,\"minimum_quality\":0.7241},\"task\":{\"ambiguity\":0.6455,\"category\":\"analysis\",\"complexity\":0.4503,\"consequence\":0.2838,\"evidence_need\":0.3688,\"family\":\"performance_tuning\",\"input_tokens\":5320,\"reversibility\":0.6678,\"risk\":0.5373,\"summary\":\"Choose between throughput and tail-latency changes for model routing.\"},\"workers\":[{\"credits_per_call\":0.815,\"id\":\"worker_a\",\"p95_latency_ms\":21593,\"pass_rate\":0.4847,\"provider_group\":\"provider_x\",\"quality\":0.7178,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.168,\"id\":\"worker_b\",\"p95_latency_ms\":76119,\"pass_rate\":0.8535,\"provider_group\":\"provider_y\",\"quality\":0.8642,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.675,\"id\":\"worker_c\",\"p95_latency_ms\":56551,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.843,\"estimated_latency_ms\":132670,\"estimated_pass_rate\":0.9099,\"estimated_quality\":0.9458,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.967,\"max_latency_ms\":191627,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6455, "category": "analysis", "complexity": 0.4503, "consequence": 0.2838, "evidence_need": 0.3688, "family": "performance_tuning", "input_tokens": 5320, "request_id": "synthetic-performance_tuning-001271", "reversibility": 0.6678, "risk": 0.5373, "summary": "Choose between throughput and tail-latency changes for model routing."}, "workers": [{"credits_per_call": 0.815, "id": "worker_a", "p95_latency_ms": 21593, "pass_rate": 0.4847, "provider_group": "provider_x", "quality": 0.7178, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.168, "id": "worker_b", "p95_latency_ms": 76119, "pass_rate": 0.8535, "provider_group": "provider_y", "quality": 0.8642, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.675, "id": "worker_c", "p95_latency_ms": 56551, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.132, "max_latency_ms": 187912, "minimum_quality": 0.8998}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001242", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.524, "estimated_latency_ms": 137566, "estimated_pass_rate": 0.7662, "estimated_quality": 0.9243, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.132, "max_latency_ms": 187912, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7662, "quality": 0.9243, "quality_target_met": true, "utility": 78.064561}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.132,\"max_latency_ms\":187912,\"minimum_quality\":0.8998},\"task\":{\"ambiguity\":0.4303,\"category\":\"review\",\"complexity\":0.8351,\"consequence\":0.7743,\"evidence_need\":0.9249,\"family\":\"policy_compliance\",\"input_tokens\":6480,\"reversibility\":0.5042,\"risk\":0.6,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.447,\"id\":\"worker_a\",\"p95_latency_ms\":57817,\"pass_rate\":0.9858,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.168,\"id\":\"worker_b\",\"p95_latency_ms\":55642,\"pass_rate\":0.7742,\"provider_group\":\"provider_z\",\"quality\":0.8699,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.909,\"id\":\"worker_c\",\"p95_latency_ms\":24107,\"pass_rate\":0.5829,\"provider_group\":\"provider_y\",\"quality\":0.7436,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.524,\"estimated_latency_ms\":137566,\"estimated_pass_rate\":0.7662,\"estimated_quality\":0.9243,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.132,\"max_latency_ms\":187912,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4303, "category": "review", "complexity": 0.8351, "consequence": 0.7743, "evidence_need": 0.9249, "family": "policy_compliance", "input_tokens": 6480, "request_id": "synthetic-policy_compliance-001242", "reversibility": 0.5042, "risk": 0.6, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 15.447, "id": "worker_a", "p95_latency_ms": 57817, "pass_rate": 0.9858, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.168, "id": "worker_b", "p95_latency_ms": 55642, "pass_rate": 0.7742, "provider_group": "provider_z", "quality": 0.8699, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.909, "id": "worker_c", "p95_latency_ms": 24107, "pass_rate": 0.5829, "provider_group": "provider_y", "quality": 0.7436, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.539, "max_latency_ms": 142995, "minimum_quality": 0.7148}, "family": "migration_planning", "id": "synthetic-migration_planning-001162", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.693, "estimated_latency_ms": 102252, "estimated_pass_rate": 0.695, "estimated_quality": 0.8508, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.539, "max_latency_ms": 142995, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.695, "quality": 0.8508, "quality_target_met": true, "utility": 62.830596}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.539,\"max_latency_ms\":142995,\"minimum_quality\":0.7148},\"task\":{\"ambiguity\":0.4799,\"category\":\"coding\",\"complexity\":0.6772,\"consequence\":0.8419,\"evidence_need\":0.7365,\"family\":\"migration_planning\",\"input_tokens\":7320,\"reversibility\":0.3686,\"risk\":0.4707,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":12.543,\"id\":\"worker_a\",\"p95_latency_ms\":78575,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.825,\"id\":\"worker_b\",\"p95_latency_ms\":72262,\"pass_rate\":0.7683,\"provider_group\":\"provider_z\",\"quality\":0.9151,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.868,\"id\":\"worker_c\",\"p95_latency_ms\":29990,\"pass_rate\":0.5446,\"provider_group\":\"provider_y\",\"quality\":0.756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.693,\"estimated_latency_ms\":102252,\"estimated_pass_rate\":0.695,\"estimated_quality\":0.8508,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.539,\"max_latency_ms\":142995,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4799, "category": "coding", "complexity": 0.6772, "consequence": 0.8419, "evidence_need": 0.7365, "family": "migration_planning", "input_tokens": 7320, "request_id": "synthetic-migration_planning-001162", "reversibility": 0.3686, "risk": 0.4707, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 12.543, "id": "worker_a", "p95_latency_ms": 78575, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.825, "id": "worker_b", "p95_latency_ms": 72262, "pass_rate": 0.7683, "provider_group": "provider_z", "quality": 0.9151, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.868, "id": "worker_c", "p95_latency_ms": 29990, "pass_rate": 0.5446, "provider_group": "provider_y", "quality": 0.756, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.507, "max_latency_ms": 193824, "minimum_quality": 0.8431}, "family": "migration_planning", "id": "synthetic-migration_planning-000733", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 20.173, "estimated_latency_ms": 137357, "estimated_pass_rate": 0.7265, "estimated_quality": 0.8943, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 24.507, "max_latency_ms": 193824, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7265, "quality": 0.8943, "quality_target_met": true, "utility": 71.625507}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.507,\"max_latency_ms\":193824,\"minimum_quality\":0.8431},\"task\":{\"ambiguity\":0.4244,\"category\":\"coding\",\"complexity\":0.7227,\"consequence\":0.7851,\"evidence_need\":0.4995,\"family\":\"migration_planning\",\"input_tokens\":1720,\"reversibility\":0.5262,\"risk\":0.5878,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.999,\"id\":\"worker_a\",\"p95_latency_ms\":50488,\"pass_rate\":0.9527,\"provider_group\":\"provider_x\",\"quality\":0.9557,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.101,\"id\":\"worker_b\",\"p95_latency_ms\":28538,\"pass_rate\":0.5175,\"provider_group\":\"provider_z\",\"quality\":0.7626,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.073,\"id\":\"worker_c\",\"p95_latency_ms\":58331,\"pass_rate\":0.7365,\"provider_group\":\"provider_y\",\"quality\":0.8763,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":20.173,\"estimated_latency_ms\":137357,\"estimated_pass_rate\":0.7265,\"estimated_quality\":0.8943,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.507,\"max_latency_ms\":193824,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4244, "category": "coding", "complexity": 0.7227, "consequence": 0.7851, "evidence_need": 0.4995, "family": "migration_planning", "input_tokens": 1720, "request_id": "synthetic-migration_planning-000733", "reversibility": 0.5262, "risk": 0.5878, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 14.999, "id": "worker_a", "p95_latency_ms": 50488, "pass_rate": 0.9527, "provider_group": "provider_x", "quality": 0.9557, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.101, "id": "worker_b", "p95_latency_ms": 28538, "pass_rate": 0.5175, "provider_group": "provider_z", "quality": 0.7626, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.073, "id": "worker_c", "p95_latency_ms": 58331, "pass_rate": 0.7365, "provider_group": "provider_y", "quality": 0.8763, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.339, "max_latency_ms": 237066, "minimum_quality": 0.8247}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000512", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.865, "estimated_latency_ms": 180371, "estimated_pass_rate": 0.7506, "estimated_quality": 0.9018, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.339, "max_latency_ms": 237066, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7506, "quality": 0.9018, "quality_target_met": true, "utility": 71.215522}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.339,\"max_latency_ms\":237066,\"minimum_quality\":0.8247},\"task\":{\"ambiguity\":0.6531,\"category\":\"analysis\",\"complexity\":0.7217,\"consequence\":0.6093,\"evidence_need\":0.3956,\"family\":\"performance_tuning\",\"input_tokens\":7920,\"reversibility\":0.8062,\"risk\":0.5682,\"summary\":\"Choose between throughput and tail-latency changes for deployment rollback.\"},\"workers\":[{\"credits_per_call\":3.418,\"id\":\"worker_a\",\"p95_latency_ms\":76025,\"pass_rate\":0.8342,\"provider_group\":\"provider_y\",\"quality\":0.8897,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.115,\"id\":\"worker_b\",\"p95_latency_ms\":18688,\"pass_rate\":0.5323,\"provider_group\":\"provider_x\",\"quality\":0.72,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.332,\"id\":\"worker_c\",\"p95_latency_ms\":85658,\"pass_rate\":0.9646,\"provider_group\":\"provider_z\",\"quality\":0.9818,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.865,\"estimated_latency_ms\":180371,\"estimated_pass_rate\":0.7506,\"estimated_quality\":0.9018,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.339,\"max_latency_ms\":237066,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6531, "category": "analysis", "complexity": 0.7217, "consequence": 0.6093, "evidence_need": 0.3956, "family": "performance_tuning", "input_tokens": 7920, "request_id": "synthetic-performance_tuning-000512", "reversibility": 0.8062, "risk": 0.5682, "summary": "Choose between throughput and tail-latency changes for deployment rollback."}, "workers": [{"credits_per_call": 3.418, "id": "worker_a", "p95_latency_ms": 76025, "pass_rate": 0.8342, "provider_group": "provider_y", "quality": 0.8897, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.115, "id": "worker_b", "p95_latency_ms": 18688, "pass_rate": 0.5323, "provider_group": "provider_x", "quality": 0.72, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.332, "id": "worker_c", "p95_latency_ms": 85658, "pass_rate": 0.9646, "provider_group": "provider_z", "quality": 0.9818, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.795, "max_latency_ms": 46734, "minimum_quality": 0.6045}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000839", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.072, "estimated_latency_ms": 28718, "estimated_pass_rate": 0.3137, "estimated_quality": 0.568, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.795, "max_latency_ms": 46734, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3137, "quality": 0.568, "quality_target_met": false, "utility": 36.155516}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.795,\"max_latency_ms\":46734,\"minimum_quality\":0.6045},\"task\":{\"ambiguity\":0.6308,\"category\":\"analysis\",\"complexity\":0.7214,\"consequence\":0.6478,\"evidence_need\":0.397,\"family\":\"performance_tuning\",\"input_tokens\":7560,\"reversibility\":0.7151,\"risk\":0.4137,\"summary\":\"Review a proposed optimization for database failover for hidden regressions.\"},\"workers\":[{\"credits_per_call\":14.654,\"id\":\"worker_a\",\"p95_latency_ms\":72447,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.693,\"id\":\"worker_b\",\"p95_latency_ms\":79920,\"pass_rate\":0.8811,\"provider_group\":\"provider_y\",\"quality\":0.8704,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.072,\"id\":\"worker_c\",\"p95_latency_ms\":28718,\"pass_rate\":0.4767,\"provider_group\":\"provider_x\",\"quality\":0.6906,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.072,\"estimated_latency_ms\":28718,\"estimated_pass_rate\":0.3137,\"estimated_quality\":0.568,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.795,\"max_latency_ms\":46734,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6308, "category": "analysis", "complexity": 0.7214, "consequence": 0.6478, "evidence_need": 0.397, "family": "performance_tuning", "input_tokens": 7560, "request_id": "synthetic-performance_tuning-000839", "reversibility": 0.7151, "risk": 0.4137, "summary": "Review a proposed optimization for database failover for hidden regressions."}, "workers": [{"credits_per_call": 14.654, "id": "worker_a", "p95_latency_ms": 72447, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.693, "id": "worker_b", "p95_latency_ms": 79920, "pass_rate": 0.8811, "provider_group": "provider_y", "quality": 0.8704, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.072, "id": "worker_c", "p95_latency_ms": 28718, "pass_rate": 0.4767, "provider_group": "provider_x", "quality": 0.6906, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.848, "max_latency_ms": 184672, "minimum_quality": 0.8928}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000566", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.207, "estimated_latency_ms": 175189, "estimated_pass_rate": 0.7189, "estimated_quality": 0.9139, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.848, "max_latency_ms": 184672, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7189, "quality": 0.9139, "quality_target_met": true, "utility": 68.795081}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.848,\"max_latency_ms\":184672,\"minimum_quality\":0.8928},\"task\":{\"ambiguity\":0.5551,\"category\":\"analysis\",\"complexity\":0.813,\"consequence\":0.6337,\"evidence_need\":0.6374,\"family\":\"performance_tuning\",\"input_tokens\":4720,\"reversibility\":0.8195,\"risk\":0.2375,\"summary\":\"Review a proposed optimization for background jobs for hidden regressions.\"},\"workers\":[{\"credits_per_call\":12.09,\"id\":\"worker_a\",\"p95_latency_ms\":77515,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.883,\"id\":\"worker_b\",\"p95_latency_ms\":24562,\"pass_rate\":0.4988,\"provider_group\":\"provider_z\",\"quality\":0.6992,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.234,\"id\":\"worker_c\",\"p95_latency_ms\":73112,\"pass_rate\":0.8165,\"provider_group\":\"provider_y\",\"quality\":0.8693,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.207,\"estimated_latency_ms\":175189,\"estimated_pass_rate\":0.7189,\"estimated_quality\":0.9139,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.848,\"max_latency_ms\":184672,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5551, "category": "analysis", "complexity": 0.813, "consequence": 0.6337, "evidence_need": 0.6374, "family": "performance_tuning", "input_tokens": 4720, "request_id": "synthetic-performance_tuning-000566", "reversibility": 0.8195, "risk": 0.2375, "summary": "Review a proposed optimization for background jobs for hidden regressions."}, "workers": [{"credits_per_call": 12.09, "id": "worker_a", "p95_latency_ms": 77515, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.883, "id": "worker_b", "p95_latency_ms": 24562, "pass_rate": 0.4988, "provider_group": "provider_z", "quality": 0.6992, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.234, "id": "worker_c", "p95_latency_ms": 73112, "pass_rate": 0.8165, "provider_group": "provider_y", "quality": 0.8693, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.564, "max_latency_ms": 43046, "minimum_quality": 0.6167}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000261", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.063, "estimated_latency_ms": 31966, "estimated_pass_rate": 0.3713, "estimated_quality": 0.645, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.564, "max_latency_ms": 43046, "policy": "lean", "reason_codes": ["high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3713, "quality": 0.645, "quality_target_met": true, "utility": 25.029441}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.564,\"max_latency_ms\":43046,\"minimum_quality\":0.6167},\"task\":{\"ambiguity\":0.7194,\"category\":\"review\",\"complexity\":0.6115,\"consequence\":0.7554,\"evidence_need\":0.9336,\"family\":\"policy_compliance\",\"input_tokens\":7440,\"reversibility\":0.3064,\"risk\":0.841,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":3.345,\"id\":\"worker_a\",\"p95_latency_ms\":69852,\"pass_rate\":0.7581,\"provider_group\":\"provider_x\",\"quality\":0.8244,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.063,\"id\":\"worker_b\",\"p95_latency_ms\":31966,\"pass_rate\":0.5632,\"provider_group\":\"provider_y\",\"quality\":0.7845,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.433,\"id\":\"worker_c\",\"p95_latency_ms\":72865,\"pass_rate\":0.9771,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.063,\"estimated_latency_ms\":31966,\"estimated_pass_rate\":0.3713,\"estimated_quality\":0.645,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.564,\"max_latency_ms\":43046,\"policy\":\"lean\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7194, "category": "review", "complexity": 0.6115, "consequence": 0.7554, "evidence_need": 0.9336, "family": "policy_compliance", "input_tokens": 7440, "request_id": "synthetic-policy_compliance-000261", "reversibility": 0.3064, "risk": 0.841, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 3.345, "id": "worker_a", "p95_latency_ms": 69852, "pass_rate": 0.7581, "provider_group": "provider_x", "quality": 0.8244, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.063, "id": "worker_b", "p95_latency_ms": 31966, "pass_rate": 0.5632, "provider_group": "provider_y", "quality": 0.7845, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.433, "id": "worker_c", "p95_latency_ms": 72865, "pass_rate": 0.9771, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.346, "max_latency_ms": 153848, "minimum_quality": 0.6989}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000816", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.989, "estimated_latency_ms": 119805, "estimated_pass_rate": 0.8529, "estimated_quality": 0.9228, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.346, "max_latency_ms": 153848, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8529, "quality": 0.9228, "quality_target_met": true, "utility": 69.652155}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.346,\"max_latency_ms\":153848,\"minimum_quality\":0.6989},\"task\":{\"ambiguity\":0.5489,\"category\":\"review\",\"complexity\":0.7731,\"consequence\":0.6603,\"evidence_need\":0.8599,\"family\":\"policy_compliance\",\"input_tokens\":2760,\"reversibility\":0.4128,\"risk\":0.6266,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.839,\"id\":\"worker_a\",\"p95_latency_ms\":28869,\"pass_rate\":0.6062,\"provider_group\":\"provider_x\",\"quality\":0.7422,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.915,\"id\":\"worker_b\",\"p95_latency_ms\":69323,\"pass_rate\":0.9374,\"provider_group\":\"provider_z\",\"quality\":0.9821,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.074,\"id\":\"worker_c\",\"p95_latency_ms\":50482,\"pass_rate\":0.8012,\"provider_group\":\"provider_y\",\"quality\":0.868,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.989,\"estimated_latency_ms\":119805,\"estimated_pass_rate\":0.8529,\"estimated_quality\":0.9228,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.346,\"max_latency_ms\":153848,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5489, "category": "review", "complexity": 0.7731, "consequence": 0.6603, "evidence_need": 0.8599, "family": "policy_compliance", "input_tokens": 2760, "request_id": "synthetic-policy_compliance-000816", "reversibility": 0.4128, "risk": 0.6266, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.839, "id": "worker_a", "p95_latency_ms": 28869, "pass_rate": 0.6062, "provider_group": "provider_x", "quality": 0.7422, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.915, "id": "worker_b", "p95_latency_ms": 69323, "pass_rate": 0.9374, "provider_group": "provider_z", "quality": 0.9821, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.074, "id": "worker_c", "p95_latency_ms": 50482, "pass_rate": 0.8012, "provider_group": "provider_y", "quality": 0.868, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.813, "max_latency_ms": 37442, "minimum_quality": 0.7419}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000518", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.921, "estimated_latency_ms": 29198, "estimated_pass_rate": 0.2466, "estimated_quality": 0.5962, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.813, "max_latency_ms": 37442, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2466, "quality": 0.5962, "quality_target_met": false, "utility": 31.316818}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.813,\"max_latency_ms\":37442,\"minimum_quality\":0.7419},\"task\":{\"ambiguity\":0.5709,\"category\":\"analysis\",\"complexity\":0.6572,\"consequence\":0.5811,\"evidence_need\":0.3885,\"family\":\"performance_tuning\",\"input_tokens\":760,\"reversibility\":0.8511,\"risk\":0.5934,\"summary\":\"Review a proposed optimization for access control for hidden regressions.\"},\"workers\":[{\"credits_per_call\":2.791,\"id\":\"worker_a\",\"p95_latency_ms\":61166,\"pass_rate\":0.8098,\"provider_group\":\"provider_x\",\"quality\":0.885,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.905,\"id\":\"worker_b\",\"p95_latency_ms\":80968,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.921,\"id\":\"worker_c\",\"p95_latency_ms\":29198,\"pass_rate\":0.4124,\"provider_group\":\"provider_y\",\"quality\":0.7161,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.921,\"estimated_latency_ms\":29198,\"estimated_pass_rate\":0.2466,\"estimated_quality\":0.5962,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.813,\"max_latency_ms\":37442,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5709, "category": "analysis", "complexity": 0.6572, "consequence": 0.5811, "evidence_need": 0.3885, "family": "performance_tuning", "input_tokens": 760, "request_id": "synthetic-performance_tuning-000518", "reversibility": 0.8511, "risk": 0.5934, "summary": "Review a proposed optimization for access control for hidden regressions."}, "workers": [{"credits_per_call": 2.791, "id": "worker_a", "p95_latency_ms": 61166, "pass_rate": 0.8098, "provider_group": "provider_x", "quality": 0.885, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.905, "id": "worker_b", "p95_latency_ms": 80968, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.921, "id": "worker_c", "p95_latency_ms": 29198, "pass_rate": 0.4124, "provider_group": "provider_y", "quality": 0.7161, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 7.142, "max_latency_ms": 224165, "minimum_quality": 0.744}, "family": "migration_planning", "id": "synthetic-migration_planning-000991", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.711, "estimated_latency_ms": 112781, "estimated_pass_rate": 0.6793, "estimated_quality": 0.7613, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.142, "max_latency_ms": 224165, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6793, "quality": 0.7613, "quality_target_met": true, "utility": 54.500954}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":7.142,\"max_latency_ms\":224165,\"minimum_quality\":0.744},\"task\":{\"ambiguity\":0.6907,\"category\":\"coding\",\"complexity\":0.8009,\"consequence\":0.6373,\"evidence_need\":0.4809,\"family\":\"migration_planning\",\"input_tokens\":3720,\"reversibility\":0.5601,\"risk\":0.4364,\"summary\":\"Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.806,\"id\":\"worker_a\",\"p95_latency_ms\":80415,\"pass_rate\":0.7688,\"provider_group\":\"provider_y\",\"quality\":0.8326,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.521,\"id\":\"worker_b\",\"p95_latency_ms\":64636,\"pass_rate\":0.9865,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.905,\"id\":\"worker_c\",\"p95_latency_ms\":32366,\"pass_rate\":0.6074,\"provider_group\":\"provider_x\",\"quality\":0.7359,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.711,\"estimated_latency_ms\":112781,\"estimated_pass_rate\":0.6793,\"estimated_quality\":0.7613,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.142,\"max_latency_ms\":224165,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6907, "category": "coding", "complexity": 0.8009, "consequence": 0.6373, "evidence_need": 0.4809, "family": "migration_planning", "input_tokens": 3720, "request_id": "synthetic-migration_planning-000991", "reversibility": 0.5601, "risk": 0.4364, "summary": "Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints."}, "workers": [{"credits_per_call": 3.806, "id": "worker_a", "p95_latency_ms": 80415, "pass_rate": 0.7688, "provider_group": "provider_y", "quality": 0.8326, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.521, "id": "worker_b", "p95_latency_ms": 64636, "pass_rate": 0.9865, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.905, "id": "worker_c", "p95_latency_ms": 32366, "pass_rate": 0.6074, "provider_group": "provider_x", "quality": 0.7359, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.14, "max_latency_ms": 193279, "minimum_quality": 0.9404}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000134", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.497, "estimated_latency_ms": 151134, "estimated_pass_rate": 0.8195, "estimated_quality": 0.9197, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.14, "max_latency_ms": 193279, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8195, "quality": 0.9197, "quality_target_met": false, "utility": 71.972376}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.14,\"max_latency_ms\":193279,\"minimum_quality\":0.9404},\"task\":{\"ambiguity\":0.5547,\"category\":\"analysis\",\"complexity\":0.8691,\"consequence\":0.5858,\"evidence_need\":0.3771,\"family\":\"performance_tuning\",\"input_tokens\":4360,\"reversibility\":0.6647,\"risk\":0.5255,\"summary\":\"Review a proposed optimization for usage metering for hidden regressions.\"},\"workers\":[{\"credits_per_call\":2.982,\"id\":\"worker_a\",\"p95_latency_ms\":68104,\"pass_rate\":0.8522,\"provider_group\":\"provider_x\",\"quality\":0.9059,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.783,\"id\":\"worker_b\",\"p95_latency_ms\":28846,\"pass_rate\":0.5268,\"provider_group\":\"provider_y\",\"quality\":0.72,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":12.732,\"id\":\"worker_c\",\"p95_latency_ms\":54184,\"pass_rate\":0.9634,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.497,\"estimated_latency_ms\":151134,\"estimated_pass_rate\":0.8195,\"estimated_quality\":0.9197,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.14,\"max_latency_ms\":193279,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5547, "category": "analysis", "complexity": 0.8691, "consequence": 0.5858, "evidence_need": 0.3771, "family": "performance_tuning", "input_tokens": 4360, "request_id": "synthetic-performance_tuning-000134", "reversibility": 0.6647, "risk": 0.5255, "summary": "Review a proposed optimization for usage metering for hidden regressions."}, "workers": [{"credits_per_call": 2.982, "id": "worker_a", "p95_latency_ms": 68104, "pass_rate": 0.8522, "provider_group": "provider_x", "quality": 0.9059, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.783, "id": "worker_b", "p95_latency_ms": 28846, "pass_rate": 0.5268, "provider_group": "provider_y", "quality": 0.72, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 12.732, "id": "worker_c", "p95_latency_ms": 54184, "pass_rate": 0.9634, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.089, "max_latency_ms": 193587, "minimum_quality": 0.6936}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000705", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.656, "estimated_latency_ms": 162719, "estimated_pass_rate": 0.9109, "estimated_quality": 0.916, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.089, "max_latency_ms": 193587, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9109, "quality": 0.916, "quality_target_met": true, "utility": 71.792068}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.089,\"max_latency_ms\":193587,\"minimum_quality\":0.6936},\"task\":{\"ambiguity\":0.3715,\"category\":\"review\",\"complexity\":0.4049,\"consequence\":0.6329,\"evidence_need\":0.7946,\"family\":\"policy_compliance\",\"input_tokens\":6800,\"reversibility\":0.5319,\"risk\":0.7146,\"summary\":\"Identify evidence required before approving model routing for production.\"},\"workers\":[{\"credits_per_call\":0.824,\"id\":\"worker_a\",\"p95_latency_ms\":32006,\"pass_rate\":0.5524,\"provider_group\":\"provider_z\",\"quality\":0.7681,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.195,\"id\":\"worker_b\",\"p95_latency_ms\":79511,\"pass_rate\":0.7574,\"provider_group\":\"provider_x\",\"quality\":0.8015,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.461,\"id\":\"worker_c\",\"p95_latency_ms\":83208,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9219,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.656,\"estimated_latency_ms\":162719,\"estimated_pass_rate\":0.9109,\"estimated_quality\":0.916,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.089,\"max_latency_ms\":193587,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3715, "category": "review", "complexity": 0.4049, "consequence": 0.6329, "evidence_need": 0.7946, "family": "policy_compliance", "input_tokens": 6800, "request_id": "synthetic-policy_compliance-000705", "reversibility": 0.5319, "risk": 0.7146, "summary": "Identify evidence required before approving model routing for production."}, "workers": [{"credits_per_call": 0.824, "id": "worker_a", "p95_latency_ms": 32006, "pass_rate": 0.5524, "provider_group": "provider_z", "quality": 0.7681, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.195, "id": "worker_b", "p95_latency_ms": 79511, "pass_rate": 0.7574, "provider_group": "provider_x", "quality": 0.8015, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.461, "id": "worker_c", "p95_latency_ms": 83208, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9219, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.225, "max_latency_ms": 163271, "minimum_quality": 0.7706}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000998", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.351, "estimated_latency_ms": 147092, "estimated_pass_rate": 0.9104, "estimated_quality": 0.9415, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.225, "max_latency_ms": 163271, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9104, "quality": 0.9415, "quality_target_met": true, "utility": 69.132951}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.225,\"max_latency_ms\":163271,\"minimum_quality\":0.7706},\"task\":{\"ambiguity\":0.5394,\"category\":\"analysis\",\"complexity\":0.6098,\"consequence\":0.4292,\"evidence_need\":0.5628,\"family\":\"performance_tuning\",\"input_tokens\":6120,\"reversibility\":0.5859,\"risk\":0.4091,\"summary\":\"Choose between throughput and tail-latency changes for model routing.\"},\"workers\":[{\"credits_per_call\":4.202,\"id\":\"worker_a\",\"p95_latency_ms\":80029,\"pass_rate\":0.826,\"provider_group\":\"provider_z\",\"quality\":0.9813,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.995,\"id\":\"worker_b\",\"p95_latency_ms\":31911,\"pass_rate\":0.4711,\"provider_group\":\"provider_x\",\"quality\":0.6975,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.149,\"id\":\"worker_c\",\"p95_latency_ms\":67063,\"pass_rate\":0.9928,\"provider_group\":\"provider_y\",\"quality\":0.9808,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.351,\"estimated_latency_ms\":147092,\"estimated_pass_rate\":0.9104,\"estimated_quality\":0.9415,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.225,\"max_latency_ms\":163271,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5394, "category": "analysis", "complexity": 0.6098, "consequence": 0.4292, "evidence_need": 0.5628, "family": "performance_tuning", "input_tokens": 6120, "request_id": "synthetic-performance_tuning-000998", "reversibility": 0.5859, "risk": 0.4091, "summary": "Choose between throughput and tail-latency changes for model routing."}, "workers": [{"credits_per_call": 4.202, "id": "worker_a", "p95_latency_ms": 80029, "pass_rate": 0.826, "provider_group": "provider_z", "quality": 0.9813, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.995, "id": "worker_b", "p95_latency_ms": 31911, "pass_rate": 0.4711, "provider_group": "provider_x", "quality": 0.6975, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.149, "id": "worker_c", "p95_latency_ms": 67063, "pass_rate": 0.9928, "provider_group": "provider_y", "quality": 0.9808, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.304, "max_latency_ms": 35660, "minimum_quality": 0.7505}, "family": "migration_planning", "id": "synthetic-migration_planning-000427", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.061, "estimated_latency_ms": 22658, "estimated_pass_rate": 0.3192, "estimated_quality": 0.6763, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.304, "max_latency_ms": 35660, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3192, "quality": 0.6763, "quality_target_met": false, "utility": 40.073022}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.304,\"max_latency_ms\":35660,\"minimum_quality\":0.7505},\"task\":{\"ambiguity\":0.5899,\"category\":\"coding\",\"complexity\":0.6312,\"consequence\":0.6757,\"evidence_need\":0.4367,\"family\":\"migration_planning\",\"input_tokens\":1640,\"reversibility\":0.5554,\"risk\":0.3576,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":4.019,\"id\":\"worker_a\",\"p95_latency_ms\":44545,\"pass_rate\":0.8251,\"provider_group\":\"provider_x\",\"quality\":0.8093,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.061,\"id\":\"worker_b\",\"p95_latency_ms\":22658,\"pass_rate\":0.4643,\"provider_group\":\"provider_y\",\"quality\":0.7758,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.052,\"id\":\"worker_c\",\"p95_latency_ms\":61206,\"pass_rate\":0.9601,\"provider_group\":\"provider_z\",\"quality\":0.9124,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.061,\"estimated_latency_ms\":22658,\"estimated_pass_rate\":0.3192,\"estimated_quality\":0.6763,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.304,\"max_latency_ms\":35660,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5899, "category": "coding", "complexity": 0.6312, "consequence": 0.6757, "evidence_need": 0.4367, "family": "migration_planning", "input_tokens": 1640, "request_id": "synthetic-migration_planning-000427", "reversibility": 0.5554, "risk": 0.3576, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 4.019, "id": "worker_a", "p95_latency_ms": 44545, "pass_rate": 0.8251, "provider_group": "provider_x", "quality": 0.8093, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.061, "id": "worker_b", "p95_latency_ms": 22658, "pass_rate": 0.4643, "provider_group": "provider_y", "quality": 0.7758, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.052, "id": "worker_c", "p95_latency_ms": 61206, "pass_rate": 0.9601, "provider_group": "provider_z", "quality": 0.9124, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.672, "max_latency_ms": 54216, "minimum_quality": 0.5577}, "family": "migration_planning", "id": "synthetic-migration_planning-000739", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 3.286, "estimated_latency_ms": 53048, "estimated_pass_rate": 0.595, "estimated_quality": 0.7735, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.672, "max_latency_ms": 54216, "policy": "lean", "reason_codes": ["complex_task", "credit_constrained", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 2, "pass_rate": 0.595, "quality": 0.7735, "quality_target_met": true, "utility": 50.974784}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.672,\"max_latency_ms\":54216,\"minimum_quality\":0.5577},\"task\":{\"ambiguity\":0.7028,\"category\":\"coding\",\"complexity\":0.7469,\"consequence\":0.4921,\"evidence_need\":0.4053,\"family\":\"migration_planning\",\"input_tokens\":1000,\"reversibility\":0.3079,\"risk\":0.5379,\"summary\":\"Design compatibility testing for a data retention schema migration.\"},\"workers\":[{\"credits_per_call\":0.959,\"id\":\"worker_a\",\"p95_latency_ms\":27815,\"pass_rate\":0.4741,\"provider_group\":\"provider_z\",\"quality\":0.7935,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.055,\"id\":\"worker_b\",\"p95_latency_ms\":80928,\"pass_rate\":0.9084,\"provider_group\":\"provider_y\",\"quality\":0.9748,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.286,\"id\":\"worker_c\",\"p95_latency_ms\":53048,\"pass_rate\":0.7764,\"provider_group\":\"provider_x\",\"quality\":0.9145,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":3.286,\"estimated_latency_ms\":53048,\"estimated_pass_rate\":0.595,\"estimated_quality\":0.7735,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.672,\"max_latency_ms\":54216,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7028, "category": "coding", "complexity": 0.7469, "consequence": 0.4921, "evidence_need": 0.4053, "family": "migration_planning", "input_tokens": 1000, "request_id": "synthetic-migration_planning-000739", "reversibility": 0.3079, "risk": 0.5379, "summary": "Design compatibility testing for a data retention schema migration."}, "workers": [{"credits_per_call": 0.959, "id": "worker_a", "p95_latency_ms": 27815, "pass_rate": 0.4741, "provider_group": "provider_z", "quality": 0.7935, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.055, "id": "worker_b", "p95_latency_ms": 80928, "pass_rate": 0.9084, "provider_group": "provider_y", "quality": 0.9748, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.286, "id": "worker_c", "p95_latency_ms": 53048, "pass_rate": 0.7764, "provider_group": "provider_x", "quality": 0.9145, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.393, "max_latency_ms": 35778, "minimum_quality": 0.6884}, "family": "migration_planning", "id": "synthetic-migration_planning-000022", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.822, "estimated_latency_ms": 23807, "estimated_pass_rate": 0.3135, "estimated_quality": 0.5245, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.393, "max_latency_ms": 35778, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3135, "quality": 0.5245, "quality_target_met": false, "utility": 20.655516}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.393,\"max_latency_ms\":35778,\"minimum_quality\":0.6884},\"task\":{\"ambiguity\":0.6189,\"category\":\"coding\",\"complexity\":0.9165,\"consequence\":0.737,\"evidence_need\":0.616,\"family\":\"migration_planning\",\"input_tokens\":1560,\"reversibility\":0.267,\"risk\":0.6221,\"summary\":\"Review a staged migration for data retention and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.822,\"id\":\"worker_a\",\"p95_latency_ms\":23807,\"pass_rate\":0.5146,\"provider_group\":\"provider_z\",\"quality\":0.6966,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.358,\"id\":\"worker_b\",\"p95_latency_ms\":81553,\"pass_rate\":0.9213,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.796,\"id\":\"worker_c\",\"p95_latency_ms\":48427,\"pass_rate\":0.7853,\"provider_group\":\"provider_x\",\"quality\":0.8157,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.822,\"estimated_latency_ms\":23807,\"estimated_pass_rate\":0.3135,\"estimated_quality\":0.5245,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.393,\"max_latency_ms\":35778,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6189, "category": "coding", "complexity": 0.9165, "consequence": 0.737, "evidence_need": 0.616, "family": "migration_planning", "input_tokens": 1560, "request_id": "synthetic-migration_planning-000022", "reversibility": 0.267, "risk": 0.6221, "summary": "Review a staged migration for data retention and identify irreversible steps."}, "workers": [{"credits_per_call": 0.822, "id": "worker_a", "p95_latency_ms": 23807, "pass_rate": 0.5146, "provider_group": "provider_z", "quality": 0.6966, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.358, "id": "worker_b", "p95_latency_ms": 81553, "pass_rate": 0.9213, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.796, "id": "worker_c", "p95_latency_ms": 48427, "pass_rate": 0.7853, "provider_group": "provider_x", "quality": 0.8157, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 12.556, "max_latency_ms": 173664, "minimum_quality": 0.7545}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001149", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.061, "estimated_latency_ms": 94567, "estimated_pass_rate": 0.6454, "estimated_quality": 0.7933, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.556, "max_latency_ms": 173664, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6454, "quality": 0.7933, "quality_target_met": true, "utility": 62.609547}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":12.556,\"max_latency_ms\":173664,\"minimum_quality\":0.7545},\"task\":{\"ambiguity\":0.4969,\"category\":\"review\",\"complexity\":0.6829,\"consequence\":0.7943,\"evidence_need\":0.6068,\"family\":\"policy_compliance\",\"input_tokens\":600,\"reversibility\":0.3461,\"risk\":0.6823,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.87,\"id\":\"worker_a\",\"p95_latency_ms\":32055,\"pass_rate\":0.555,\"provider_group\":\"provider_z\",\"quality\":0.7244,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.527,\"id\":\"worker_b\",\"p95_latency_ms\":87095,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9264,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.191,\"id\":\"worker_c\",\"p95_latency_ms\":62512,\"pass_rate\":0.8825,\"provider_group\":\"provider_x\",\"quality\":0.8126,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.061,\"estimated_latency_ms\":94567,\"estimated_pass_rate\":0.6454,\"estimated_quality\":0.7933,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.556,\"max_latency_ms\":173664,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4969, "category": "review", "complexity": 0.6829, "consequence": 0.7943, "evidence_need": 0.6068, "family": "policy_compliance", "input_tokens": 600, "request_id": "synthetic-policy_compliance-001149", "reversibility": 0.3461, "risk": 0.6823, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.87, "id": "worker_a", "p95_latency_ms": 32055, "pass_rate": 0.555, "provider_group": "provider_z", "quality": 0.7244, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.527, "id": "worker_b", "p95_latency_ms": 87095, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9264, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.191, "id": "worker_c", "p95_latency_ms": 62512, "pass_rate": 0.8825, "provider_group": "provider_x", "quality": 0.8126, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 17.764, "max_latency_ms": 214992, "minimum_quality": 0.7257}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000870", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.199, "estimated_latency_ms": 133016, "estimated_pass_rate": 0.8335, "estimated_quality": 0.9206, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.764, "max_latency_ms": 214992, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8335, "quality": 0.9206, "quality_target_met": true, "utility": 70.977885}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":17.764,\"max_latency_ms\":214992,\"minimum_quality\":0.7257},\"task\":{\"ambiguity\":0.5734,\"category\":\"review\",\"complexity\":0.801,\"consequence\":0.6765,\"evidence_need\":0.7286,\"family\":\"policy_compliance\",\"input_tokens\":4840,\"reversibility\":0.5721,\"risk\":0.7574,\"summary\":\"Identify evidence required before approving background jobs for production.\"},\"workers\":[{\"credits_per_call\":0.945,\"id\":\"worker_a\",\"p95_latency_ms\":26396,\"pass_rate\":0.5473,\"provider_group\":\"provider_z\",\"quality\":0.7434,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.793,\"id\":\"worker_b\",\"p95_latency_ms\":55231,\"pass_rate\":0.9007,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.406,\"id\":\"worker_c\",\"p95_latency_ms\":77785,\"pass_rate\":0.8432,\"provider_group\":\"provider_x\",\"quality\":0.8431,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.199,\"estimated_latency_ms\":133016,\"estimated_pass_rate\":0.8335,\"estimated_quality\":0.9206,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.764,\"max_latency_ms\":214992,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5734, "category": "review", "complexity": 0.801, "consequence": 0.6765, "evidence_need": 0.7286, "family": "policy_compliance", "input_tokens": 4840, "request_id": "synthetic-policy_compliance-000870", "reversibility": 0.5721, "risk": 0.7574, "summary": "Identify evidence required before approving background jobs for production."}, "workers": [{"credits_per_call": 0.945, "id": "worker_a", "p95_latency_ms": 26396, "pass_rate": 0.5473, "provider_group": "provider_z", "quality": 0.7434, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.793, "id": "worker_b", "p95_latency_ms": 55231, "pass_rate": 0.9007, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.406, "id": "worker_c", "p95_latency_ms": 77785, "pass_rate": 0.8432, "provider_group": "provider_x", "quality": 0.8431, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.302, "max_latency_ms": 34334, "minimum_quality": 0.4984}, "family": "migration_planning", "id": "synthetic-migration_planning-000484", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.172, "estimated_latency_ms": 24327, "estimated_pass_rate": 0.4239, "estimated_quality": 0.6355, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.302, "max_latency_ms": 34334, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4239, "quality": 0.6355, "quality_target_met": true, "utility": 42.398414}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.302,\"max_latency_ms\":34334,\"minimum_quality\":0.4984},\"task\":{\"ambiguity\":0.6831,\"category\":\"coding\",\"complexity\":0.8073,\"consequence\":0.6263,\"evidence_need\":0.7366,\"family\":\"migration_planning\",\"input_tokens\":1880,\"reversibility\":0.5054,\"risk\":0.4432,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":12.575,\"id\":\"worker_a\",\"p95_latency_ms\":59813,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9808,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.742,\"id\":\"worker_b\",\"p95_latency_ms\":48264,\"pass_rate\":0.7592,\"provider_group\":\"provider_x\",\"quality\":0.8157,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.172,\"id\":\"worker_c\",\"p95_latency_ms\":24327,\"pass_rate\":0.603,\"provider_group\":\"provider_y\",\"quality\":0.7793,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.172,\"estimated_latency_ms\":24327,\"estimated_pass_rate\":0.4239,\"estimated_quality\":0.6355,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.302,\"max_latency_ms\":34334,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6831, "category": "coding", "complexity": 0.8073, "consequence": 0.6263, "evidence_need": 0.7366, "family": "migration_planning", "input_tokens": 1880, "request_id": "synthetic-migration_planning-000484", "reversibility": 0.5054, "risk": 0.4432, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 12.575, "id": "worker_a", "p95_latency_ms": 59813, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9808, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.742, "id": "worker_b", "p95_latency_ms": 48264, "pass_rate": 0.7592, "provider_group": "provider_x", "quality": 0.8157, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.172, "id": "worker_c", "p95_latency_ms": 24327, "pass_rate": 0.603, "provider_group": "provider_y", "quality": 0.7793, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.732, "max_latency_ms": 33717, "minimum_quality": 0.7035}, "family": "migration_planning", "id": "synthetic-migration_planning-000391", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.908, "estimated_latency_ms": 23846, "estimated_pass_rate": 0.2931, "estimated_quality": 0.6208, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.732, "max_latency_ms": 33717, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2931, "quality": 0.6208, "quality_target_met": false, "utility": 36.871813}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.732,\"max_latency_ms\":33717,\"minimum_quality\":0.7035},\"task\":{\"ambiguity\":0.634,\"category\":\"coding\",\"complexity\":0.7123,\"consequence\":0.6379,\"evidence_need\":0.5612,\"family\":\"migration_planning\",\"input_tokens\":4520,\"reversibility\":0.4797,\"risk\":0.5613,\"summary\":\"Review a staged migration for prompt caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.524,\"id\":\"worker_a\",\"p95_latency_ms\":92239,\"pass_rate\":0.9432,\"provider_group\":\"provider_z\",\"quality\":0.9306,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.791,\"id\":\"worker_b\",\"p95_latency_ms\":55778,\"pass_rate\":0.8614,\"provider_group\":\"provider_y\",\"quality\":0.8577,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.908,\"id\":\"worker_c\",\"p95_latency_ms\":23846,\"pass_rate\":0.4673,\"provider_group\":\"provider_x\",\"quality\":0.7526,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.908,\"estimated_latency_ms\":23846,\"estimated_pass_rate\":0.2931,\"estimated_quality\":0.6208,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.732,\"max_latency_ms\":33717,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.634, "category": "coding", "complexity": 0.7123, "consequence": 0.6379, "evidence_need": 0.5612, "family": "migration_planning", "input_tokens": 4520, "request_id": "synthetic-migration_planning-000391", "reversibility": 0.4797, "risk": 0.5613, "summary": "Review a staged migration for prompt caching and identify irreversible steps."}, "workers": [{"credits_per_call": 14.524, "id": "worker_a", "p95_latency_ms": 92239, "pass_rate": 0.9432, "provider_group": "provider_z", "quality": 0.9306, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.791, "id": "worker_b", "p95_latency_ms": 55778, "pass_rate": 0.8614, "provider_group": "provider_y", "quality": 0.8577, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.908, "id": "worker_c", "p95_latency_ms": 23846, "pass_rate": 0.4673, "provider_group": "provider_x", "quality": 0.7526, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.125, "max_latency_ms": 152135, "minimum_quality": 0.8796}, "family": "migration_planning", "id": "synthetic-migration_planning-000976", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.008, "estimated_latency_ms": 146196, "estimated_pass_rate": 0.7619, "estimated_quality": 0.9007, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 20.125, "max_latency_ms": 152135, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7619, "quality": 0.9007, "quality_target_met": true, "utility": 71.694804}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.125,\"max_latency_ms\":152135,\"minimum_quality\":0.8796},\"task\":{\"ambiguity\":0.3904,\"category\":\"coding\",\"complexity\":0.732,\"consequence\":0.6356,\"evidence_need\":0.6417,\"family\":\"migration_planning\",\"input_tokens\":560,\"reversibility\":0.6207,\"risk\":0.6024,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":12.817,\"id\":\"worker_a\",\"p95_latency_ms\":52239,\"pass_rate\":0.9288,\"provider_group\":\"provider_z\",\"quality\":0.9789,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.966,\"id\":\"worker_b\",\"p95_latency_ms\":29869,\"pass_rate\":0.586,\"provider_group\":\"provider_x\",\"quality\":0.6802,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.225,\"id\":\"worker_c\",\"p95_latency_ms\":64088,\"pass_rate\":0.8348,\"provider_group\":\"provider_y\",\"quality\":0.8677,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.008,\"estimated_latency_ms\":146196,\"estimated_pass_rate\":0.7619,\"estimated_quality\":0.9007,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.125,\"max_latency_ms\":152135,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3904, "category": "coding", "complexity": 0.732, "consequence": 0.6356, "evidence_need": 0.6417, "family": "migration_planning", "input_tokens": 560, "request_id": "synthetic-migration_planning-000976", "reversibility": 0.6207, "risk": 0.6024, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 12.817, "id": "worker_a", "p95_latency_ms": 52239, "pass_rate": 0.9288, "provider_group": "provider_z", "quality": 0.9789, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.966, "id": "worker_b", "p95_latency_ms": 29869, "pass_rate": 0.586, "provider_group": "provider_x", "quality": 0.6802, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.225, "id": "worker_c", "p95_latency_ms": 64088, "pass_rate": 0.8348, "provider_group": "provider_y", "quality": 0.8677, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.546, "max_latency_ms": 199776, "minimum_quality": 0.8237}, "family": "migration_planning", "id": "synthetic-migration_planning-000463", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.715, "estimated_latency_ms": 156390, "estimated_pass_rate": 0.7368, "estimated_quality": 0.8926, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 22.546, "max_latency_ms": 199776, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7368, "quality": 0.8926, "quality_target_met": true, "utility": 73.926843}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.546,\"max_latency_ms\":199776,\"minimum_quality\":0.8237},\"task\":{\"ambiguity\":0.5416,\"category\":\"coding\",\"complexity\":0.7842,\"consequence\":0.7482,\"evidence_need\":0.5285,\"family\":\"migration_planning\",\"input_tokens\":3240,\"reversibility\":0.2844,\"risk\":0.612,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":12.398,\"id\":\"worker_a\",\"p95_latency_ms\":86002,\"pass_rate\":0.9703,\"provider_group\":\"provider_y\",\"quality\":0.9706,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.315,\"id\":\"worker_b\",\"p95_latency_ms\":49530,\"pass_rate\":0.7446,\"provider_group\":\"provider_z\",\"quality\":0.8887,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.002,\"id\":\"worker_c\",\"p95_latency_ms\":20858,\"pass_rate\":0.5294,\"provider_group\":\"provider_x\",\"quality\":0.7151,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.715,\"estimated_latency_ms\":156390,\"estimated_pass_rate\":0.7368,\"estimated_quality\":0.8926,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.546,\"max_latency_ms\":199776,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5416, "category": "coding", "complexity": 0.7842, "consequence": 0.7482, "evidence_need": 0.5285, "family": "migration_planning", "input_tokens": 3240, "request_id": "synthetic-migration_planning-000463", "reversibility": 0.2844, "risk": 0.612, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 12.398, "id": "worker_a", "p95_latency_ms": 86002, "pass_rate": 0.9703, "provider_group": "provider_y", "quality": 0.9706, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.315, "id": "worker_b", "p95_latency_ms": 49530, "pass_rate": 0.7446, "provider_group": "provider_z", "quality": 0.8887, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.002, "id": "worker_c", "p95_latency_ms": 20858, "pass_rate": 0.5294, "provider_group": "provider_x", "quality": 0.7151, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.018, "max_latency_ms": 31532, "minimum_quality": 0.5616}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000339", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.913, "estimated_latency_ms": 22171, "estimated_pass_rate": 0.4157, "estimated_quality": 0.6214, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.018, "max_latency_ms": 31532, "policy": "lean", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4157, "quality": 0.6214, "quality_target_met": true, "utility": 23.064816}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.018,\"max_latency_ms\":31532,\"minimum_quality\":0.5616},\"task\":{\"ambiguity\":0.4121,\"category\":\"review\",\"complexity\":0.6189,\"consequence\":0.9235,\"evidence_need\":0.6967,\"family\":\"policy_compliance\",\"input_tokens\":960,\"reversibility\":0.6122,\"risk\":0.6978,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":0.913,\"id\":\"worker_a\",\"p95_latency_ms\":22171,\"pass_rate\":0.5736,\"provider_group\":\"provider_x\",\"quality\":0.7312,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.049,\"id\":\"worker_b\",\"p95_latency_ms\":74406,\"pass_rate\":0.9264,\"provider_group\":\"provider_y\",\"quality\":0.9127,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.309,\"id\":\"worker_c\",\"p95_latency_ms\":51737,\"pass_rate\":0.853,\"provider_group\":\"provider_z\",\"quality\":0.9154,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.913,\"estimated_latency_ms\":22171,\"estimated_pass_rate\":0.4157,\"estimated_quality\":0.6214,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.018,\"max_latency_ms\":31532,\"policy\":\"lean\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4121, "category": "review", "complexity": 0.6189, "consequence": 0.9235, "evidence_need": 0.6967, "family": "policy_compliance", "input_tokens": 960, "request_id": "synthetic-policy_compliance-000339", "reversibility": 0.6122, "risk": 0.6978, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 0.913, "id": "worker_a", "p95_latency_ms": 22171, "pass_rate": 0.5736, "provider_group": "provider_x", "quality": 0.7312, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.049, "id": "worker_b", "p95_latency_ms": 74406, "pass_rate": 0.9264, "provider_group": "provider_y", "quality": 0.9127, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.309, "id": "worker_c", "p95_latency_ms": 51737, "pass_rate": 0.853, "provider_group": "provider_z", "quality": 0.9154, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.362, "max_latency_ms": 100256, "minimum_quality": 0.885}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000018", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.397, "estimated_latency_ms": 85730, "estimated_pass_rate": 0.6915, "estimated_quality": 0.8597, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.362, "max_latency_ms": 100256, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6915, "quality": 0.8597, "quality_target_met": false, "utility": 60.701803}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.362,\"max_latency_ms\":100256,\"minimum_quality\":0.885},\"task\":{\"ambiguity\":0.5138,\"category\":\"review\",\"complexity\":0.7999,\"consequence\":0.5208,\"evidence_need\":0.5828,\"family\":\"policy_compliance\",\"input_tokens\":2200,\"reversibility\":0.3787,\"risk\":0.7004,\"summary\":\"Review whether provider failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.634,\"id\":\"worker_a\",\"p95_latency_ms\":74962,\"pass_rate\":0.802,\"provider_group\":\"provider_z\",\"quality\":0.9218,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.936,\"id\":\"worker_b\",\"p95_latency_ms\":28076,\"pass_rate\":0.4618,\"provider_group\":\"provider_x\",\"quality\":0.7492,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.461,\"id\":\"worker_c\",\"p95_latency_ms\":57654,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9647,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.397,\"estimated_latency_ms\":85730,\"estimated_pass_rate\":0.6915,\"estimated_quality\":0.8597,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.362,\"max_latency_ms\":100256,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5138, "category": "review", "complexity": 0.7999, "consequence": 0.5208, "evidence_need": 0.5828, "family": "policy_compliance", "input_tokens": 2200, "request_id": "synthetic-policy_compliance-000018", "reversibility": 0.3787, "risk": 0.7004, "summary": "Review whether provider failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.634, "id": "worker_a", "p95_latency_ms": 74962, "pass_rate": 0.802, "provider_group": "provider_z", "quality": 0.9218, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.936, "id": "worker_b", "p95_latency_ms": 28076, "pass_rate": 0.4618, "provider_group": "provider_x", "quality": 0.7492, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.461, "id": "worker_c", "p95_latency_ms": 57654, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9647, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.892, "max_latency_ms": 190427, "minimum_quality": 0.8274}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001182", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.774, "estimated_latency_ms": 169868, "estimated_pass_rate": 0.7614, "estimated_quality": 0.9201, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.892, "max_latency_ms": 190427, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7614, "quality": 0.9201, "quality_target_met": true, "utility": 75.974793}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.892,\"max_latency_ms\":190427,\"minimum_quality\":0.8274},\"task\":{\"ambiguity\":0.5819,\"category\":\"review\",\"complexity\":0.7483,\"consequence\":0.6362,\"evidence_need\":0.6359,\"family\":\"policy_compliance\",\"input_tokens\":7000,\"reversibility\":0.4463,\"risk\":0.679,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":12.363,\"id\":\"worker_a\",\"p95_latency_ms\":76919,\"pass_rate\":0.9709,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.228,\"id\":\"worker_b\",\"p95_latency_ms\":59940,\"pass_rate\":0.8629,\"provider_group\":\"provider_y\",\"quality\":0.9011,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.183,\"id\":\"worker_c\",\"p95_latency_ms\":33009,\"pass_rate\":0.5055,\"provider_group\":\"provider_x\",\"quality\":0.7178,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.774,\"estimated_latency_ms\":169868,\"estimated_pass_rate\":0.7614,\"estimated_quality\":0.9201,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.892,\"max_latency_ms\":190427,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5819, "category": "review", "complexity": 0.7483, "consequence": 0.6362, "evidence_need": 0.6359, "family": "policy_compliance", "input_tokens": 7000, "request_id": "synthetic-policy_compliance-001182", "reversibility": 0.4463, "risk": 0.679, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 12.363, "id": "worker_a", "p95_latency_ms": 76919, "pass_rate": 0.9709, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.228, "id": "worker_b", "p95_latency_ms": 59940, "pass_rate": 0.8629, "provider_group": "provider_y", "quality": 0.9011, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.183, "id": "worker_c", "p95_latency_ms": 33009, "pass_rate": 0.5055, "provider_group": "provider_x", "quality": 0.7178, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 7.449, "max_latency_ms": 133082, "minimum_quality": 0.8253}, "family": "migration_planning", "id": "synthetic-migration_planning-000832", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.163, "estimated_latency_ms": 89757, "estimated_pass_rate": 0.5988, "estimated_quality": 0.7231, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.449, "max_latency_ms": 133082, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5988, "quality": 0.7231, "quality_target_met": false, "utility": 49.129705}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":7.449,\"max_latency_ms\":133082,\"minimum_quality\":0.8253},\"task\":{\"ambiguity\":0.7,\"category\":\"coding\",\"complexity\":0.8548,\"consequence\":0.6217,\"evidence_need\":0.4821,\"family\":\"migration_planning\",\"input_tokens\":2240,\"reversibility\":0.3399,\"risk\":0.7348,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.954,\"id\":\"worker_a\",\"p95_latency_ms\":57716,\"pass_rate\":0.8079,\"provider_group\":\"provider_y\",\"quality\":0.816,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.174,\"id\":\"worker_b\",\"p95_latency_ms\":86005,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.209,\"id\":\"worker_c\",\"p95_latency_ms\":32041,\"pass_rate\":0.5398,\"provider_group\":\"provider_x\",\"quality\":0.6969,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.163,\"estimated_latency_ms\":89757,\"estimated_pass_rate\":0.5988,\"estimated_quality\":0.7231,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.449,\"max_latency_ms\":133082,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7, "category": "coding", "complexity": 0.8548, "consequence": 0.6217, "evidence_need": 0.4821, "family": "migration_planning", "input_tokens": 2240, "request_id": "synthetic-migration_planning-000832", "reversibility": 0.3399, "risk": 0.7348, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 3.954, "id": "worker_a", "p95_latency_ms": 57716, "pass_rate": 0.8079, "provider_group": "provider_y", "quality": 0.816, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.174, "id": "worker_b", "p95_latency_ms": 86005, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.209, "id": "worker_c", "p95_latency_ms": 32041, "pass_rate": 0.5398, "provider_group": "provider_x", "quality": 0.6969, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.467, "max_latency_ms": 159595, "minimum_quality": 0.8615}, "family": "migration_planning", "id": "synthetic-migration_planning-000751", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.145, "estimated_latency_ms": 139926, "estimated_pass_rate": 0.7257, "estimated_quality": 0.9179, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.467, "max_latency_ms": 159595, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7257, "quality": 0.9179, "quality_target_met": true, "utility": 69.908899}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.467,\"max_latency_ms\":159595,\"minimum_quality\":0.8615},\"task\":{\"ambiguity\":0.5266,\"category\":\"coding\",\"complexity\":0.7004,\"consequence\":0.421,\"evidence_need\":0.8265,\"family\":\"migration_planning\",\"input_tokens\":1560,\"reversibility\":0.4618,\"risk\":0.3894,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":1.134,\"id\":\"worker_a\",\"p95_latency_ms\":26256,\"pass_rate\":0.5076,\"provider_group\":\"provider_y\",\"quality\":0.7128,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.99,\"id\":\"worker_b\",\"p95_latency_ms\":65117,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.021,\"id\":\"worker_c\",\"p95_latency_ms\":48553,\"pass_rate\":0.7826,\"provider_group\":\"provider_z\",\"quality\":0.8257,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.145,\"estimated_latency_ms\":139926,\"estimated_pass_rate\":0.7257,\"estimated_quality\":0.9179,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.467,\"max_latency_ms\":159595,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5266, "category": "coding", "complexity": 0.7004, "consequence": 0.421, "evidence_need": 0.8265, "family": "migration_planning", "input_tokens": 1560, "request_id": "synthetic-migration_planning-000751", "reversibility": 0.4618, "risk": 0.3894, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 1.134, "id": "worker_a", "p95_latency_ms": 26256, "pass_rate": 0.5076, "provider_group": "provider_y", "quality": 0.7128, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.99, "id": "worker_b", "p95_latency_ms": 65117, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.021, "id": "worker_c", "p95_latency_ms": 48553, "pass_rate": 0.7826, "provider_group": "provider_z", "quality": 0.8257, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.412, "max_latency_ms": 136559, "minimum_quality": 0.6911}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000234", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.437, "estimated_latency_ms": 90910, "estimated_pass_rate": 0.6571, "estimated_quality": 0.8477, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.412, "max_latency_ms": 136559, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6571, "quality": 0.8477, "quality_target_met": true, "utility": 61.221613}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.412,\"max_latency_ms\":136559,\"minimum_quality\":0.6911},\"task\":{\"ambiguity\":0.6193,\"category\":\"review\",\"complexity\":0.7572,\"consequence\":0.5652,\"evidence_need\":0.8215,\"family\":\"policy_compliance\",\"input_tokens\":6280,\"reversibility\":0.6669,\"risk\":0.7424,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.26,\"id\":\"worker_a\",\"p95_latency_ms\":78338,\"pass_rate\":0.783,\"provider_group\":\"provider_x\",\"quality\":0.8625,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.878,\"id\":\"worker_b\",\"p95_latency_ms\":32837,\"pass_rate\":0.4964,\"provider_group\":\"provider_y\",\"quality\":0.7262,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.559,\"id\":\"worker_c\",\"p95_latency_ms\":58073,\"pass_rate\":0.8947,\"provider_group\":\"provider_z\",\"quality\":0.9269,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.437,\"estimated_latency_ms\":90910,\"estimated_pass_rate\":0.6571,\"estimated_quality\":0.8477,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.412,\"max_latency_ms\":136559,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6193, "category": "review", "complexity": 0.7572, "consequence": 0.5652, "evidence_need": 0.8215, "family": "policy_compliance", "input_tokens": 6280, "request_id": "synthetic-policy_compliance-000234", "reversibility": 0.6669, "risk": 0.7424, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 4.26, "id": "worker_a", "p95_latency_ms": 78338, "pass_rate": 0.783, "provider_group": "provider_x", "quality": 0.8625, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.878, "id": "worker_b", "p95_latency_ms": 32837, "pass_rate": 0.4964, "provider_group": "provider_y", "quality": 0.7262, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.559, "id": "worker_c", "p95_latency_ms": 58073, "pass_rate": 0.8947, "provider_group": "provider_z", "quality": 0.9269, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.743, "max_latency_ms": 232117, "minimum_quality": 0.7856}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000918", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.743, "estimated_latency_ms": 167588, "estimated_pass_rate": 0.7698, "estimated_quality": 0.9076, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 16.743, "max_latency_ms": 232117, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7698, "quality": 0.9076, "quality_target_met": true, "utility": 76.382636}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.743,\"max_latency_ms\":232117,\"minimum_quality\":0.7856},\"task\":{\"ambiguity\":0.4977,\"category\":\"review\",\"complexity\":0.6855,\"consequence\":0.823,\"evidence_need\":0.8293,\"family\":\"policy_compliance\",\"input_tokens\":7880,\"reversibility\":0.4653,\"risk\":0.6369,\"summary\":\"Identify evidence required before approving provider failover for production.\"},\"workers\":[{\"credits_per_call\":0.932,\"id\":\"worker_a\",\"p95_latency_ms\":25988,\"pass_rate\":0.5344,\"provider_group\":\"provider_z\",\"quality\":0.7438,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.94,\"id\":\"worker_b\",\"p95_latency_ms\":87750,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9478,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.871,\"id\":\"worker_c\",\"p95_latency_ms\":53850,\"pass_rate\":0.8035,\"provider_group\":\"provider_x\",\"quality\":0.9033,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.743,\"estimated_latency_ms\":167588,\"estimated_pass_rate\":0.7698,\"estimated_quality\":0.9076,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.743,\"max_latency_ms\":232117,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4977, "category": "review", "complexity": 0.6855, "consequence": 0.823, "evidence_need": 0.8293, "family": "policy_compliance", "input_tokens": 7880, "request_id": "synthetic-policy_compliance-000918", "reversibility": 0.4653, "risk": 0.6369, "summary": "Identify evidence required before approving provider failover for production."}, "workers": [{"credits_per_call": 0.932, "id": "worker_a", "p95_latency_ms": 25988, "pass_rate": 0.5344, "provider_group": "provider_z", "quality": 0.7438, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.94, "id": "worker_b", "p95_latency_ms": 87750, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9478, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.871, "id": "worker_c", "p95_latency_ms": 53850, "pass_rate": 0.8035, "provider_group": "provider_x", "quality": 0.9033, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 2.983, "max_latency_ms": 41480, "minimum_quality": 0.518}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000080", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.94, "estimated_latency_ms": 25962, "estimated_pass_rate": 0.3932, "estimated_quality": 0.6, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.983, "max_latency_ms": 41480, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3932, "quality": 0.6, "quality_target_met": true, "utility": 41.221253}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":2.983,\"max_latency_ms\":41480,\"minimum_quality\":0.518},\"task\":{\"ambiguity\":0.6044,\"category\":\"analysis\",\"complexity\":0.6441,\"consequence\":0.4904,\"evidence_need\":0.6942,\"family\":\"performance_tuning\",\"input_tokens\":5880,\"reversibility\":0.8051,\"risk\":0.5018,\"summary\":\"Choose between throughput and tail-latency changes for rate limiting.\"},\"workers\":[{\"credits_per_call\":3.925,\"id\":\"worker_a\",\"p95_latency_ms\":57061,\"pass_rate\":0.8126,\"provider_group\":\"provider_x\",\"quality\":0.9795,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.94,\"id\":\"worker_b\",\"p95_latency_ms\":25962,\"pass_rate\":0.5526,\"provider_group\":\"provider_z\",\"quality\":0.7132,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.423,\"id\":\"worker_c\",\"p95_latency_ms\":81482,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.94,\"estimated_latency_ms\":25962,\"estimated_pass_rate\":0.3932,\"estimated_quality\":0.6,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.983,\"max_latency_ms\":41480,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6044, "category": "analysis", "complexity": 0.6441, "consequence": 0.4904, "evidence_need": 0.6942, "family": "performance_tuning", "input_tokens": 5880, "request_id": "synthetic-performance_tuning-000080", "reversibility": 0.8051, "risk": 0.5018, "summary": "Choose between throughput and tail-latency changes for rate limiting."}, "workers": [{"credits_per_call": 3.925, "id": "worker_a", "p95_latency_ms": 57061, "pass_rate": 0.8126, "provider_group": "provider_x", "quality": 0.9795, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.94, "id": "worker_b", "p95_latency_ms": 25962, "pass_rate": 0.5526, "provider_group": "provider_z", "quality": 0.7132, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.423, "id": "worker_c", "p95_latency_ms": 81482, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.214, "max_latency_ms": 153378, "minimum_quality": 0.848}, "family": "migration_planning", "id": "synthetic-migration_planning-000157", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.253, "estimated_latency_ms": 135019, "estimated_pass_rate": 0.6837, "estimated_quality": 0.9188, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.214, "max_latency_ms": 153378, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6837, "quality": 0.9188, "quality_target_met": true, "utility": 73.829377}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.214,\"max_latency_ms\":153378,\"minimum_quality\":0.848},\"task\":{\"ambiguity\":0.3716,\"category\":\"coding\",\"complexity\":0.8859,\"consequence\":0.7648,\"evidence_need\":0.8481,\"family\":\"migration_planning\",\"input_tokens\":6320,\"reversibility\":0.3983,\"risk\":0.5481,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":1.104,\"id\":\"worker_a\",\"p95_latency_ms\":24287,\"pass_rate\":0.5015,\"provider_group\":\"provider_x\",\"quality\":0.7969,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.831,\"id\":\"worker_b\",\"p95_latency_ms\":57804,\"pass_rate\":0.9089,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.318,\"id\":\"worker_c\",\"p95_latency_ms\":52928,\"pass_rate\":0.7465,\"provider_group\":\"provider_y\",\"quality\":0.7964,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.253,\"estimated_latency_ms\":135019,\"estimated_pass_rate\":0.6837,\"estimated_quality\":0.9188,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.214,\"max_latency_ms\":153378,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3716, "category": "coding", "complexity": 0.8859, "consequence": 0.7648, "evidence_need": 0.8481, "family": "migration_planning", "input_tokens": 6320, "request_id": "synthetic-migration_planning-000157", "reversibility": 0.3983, "risk": 0.5481, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 1.104, "id": "worker_a", "p95_latency_ms": 24287, "pass_rate": 0.5015, "provider_group": "provider_x", "quality": 0.7969, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.831, "id": "worker_b", "p95_latency_ms": 57804, "pass_rate": 0.9089, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.318, "id": "worker_c", "p95_latency_ms": 52928, "pass_rate": 0.7465, "provider_group": "provider_y", "quality": 0.7964, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.364, "max_latency_ms": 225538, "minimum_quality": 0.8723}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001121", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.153, "estimated_latency_ms": 147534, "estimated_pass_rate": 0.9017, "estimated_quality": 0.924, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.364, "max_latency_ms": 225538, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9017, "quality": 0.924, "quality_target_met": true, "utility": 67.880117}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.364,\"max_latency_ms\":225538,\"minimum_quality\":0.8723},\"task\":{\"ambiguity\":0.4235,\"category\":\"analysis\",\"complexity\":0.894,\"consequence\":0.3322,\"evidence_need\":0.5624,\"family\":\"performance_tuning\",\"input_tokens\":2200,\"reversibility\":0.6115,\"risk\":0.375,\"summary\":\"Diagnose latency variance in streaming inference and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":11.152,\"id\":\"worker_a\",\"p95_latency_ms\":68783,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9802,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.07,\"id\":\"worker_b\",\"p95_latency_ms\":29033,\"pass_rate\":0.3889,\"provider_group\":\"provider_y\",\"quality\":0.7117,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.001,\"id\":\"worker_c\",\"p95_latency_ms\":78751,\"pass_rate\":0.9017,\"provider_group\":\"provider_x\",\"quality\":0.9534,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.153,\"estimated_latency_ms\":147534,\"estimated_pass_rate\":0.9017,\"estimated_quality\":0.924,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.364,\"max_latency_ms\":225538,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4235, "category": "analysis", "complexity": 0.894, "consequence": 0.3322, "evidence_need": 0.5624, "family": "performance_tuning", "input_tokens": 2200, "request_id": "synthetic-performance_tuning-001121", "reversibility": 0.6115, "risk": 0.375, "summary": "Diagnose latency variance in streaming inference and rank the next experiments."}, "workers": [{"credits_per_call": 11.152, "id": "worker_a", "p95_latency_ms": 68783, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9802, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.07, "id": "worker_b", "p95_latency_ms": 29033, "pass_rate": 0.3889, "provider_group": "provider_y", "quality": 0.7117, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.001, "id": "worker_c", "p95_latency_ms": 78751, "pass_rate": 0.9017, "provider_group": "provider_x", "quality": 0.9534, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.129, "max_latency_ms": 120746, "minimum_quality": 0.7394}, "family": "migration_planning", "id": "synthetic-migration_planning-000481", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.832, "estimated_latency_ms": 100102, "estimated_pass_rate": 0.8593, "estimated_quality": 0.8854, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.129, "max_latency_ms": 120746, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.8593, "quality": 0.8854, "quality_target_met": true, "utility": 64.286696}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.129,\"max_latency_ms\":120746,\"minimum_quality\":0.7394},\"task\":{\"ambiguity\":0.5173,\"category\":\"coding\",\"complexity\":0.7705,\"consequence\":0.7313,\"evidence_need\":0.4878,\"family\":\"migration_planning\",\"input_tokens\":6320,\"reversibility\":0.4982,\"risk\":0.6374,\"summary\":\"Design compatibility testing for a data retention schema migration.\"},\"workers\":[{\"credits_per_call\":10.67,\"id\":\"worker_a\",\"p95_latency_ms\":77588,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.162,\"id\":\"worker_b\",\"p95_latency_ms\":22514,\"pass_rate\":0.6035,\"provider_group\":\"provider_x\",\"quality\":0.7442,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.596,\"id\":\"worker_c\",\"p95_latency_ms\":63109,\"pass_rate\":0.7457,\"provider_group\":\"provider_z\",\"quality\":0.878,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.832,\"estimated_latency_ms\":100102,\"estimated_pass_rate\":0.8593,\"estimated_quality\":0.8854,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.129,\"max_latency_ms\":120746,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5173, "category": "coding", "complexity": 0.7705, "consequence": 0.7313, "evidence_need": 0.4878, "family": "migration_planning", "input_tokens": 6320, "request_id": "synthetic-migration_planning-000481", "reversibility": 0.4982, "risk": 0.6374, "summary": "Design compatibility testing for a data retention schema migration."}, "workers": [{"credits_per_call": 10.67, "id": "worker_a", "p95_latency_ms": 77588, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.162, "id": "worker_b", "p95_latency_ms": 22514, "pass_rate": 0.6035, "provider_group": "provider_x", "quality": 0.7442, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.596, "id": "worker_c", "p95_latency_ms": 63109, "pass_rate": 0.7457, "provider_group": "provider_z", "quality": 0.878, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.41, "max_latency_ms": 193705, "minimum_quality": 0.7891}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000609", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.478, "estimated_latency_ms": 174444, "estimated_pass_rate": 0.7407, "estimated_quality": 0.9193, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 25.41, "max_latency_ms": 193705, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7407, "quality": 0.9193, "quality_target_met": true, "utility": 75.942195}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.41,\"max_latency_ms\":193705,\"minimum_quality\":0.7891},\"task\":{\"ambiguity\":0.7501,\"category\":\"review\",\"complexity\":0.7004,\"consequence\":0.8725,\"evidence_need\":0.7156,\"family\":\"policy_compliance\",\"input_tokens\":3760,\"reversibility\":0.4865,\"risk\":0.5008,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":14.731,\"id\":\"worker_a\",\"p95_latency_ms\":73420,\"pass_rate\":0.9279,\"provider_group\":\"provider_x\",\"quality\":0.9946,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.106,\"id\":\"worker_b\",\"p95_latency_ms\":26722,\"pass_rate\":0.5597,\"provider_group\":\"provider_z\",\"quality\":0.7253,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.641,\"id\":\"worker_c\",\"p95_latency_ms\":74302,\"pass_rate\":0.8806,\"provider_group\":\"provider_y\",\"quality\":0.8449,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.478,\"estimated_latency_ms\":174444,\"estimated_pass_rate\":0.7407,\"estimated_quality\":0.9193,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.41,\"max_latency_ms\":193705,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7501, "category": "review", "complexity": 0.7004, "consequence": 0.8725, "evidence_need": 0.7156, "family": "policy_compliance", "input_tokens": 3760, "request_id": "synthetic-policy_compliance-000609", "reversibility": 0.4865, "risk": 0.5008, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 14.731, "id": "worker_a", "p95_latency_ms": 73420, "pass_rate": 0.9279, "provider_group": "provider_x", "quality": 0.9946, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.106, "id": "worker_b", "p95_latency_ms": 26722, "pass_rate": 0.5597, "provider_group": "provider_z", "quality": 0.7253, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.641, "id": "worker_c", "p95_latency_ms": 74302, "pass_rate": 0.8806, "provider_group": "provider_y", "quality": 0.8449, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.607, "max_latency_ms": 180486, "minimum_quality": 0.6878}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000282", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.348, "estimated_latency_ms": 74338, "estimated_pass_rate": 0.5816, "estimated_quality": 0.8316, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.607, "max_latency_ms": 180486, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 5, "pass_rate": 0.5816, "quality": 0.8316, "quality_target_met": true, "utility": 65.537911}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.607,\"max_latency_ms\":180486,\"minimum_quality\":0.6878},\"task\":{\"ambiguity\":0.4489,\"category\":\"review\",\"complexity\":0.8382,\"consequence\":0.8788,\"evidence_need\":0.8011,\"family\":\"policy_compliance\",\"input_tokens\":5040,\"reversibility\":0.3476,\"risk\":0.6449,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":3.398,\"id\":\"worker_a\",\"p95_latency_ms\":51110,\"pass_rate\":0.7846,\"provider_group\":\"provider_x\",\"quality\":0.9226,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.544,\"id\":\"worker_b\",\"p95_latency_ms\":80389,\"pass_rate\":0.9177,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.95,\"id\":\"worker_c\",\"p95_latency_ms\":23228,\"pass_rate\":0.4612,\"provider_group\":\"provider_z\",\"quality\":0.7831,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.348,\"estimated_latency_ms\":74338,\"estimated_pass_rate\":0.5816,\"estimated_quality\":0.8316,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.607,\"max_latency_ms\":180486,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4489, "category": "review", "complexity": 0.8382, "consequence": 0.8788, "evidence_need": 0.8011, "family": "policy_compliance", "input_tokens": 5040, "request_id": "synthetic-policy_compliance-000282", "reversibility": 0.3476, "risk": 0.6449, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 3.398, "id": "worker_a", "p95_latency_ms": 51110, "pass_rate": 0.7846, "provider_group": "provider_x", "quality": 0.9226, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.544, "id": "worker_b", "p95_latency_ms": 80389, "pass_rate": 0.9177, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.95, "id": "worker_c", "p95_latency_ms": 23228, "pass_rate": 0.4612, "provider_group": "provider_z", "quality": 0.7831, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.402, "max_latency_ms": 173215, "minimum_quality": 0.8251}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000849", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.668, "estimated_latency_ms": 138824, "estimated_pass_rate": 0.7535, "estimated_quality": 0.8791, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.402, "max_latency_ms": 173215, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7535, "quality": 0.8791, "quality_target_met": true, "utility": 74.790108}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.402,\"max_latency_ms\":173215,\"minimum_quality\":0.8251},\"task\":{\"ambiguity\":0.7048,\"category\":\"review\",\"complexity\":0.6268,\"consequence\":0.8881,\"evidence_need\":0.6371,\"family\":\"policy_compliance\",\"input_tokens\":2160,\"reversibility\":0.3283,\"risk\":0.7436,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":4.265,\"id\":\"worker_a\",\"p95_latency_ms\":48774,\"pass_rate\":0.7877,\"provider_group\":\"provider_y\",\"quality\":0.8514,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.163,\"id\":\"worker_b\",\"p95_latency_ms\":26086,\"pass_rate\":0.5161,\"provider_group\":\"provider_x\",\"quality\":0.7552,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.24,\"id\":\"worker_c\",\"p95_latency_ms\":63964,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9117,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.668,\"estimated_latency_ms\":138824,\"estimated_pass_rate\":0.7535,\"estimated_quality\":0.8791,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.402,\"max_latency_ms\":173215,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7048, "category": "review", "complexity": 0.6268, "consequence": 0.8881, "evidence_need": 0.6371, "family": "policy_compliance", "input_tokens": 2160, "request_id": "synthetic-policy_compliance-000849", "reversibility": 0.3283, "risk": 0.7436, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 4.265, "id": "worker_a", "p95_latency_ms": 48774, "pass_rate": 0.7877, "provider_group": "provider_y", "quality": 0.8514, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.163, "id": "worker_b", "p95_latency_ms": 26086, "pass_rate": 0.5161, "provider_group": "provider_x", "quality": 0.7552, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.24, "id": "worker_c", "p95_latency_ms": 63964, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9117, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.615, "max_latency_ms": 55918, "minimum_quality": 0.6488}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000552", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 2.768, "estimated_latency_ms": 48257, "estimated_pass_rate": 0.5711, "estimated_quality": 0.7561, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.615, "max_latency_ms": 55918, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 2, "pass_rate": 0.5711, "quality": 0.7561, "quality_target_met": true, "utility": 50.770129}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.615,\"max_latency_ms\":55918,\"minimum_quality\":0.6488},\"task\":{\"ambiguity\":0.7619,\"category\":\"review\",\"complexity\":0.5821,\"consequence\":0.5849,\"evidence_need\":0.9159,\"family\":\"policy_compliance\",\"input_tokens\":3080,\"reversibility\":0.435,\"risk\":0.7231,\"summary\":\"Review whether credit reservation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.768,\"id\":\"worker_a\",\"p95_latency_ms\":48257,\"pass_rate\":0.7533,\"provider_group\":\"provider_y\",\"quality\":0.8846,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.015,\"id\":\"worker_b\",\"p95_latency_ms\":28568,\"pass_rate\":0.4676,\"provider_group\":\"provider_x\",\"quality\":0.6905,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.006,\"id\":\"worker_c\",\"p95_latency_ms\":69866,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":2.768,\"estimated_latency_ms\":48257,\"estimated_pass_rate\":0.5711,\"estimated_quality\":0.7561,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.615,\"max_latency_ms\":55918,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7619, "category": "review", "complexity": 0.5821, "consequence": 0.5849, "evidence_need": 0.9159, "family": "policy_compliance", "input_tokens": 3080, "request_id": "synthetic-policy_compliance-000552", "reversibility": 0.435, "risk": 0.7231, "summary": "Review whether credit reservation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.768, "id": "worker_a", "p95_latency_ms": 48257, "pass_rate": 0.7533, "provider_group": "provider_y", "quality": 0.8846, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.015, "id": "worker_b", "p95_latency_ms": 28568, "pass_rate": 0.4676, "provider_group": "provider_x", "quality": 0.6905, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.006, "id": "worker_c", "p95_latency_ms": 69866, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.838, "max_latency_ms": 202124, "minimum_quality": 0.8217}, "family": "migration_planning", "id": "synthetic-migration_planning-000865", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.586, "estimated_latency_ms": 103843, "estimated_pass_rate": 0.5885, "estimated_quality": 0.7693, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.838, "max_latency_ms": 202124, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5885, "quality": 0.7693, "quality_target_met": false, "utility": 56.192256}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.838,\"max_latency_ms\":202124,\"minimum_quality\":0.8217},\"task\":{\"ambiguity\":0.572,\"category\":\"coding\",\"complexity\":0.9315,\"consequence\":0.6163,\"evidence_need\":0.8074,\"family\":\"migration_planning\",\"input_tokens\":6600,\"reversibility\":0.5415,\"risk\":0.5455,\"summary\":\"Plan a zero-downtime migration of deployment rollback with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":15.296,\"id\":\"worker_a\",\"p95_latency_ms\":71383,\"pass_rate\":0.9096,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.669,\"id\":\"worker_b\",\"p95_latency_ms\":71122,\"pass_rate\":0.7197,\"provider_group\":\"provider_x\",\"quality\":0.8137,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.917,\"id\":\"worker_c\",\"p95_latency_ms\":32721,\"pass_rate\":0.5904,\"provider_group\":\"provider_z\",\"quality\":0.7697,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.586,\"estimated_latency_ms\":103843,\"estimated_pass_rate\":0.5885,\"estimated_quality\":0.7693,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.838,\"max_latency_ms\":202124,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.572, "category": "coding", "complexity": 0.9315, "consequence": 0.6163, "evidence_need": 0.8074, "family": "migration_planning", "input_tokens": 6600, "request_id": "synthetic-migration_planning-000865", "reversibility": 0.5415, "risk": 0.5455, "summary": "Plan a zero-downtime migration of deployment rollback with rollback checkpoints."}, "workers": [{"credits_per_call": 15.296, "id": "worker_a", "p95_latency_ms": 71383, "pass_rate": 0.9096, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.669, "id": "worker_b", "p95_latency_ms": 71122, "pass_rate": 0.7197, "provider_group": "provider_x", "quality": 0.8137, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.917, "id": "worker_c", "p95_latency_ms": 32721, "pass_rate": 0.5904, "provider_group": "provider_z", "quality": 0.7697, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.717, "max_latency_ms": 174003, "minimum_quality": 0.8926}, "family": "migration_planning", "id": "synthetic-migration_planning-001078", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.562, "estimated_latency_ms": 157123, "estimated_pass_rate": 0.6981, "estimated_quality": 0.8971, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.717, "max_latency_ms": 174003, "policy": "strict", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.6981, "quality": 0.8971, "quality_target_met": true, "utility": 74.420573}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.717,\"max_latency_ms\":174003,\"minimum_quality\":0.8926},\"task\":{\"ambiguity\":0.6822,\"category\":\"coding\",\"complexity\":0.7845,\"consequence\":0.6888,\"evidence_need\":0.7625,\"family\":\"migration_planning\",\"input_tokens\":7040,\"reversibility\":0.3959,\"risk\":0.6027,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":3.16,\"id\":\"worker_a\",\"p95_latency_ms\":52386,\"pass_rate\":0.7312,\"provider_group\":\"provider_y\",\"quality\":0.8655,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.215,\"id\":\"worker_b\",\"p95_latency_ms\":32718,\"pass_rate\":0.5288,\"provider_group\":\"provider_z\",\"quality\":0.799,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.187,\"id\":\"worker_c\",\"p95_latency_ms\":72019,\"pass_rate\":0.8992,\"provider_group\":\"provider_x\",\"quality\":0.9483,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.562,\"estimated_latency_ms\":157123,\"estimated_pass_rate\":0.6981,\"estimated_quality\":0.8971,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.717,\"max_latency_ms\":174003,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6822, "category": "coding", "complexity": 0.7845, "consequence": 0.6888, "evidence_need": 0.7625, "family": "migration_planning", "input_tokens": 7040, "request_id": "synthetic-migration_planning-001078", "reversibility": 0.3959, "risk": 0.6027, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 3.16, "id": "worker_a", "p95_latency_ms": 52386, "pass_rate": 0.7312, "provider_group": "provider_y", "quality": 0.8655, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.215, "id": "worker_b", "p95_latency_ms": 32718, "pass_rate": 0.5288, "provider_group": "provider_z", "quality": 0.799, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.187, "id": "worker_c", "p95_latency_ms": 72019, "pass_rate": 0.8992, "provider_group": "provider_x", "quality": 0.9483, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.162, "max_latency_ms": 248522, "minimum_quality": 0.886}, "family": "migration_planning", "id": "synthetic-migration_planning-000004", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.963, "estimated_latency_ms": 182653, "estimated_pass_rate": 0.7716, "estimated_quality": 0.8718, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.162, "max_latency_ms": 248522, "policy": "strict", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7716, "quality": 0.8718, "quality_target_met": false, "utility": 72.367295}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.162,\"max_latency_ms\":248522,\"minimum_quality\":0.886},\"task\":{\"ambiguity\":0.6659,\"category\":\"coding\",\"complexity\":0.6003,\"consequence\":0.8281,\"evidence_need\":0.4083,\"family\":\"migration_planning\",\"input_tokens\":1800,\"reversibility\":0.5833,\"risk\":0.7184,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":0.781,\"id\":\"worker_a\",\"p95_latency_ms\":27252,\"pass_rate\":0.6033,\"provider_group\":\"provider_z\",\"quality\":0.7002,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.083,\"id\":\"worker_b\",\"p95_latency_ms\":86878,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9457,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.099,\"id\":\"worker_c\",\"p95_latency_ms\":68523,\"pass_rate\":0.7615,\"provider_group\":\"provider_y\",\"quality\":0.7939,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.963,\"estimated_latency_ms\":182653,\"estimated_pass_rate\":0.7716,\"estimated_quality\":0.8718,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.162,\"max_latency_ms\":248522,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6659, "category": "coding", "complexity": 0.6003, "consequence": 0.8281, "evidence_need": 0.4083, "family": "migration_planning", "input_tokens": 1800, "request_id": "synthetic-migration_planning-000004", "reversibility": 0.5833, "risk": 0.7184, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 0.781, "id": "worker_a", "p95_latency_ms": 27252, "pass_rate": 0.6033, "provider_group": "provider_z", "quality": 0.7002, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.083, "id": "worker_b", "p95_latency_ms": 86878, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9457, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.099, "id": "worker_c", "p95_latency_ms": 68523, "pass_rate": 0.7615, "provider_group": "provider_y", "quality": 0.7939, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.011, "max_latency_ms": 163972, "minimum_quality": 0.7316}, "family": "migration_planning", "id": "synthetic-migration_planning-000187", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.023, "estimated_latency_ms": 97556, "estimated_pass_rate": 0.6495, "estimated_quality": 0.8278, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.011, "max_latency_ms": 163972, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6495, "quality": 0.8278, "quality_target_met": true, "utility": 59.50933}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.011,\"max_latency_ms\":163972,\"minimum_quality\":0.7316},\"task\":{\"ambiguity\":0.4011,\"category\":\"coding\",\"complexity\":0.759,\"consequence\":0.8009,\"evidence_need\":0.4203,\"family\":\"migration_planning\",\"input_tokens\":5600,\"reversibility\":0.4724,\"risk\":0.5269,\"summary\":\"Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.174,\"id\":\"worker_a\",\"p95_latency_ms\":26181,\"pass_rate\":0.5317,\"provider_group\":\"provider_z\",\"quality\":0.7035,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.236,\"id\":\"worker_b\",\"p95_latency_ms\":91871,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9751,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.849,\"id\":\"worker_c\",\"p95_latency_ms\":71375,\"pass_rate\":0.7256,\"provider_group\":\"provider_y\",\"quality\":0.9181,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.023,\"estimated_latency_ms\":97556,\"estimated_pass_rate\":0.6495,\"estimated_quality\":0.8278,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.011,\"max_latency_ms\":163972,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4011, "category": "coding", "complexity": 0.759, "consequence": 0.8009, "evidence_need": 0.4203, "family": "migration_planning", "input_tokens": 5600, "request_id": "synthetic-migration_planning-000187", "reversibility": 0.4724, "risk": 0.5269, "summary": "Plan a zero-downtime migration of OAuth callbacks with rollback checkpoints."}, "workers": [{"credits_per_call": 1.174, "id": "worker_a", "p95_latency_ms": 26181, "pass_rate": 0.5317, "provider_group": "provider_z", "quality": 0.7035, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.236, "id": "worker_b", "p95_latency_ms": 91871, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9751, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.849, "id": "worker_c", "p95_latency_ms": 71375, "pass_rate": 0.7256, "provider_group": "provider_y", "quality": 0.9181, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.849, "max_latency_ms": 25851, "minimum_quality": 0.6736}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000404", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.985, "estimated_latency_ms": 18064, "estimated_pass_rate": 0.3147, "estimated_quality": 0.5571, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.849, "max_latency_ms": 25851, "policy": "lean", "reason_codes": ["best_feasible_below_target"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3147, "quality": 0.5571, "quality_target_met": false, "utility": 31.949444}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.849,\"max_latency_ms\":25851,\"minimum_quality\":0.6736},\"task\":{\"ambiguity\":0.6319,\"category\":\"analysis\",\"complexity\":0.591,\"consequence\":0.4464,\"evidence_need\":0.5909,\"family\":\"performance_tuning\",\"input_tokens\":3040,\"reversibility\":0.6031,\"risk\":0.3952,\"summary\":\"Diagnose latency variance in webhook delivery and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.985,\"id\":\"worker_a\",\"p95_latency_ms\":18064,\"pass_rate\":0.4615,\"provider_group\":\"provider_y\",\"quality\":0.6549,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.98,\"id\":\"worker_b\",\"p95_latency_ms\":62194,\"pass_rate\":0.9639,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.963,\"id\":\"worker_c\",\"p95_latency_ms\":56849,\"pass_rate\":0.9055,\"provider_group\":\"provider_x\",\"quality\":0.9191,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.985,\"estimated_latency_ms\":18064,\"estimated_pass_rate\":0.3147,\"estimated_quality\":0.5571,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.849,\"max_latency_ms\":25851,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6319, "category": "analysis", "complexity": 0.591, "consequence": 0.4464, "evidence_need": 0.5909, "family": "performance_tuning", "input_tokens": 3040, "request_id": "synthetic-performance_tuning-000404", "reversibility": 0.6031, "risk": 0.3952, "summary": "Diagnose latency variance in webhook delivery and rank the next experiments."}, "workers": [{"credits_per_call": 0.985, "id": "worker_a", "p95_latency_ms": 18064, "pass_rate": 0.4615, "provider_group": "provider_y", "quality": 0.6549, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.98, "id": "worker_b", "p95_latency_ms": 62194, "pass_rate": 0.9639, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.963, "id": "worker_c", "p95_latency_ms": 56849, "pass_rate": 0.9055, "provider_group": "provider_x", "quality": 0.9191, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.091, "max_latency_ms": 32186, "minimum_quality": 0.695}, "family": "migration_planning", "id": "synthetic-migration_planning-000514", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.112, "estimated_latency_ms": 29652, "estimated_pass_rate": 0.3407, "estimated_quality": 0.6727, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.091, "max_latency_ms": 32186, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3407, "quality": 0.6727, "quality_target_met": false, "utility": 41.942906}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.091,\"max_latency_ms\":32186,\"minimum_quality\":0.695},\"task\":{\"ambiguity\":0.4608,\"category\":\"coding\",\"complexity\":0.6593,\"consequence\":0.5849,\"evidence_need\":0.7847,\"family\":\"migration_planning\",\"input_tokens\":160,\"reversibility\":0.5024,\"risk\":0.4005,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":14.211,\"id\":\"worker_a\",\"p95_latency_ms\":76719,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.112,\"id\":\"worker_b\",\"p95_latency_ms\":29652,\"pass_rate\":0.4827,\"provider_group\":\"provider_z\",\"quality\":0.7717,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.321,\"id\":\"worker_c\",\"p95_latency_ms\":74624,\"pass_rate\":0.7882,\"provider_group\":\"provider_y\",\"quality\":0.8581,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.112,\"estimated_latency_ms\":29652,\"estimated_pass_rate\":0.3407,\"estimated_quality\":0.6727,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.091,\"max_latency_ms\":32186,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4608, "category": "coding", "complexity": 0.6593, "consequence": 0.5849, "evidence_need": 0.7847, "family": "migration_planning", "input_tokens": 160, "request_id": "synthetic-migration_planning-000514", "reversibility": 0.5024, "risk": 0.4005, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 14.211, "id": "worker_a", "p95_latency_ms": 76719, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.112, "id": "worker_b", "p95_latency_ms": 29652, "pass_rate": 0.4827, "provider_group": "provider_z", "quality": 0.7717, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.321, "id": "worker_c", "p95_latency_ms": 74624, "pass_rate": 0.7882, "provider_group": "provider_y", "quality": 0.8581, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.932, "max_latency_ms": 29829, "minimum_quality": 0.7719}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000624", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.197, "estimated_latency_ms": 21837, "estimated_pass_rate": 0.2912, "estimated_quality": 0.5473, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.932, "max_latency_ms": 29829, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2912, "quality": 0.5473, "quality_target_met": false, "utility": 23.565716}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.932,\"max_latency_ms\":29829,\"minimum_quality\":0.7719},\"task\":{\"ambiguity\":0.451,\"category\":\"review\",\"complexity\":0.8017,\"consequence\":0.6571,\"evidence_need\":0.6801,\"family\":\"policy_compliance\",\"input_tokens\":4680,\"reversibility\":0.4029,\"risk\":0.4856,\"summary\":\"Audit a workflow for observability against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.241,\"id\":\"worker_a\",\"p95_latency_ms\":78261,\"pass_rate\":0.9296,\"provider_group\":\"provider_z\",\"quality\":0.9135,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.537,\"id\":\"worker_b\",\"p95_latency_ms\":79718,\"pass_rate\":0.7521,\"provider_group\":\"provider_x\",\"quality\":0.8796,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.197,\"id\":\"worker_c\",\"p95_latency_ms\":21837,\"pass_rate\":0.4555,\"provider_group\":\"provider_y\",\"quality\":0.6775,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.197,\"estimated_latency_ms\":21837,\"estimated_pass_rate\":0.2912,\"estimated_quality\":0.5473,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.932,\"max_latency_ms\":29829,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.451, "category": "review", "complexity": 0.8017, "consequence": 0.6571, "evidence_need": 0.6801, "family": "policy_compliance", "input_tokens": 4680, "request_id": "synthetic-policy_compliance-000624", "reversibility": 0.4029, "risk": 0.4856, "summary": "Audit a workflow for observability against explicit access requirements."}, "workers": [{"credits_per_call": 13.241, "id": "worker_a", "p95_latency_ms": 78261, "pass_rate": 0.9296, "provider_group": "provider_z", "quality": 0.9135, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.537, "id": "worker_b", "p95_latency_ms": 79718, "pass_rate": 0.7521, "provider_group": "provider_x", "quality": 0.8796, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.197, "id": "worker_c", "p95_latency_ms": 21837, "pass_rate": 0.4555, "provider_group": "provider_y", "quality": 0.6775, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.318, "max_latency_ms": 19783, "minimum_quality": 0.6532}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000657", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.15, "estimated_latency_ms": 17648, "estimated_pass_rate": 0.3569, "estimated_quality": 0.6597, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.318, "max_latency_ms": 19783, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3569, "quality": 0.6597, "quality_target_met": true, "utility": 41.622923}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.318,\"max_latency_ms\":19783,\"minimum_quality\":0.6532},\"task\":{\"ambiguity\":0.5528,\"category\":\"review\",\"complexity\":0.8276,\"consequence\":0.7246,\"evidence_need\":0.7494,\"family\":\"policy_compliance\",\"input_tokens\":720,\"reversibility\":0.3938,\"risk\":0.4646,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":1.15,\"id\":\"worker_a\",\"p95_latency_ms\":17648,\"pass_rate\":0.5302,\"provider_group\":\"provider_x\",\"quality\":0.7999,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.485,\"id\":\"worker_b\",\"p95_latency_ms\":91155,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9466,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.484,\"id\":\"worker_c\",\"p95_latency_ms\":73944,\"pass_rate\":0.7458,\"provider_group\":\"provider_y\",\"quality\":0.8752,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.15,\"estimated_latency_ms\":17648,\"estimated_pass_rate\":0.3569,\"estimated_quality\":0.6597,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.318,\"max_latency_ms\":19783,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5528, "category": "review", "complexity": 0.8276, "consequence": 0.7246, "evidence_need": 0.7494, "family": "policy_compliance", "input_tokens": 720, "request_id": "synthetic-policy_compliance-000657", "reversibility": 0.3938, "risk": 0.4646, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 1.15, "id": "worker_a", "p95_latency_ms": 17648, "pass_rate": 0.5302, "provider_group": "provider_x", "quality": 0.7999, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.485, "id": "worker_b", "p95_latency_ms": 91155, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9466, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.484, "id": "worker_c", "p95_latency_ms": 73944, "pass_rate": 0.7458, "provider_group": "provider_y", "quality": 0.8752, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.713, "max_latency_ms": 199257, "minimum_quality": 0.788}, "family": "migration_planning", "id": "synthetic-migration_planning-000928", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.623, "estimated_latency_ms": 144989, "estimated_pass_rate": 0.7333, "estimated_quality": 0.9111, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.713, "max_latency_ms": 199257, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7333, "quality": 0.9111, "quality_target_met": true, "utility": 75.571356}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.713,\"max_latency_ms\":199257,\"minimum_quality\":0.788},\"task\":{\"ambiguity\":0.7265,\"category\":\"coding\",\"complexity\":0.9396,\"consequence\":0.6986,\"evidence_need\":0.5153,\"family\":\"migration_planning\",\"input_tokens\":1400,\"reversibility\":0.6147,\"risk\":0.4877,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.733,\"id\":\"worker_a\",\"p95_latency_ms\":50599,\"pass_rate\":0.8268,\"provider_group\":\"provider_z\",\"quality\":0.8753,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.782,\"id\":\"worker_b\",\"p95_latency_ms\":26435,\"pass_rate\":0.5223,\"provider_group\":\"provider_x\",\"quality\":0.7435,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.108,\"id\":\"worker_c\",\"p95_latency_ms\":67955,\"pass_rate\":0.985,\"provider_group\":\"provider_y\",\"quality\":0.9925,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.623,\"estimated_latency_ms\":144989,\"estimated_pass_rate\":0.7333,\"estimated_quality\":0.9111,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.713,\"max_latency_ms\":199257,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7265, "category": "coding", "complexity": 0.9396, "consequence": 0.6986, "evidence_need": 0.5153, "family": "migration_planning", "input_tokens": 1400, "request_id": "synthetic-migration_planning-000928", "reversibility": 0.6147, "risk": 0.4877, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 3.733, "id": "worker_a", "p95_latency_ms": 50599, "pass_rate": 0.8268, "provider_group": "provider_z", "quality": 0.8753, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.782, "id": "worker_b", "p95_latency_ms": 26435, "pass_rate": 0.5223, "provider_group": "provider_x", "quality": 0.7435, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.108, "id": "worker_c", "p95_latency_ms": 67955, "pass_rate": 0.985, "provider_group": "provider_y", "quality": 0.9925, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.209, "max_latency_ms": 190095, "minimum_quality": 0.895}, "family": "migration_planning", "id": "synthetic-migration_planning-000259", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.995, "estimated_latency_ms": 142755, "estimated_pass_rate": 0.7283, "estimated_quality": 0.8829, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.209, "max_latency_ms": 190095, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7283, "quality": 0.8829, "quality_target_met": false, "utility": 68.010939}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.209,\"max_latency_ms\":190095,\"minimum_quality\":0.895},\"task\":{\"ambiguity\":0.509,\"category\":\"coding\",\"complexity\":0.7652,\"consequence\":0.6629,\"evidence_need\":0.4267,\"family\":\"migration_planning\",\"input_tokens\":5040,\"reversibility\":0.403,\"risk\":0.4274,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":3.562,\"id\":\"worker_a\",\"p95_latency_ms\":44619,\"pass_rate\":0.8198,\"provider_group\":\"provider_z\",\"quality\":0.8085,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.023,\"id\":\"worker_b\",\"p95_latency_ms\":30267,\"pass_rate\":0.5806,\"provider_group\":\"provider_y\",\"quality\":0.7741,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.41,\"id\":\"worker_c\",\"p95_latency_ms\":67869,\"pass_rate\":0.9151,\"provider_group\":\"provider_x\",\"quality\":0.9333,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.995,\"estimated_latency_ms\":142755,\"estimated_pass_rate\":0.7283,\"estimated_quality\":0.8829,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.209,\"max_latency_ms\":190095,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.509, "category": "coding", "complexity": 0.7652, "consequence": 0.6629, "evidence_need": 0.4267, "family": "migration_planning", "input_tokens": 5040, "request_id": "synthetic-migration_planning-000259", "reversibility": 0.403, "risk": 0.4274, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 3.562, "id": "worker_a", "p95_latency_ms": 44619, "pass_rate": 0.8198, "provider_group": "provider_z", "quality": 0.8085, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.023, "id": "worker_b", "p95_latency_ms": 30267, "pass_rate": 0.5806, "provider_group": "provider_y", "quality": 0.7741, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.41, "id": "worker_c", "p95_latency_ms": 67869, "pass_rate": 0.9151, "provider_group": "provider_x", "quality": 0.9333, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.22, "max_latency_ms": 26282, "minimum_quality": 0.4917}, "family": "migration_planning", "id": "synthetic-migration_planning-000139", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.794, "estimated_latency_ms": 18730, "estimated_pass_rate": 0.3051, "estimated_quality": 0.6014, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.22, "max_latency_ms": 26282, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3051, "quality": 0.6014, "quality_target_met": true, "utility": 36.541415}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.22,\"max_latency_ms\":26282,\"minimum_quality\":0.4917},\"task\":{\"ambiguity\":0.3739,\"category\":\"coding\",\"complexity\":0.7531,\"consequence\":0.8069,\"evidence_need\":0.7198,\"family\":\"migration_planning\",\"input_tokens\":2320,\"reversibility\":0.5204,\"risk\":0.4151,\"summary\":\"Plan a zero-downtime migration of credit reservation with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":9.988,\"id\":\"worker_a\",\"p95_latency_ms\":78331,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.968,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.794,\"id\":\"worker_b\",\"p95_latency_ms\":18730,\"pass_rate\":0.4524,\"provider_group\":\"provider_x\",\"quality\":0.7126,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.067,\"id\":\"worker_c\",\"p95_latency_ms\":48530,\"pass_rate\":0.8834,\"provider_group\":\"provider_z\",\"quality\":0.8392,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.794,\"estimated_latency_ms\":18730,\"estimated_pass_rate\":0.3051,\"estimated_quality\":0.6014,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.22,\"max_latency_ms\":26282,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3739, "category": "coding", "complexity": 0.7531, "consequence": 0.8069, "evidence_need": 0.7198, "family": "migration_planning", "input_tokens": 2320, "request_id": "synthetic-migration_planning-000139", "reversibility": 0.5204, "risk": 0.4151, "summary": "Plan a zero-downtime migration of credit reservation with rollback checkpoints."}, "workers": [{"credits_per_call": 9.988, "id": "worker_a", "p95_latency_ms": 78331, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.968, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.794, "id": "worker_b", "p95_latency_ms": 18730, "pass_rate": 0.4524, "provider_group": "provider_x", "quality": 0.7126, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.067, "id": "worker_c", "p95_latency_ms": 48530, "pass_rate": 0.8834, "provider_group": "provider_z", "quality": 0.8392, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.807, "max_latency_ms": 220439, "minimum_quality": 0.7463}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000828", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.24, "estimated_latency_ms": 127495, "estimated_pass_rate": 0.8233, "estimated_quality": 0.9218, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.807, "max_latency_ms": 220439, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8233, "quality": 0.9218, "quality_target_met": true, "utility": 70.981249}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.807,\"max_latency_ms\":220439,\"minimum_quality\":0.7463},\"task\":{\"ambiguity\":0.5801,\"category\":\"review\",\"complexity\":0.664,\"consequence\":0.5132,\"evidence_need\":0.7598,\"family\":\"policy_compliance\",\"input_tokens\":5800,\"reversibility\":0.5744,\"risk\":0.8248,\"summary\":\"Review whether streaming inference satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.863,\"id\":\"worker_a\",\"p95_latency_ms\":58722,\"pass_rate\":0.7498,\"provider_group\":\"provider_x\",\"quality\":0.8646,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.055,\"id\":\"worker_b\",\"p95_latency_ms\":19709,\"pass_rate\":0.4736,\"provider_group\":\"provider_y\",\"quality\":0.7697,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.377,\"id\":\"worker_c\",\"p95_latency_ms\":68773,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9335,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.24,\"estimated_latency_ms\":127495,\"estimated_pass_rate\":0.8233,\"estimated_quality\":0.9218,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.807,\"max_latency_ms\":220439,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5801, "category": "review", "complexity": 0.664, "consequence": 0.5132, "evidence_need": 0.7598, "family": "policy_compliance", "input_tokens": 5800, "request_id": "synthetic-policy_compliance-000828", "reversibility": 0.5744, "risk": 0.8248, "summary": "Review whether streaming inference satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.863, "id": "worker_a", "p95_latency_ms": 58722, "pass_rate": 0.7498, "provider_group": "provider_x", "quality": 0.8646, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.055, "id": "worker_b", "p95_latency_ms": 19709, "pass_rate": 0.4736, "provider_group": "provider_y", "quality": 0.7697, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.377, "id": "worker_c", "p95_latency_ms": 68773, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9335, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.512, "max_latency_ms": 189592, "minimum_quality": 0.8177}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000570", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.464, "estimated_latency_ms": 162054, "estimated_pass_rate": 0.7376, "estimated_quality": 0.8938, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 23.512, "max_latency_ms": 189592, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7376, "quality": 0.8938, "quality_target_met": true, "utility": 73.718775}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.512,\"max_latency_ms\":189592,\"minimum_quality\":0.8177},\"task\":{\"ambiguity\":0.7768,\"category\":\"review\",\"complexity\":0.5735,\"consequence\":0.8404,\"evidence_need\":0.7433,\"family\":\"policy_compliance\",\"input_tokens\":7520,\"reversibility\":0.5975,\"risk\":0.5648,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.534,\"id\":\"worker_a\",\"p95_latency_ms\":74494,\"pass_rate\":0.7785,\"provider_group\":\"provider_y\",\"quality\":0.9002,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.883,\"id\":\"worker_b\",\"p95_latency_ms\":30124,\"pass_rate\":0.5567,\"provider_group\":\"provider_z\",\"quality\":0.6937,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.047,\"id\":\"worker_c\",\"p95_latency_ms\":57436,\"pass_rate\":0.9235,\"provider_group\":\"provider_x\",\"quality\":0.9539,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.464,\"estimated_latency_ms\":162054,\"estimated_pass_rate\":0.7376,\"estimated_quality\":0.8938,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.512,\"max_latency_ms\":189592,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7768, "category": "review", "complexity": 0.5735, "consequence": 0.8404, "evidence_need": 0.7433, "family": "policy_compliance", "input_tokens": 7520, "request_id": "synthetic-policy_compliance-000570", "reversibility": 0.5975, "risk": 0.5648, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.534, "id": "worker_a", "p95_latency_ms": 74494, "pass_rate": 0.7785, "provider_group": "provider_y", "quality": 0.9002, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.883, "id": "worker_b", "p95_latency_ms": 30124, "pass_rate": 0.5567, "provider_group": "provider_z", "quality": 0.6937, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.047, "id": "worker_c", "p95_latency_ms": 57436, "pass_rate": 0.9235, "provider_group": "provider_x", "quality": 0.9539, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 14.193, "max_latency_ms": 254260, "minimum_quality": 0.8449}, "family": "migration_planning", "id": "synthetic-migration_planning-000895", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 11.324, "estimated_latency_ms": 97898, "estimated_pass_rate": 0.6596, "estimated_quality": 0.8807, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 14.193, "max_latency_ms": 254260, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6596, "quality": 0.8807, "quality_target_met": true, "utility": 64.322677}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":14.193,\"max_latency_ms\":254260,\"minimum_quality\":0.8449},\"task\":{\"ambiguity\":0.5986,\"category\":\"coding\",\"complexity\":0.6116,\"consequence\":0.4625,\"evidence_need\":0.8336,\"family\":\"migration_planning\",\"input_tokens\":7080,\"reversibility\":0.5298,\"risk\":0.6556,\"summary\":\"Design compatibility testing for a data retention schema migration.\"},\"workers\":[{\"credits_per_call\":10.471,\"id\":\"worker_a\",\"p95_latency_ms\":72182,\"pass_rate\":0.9434,\"provider_group\":\"provider_z\",\"quality\":0.9494,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.853,\"id\":\"worker_b\",\"p95_latency_ms\":25716,\"pass_rate\":0.4555,\"provider_group\":\"provider_y\",\"quality\":0.7417,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.142,\"id\":\"worker_c\",\"p95_latency_ms\":73496,\"pass_rate\":0.7416,\"provider_group\":\"provider_x\",\"quality\":0.8416,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":11.324,\"estimated_latency_ms\":97898,\"estimated_pass_rate\":0.6596,\"estimated_quality\":0.8807,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":14.193,\"max_latency_ms\":254260,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5986, "category": "coding", "complexity": 0.6116, "consequence": 0.4625, "evidence_need": 0.8336, "family": "migration_planning", "input_tokens": 7080, "request_id": "synthetic-migration_planning-000895", "reversibility": 0.5298, "risk": 0.6556, "summary": "Design compatibility testing for a data retention schema migration."}, "workers": [{"credits_per_call": 10.471, "id": "worker_a", "p95_latency_ms": 72182, "pass_rate": 0.9434, "provider_group": "provider_z", "quality": 0.9494, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.853, "id": "worker_b", "p95_latency_ms": 25716, "pass_rate": 0.4555, "provider_group": "provider_y", "quality": 0.7417, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.142, "id": "worker_c", "p95_latency_ms": 73496, "pass_rate": 0.7416, "provider_group": "provider_x", "quality": 0.8416, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 22.1, "max_latency_ms": 198315, "minimum_quality": 0.8077}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000632", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.77, "estimated_latency_ms": 138653, "estimated_pass_rate": 0.8955, "estimated_quality": 0.9192, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 22.1, "max_latency_ms": 198315, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8955, "quality": 0.9192, "quality_target_met": true, "utility": 69.77142}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":22.1,\"max_latency_ms\":198315,\"minimum_quality\":0.8077},\"task\":{\"ambiguity\":0.503,\"category\":\"analysis\",\"complexity\":0.8237,\"consequence\":0.6817,\"evidence_need\":0.3658,\"family\":\"performance_tuning\",\"input_tokens\":4720,\"reversibility\":0.7369,\"risk\":0.3665,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":13.677,\"id\":\"worker_a\",\"p95_latency_ms\":59971,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.093,\"id\":\"worker_b\",\"p95_latency_ms\":78682,\"pass_rate\":0.8486,\"provider_group\":\"provider_y\",\"quality\":0.9118,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.955,\"id\":\"worker_c\",\"p95_latency_ms\":18224,\"pass_rate\":0.467,\"provider_group\":\"provider_x\",\"quality\":0.6409,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.77,\"estimated_latency_ms\":138653,\"estimated_pass_rate\":0.8955,\"estimated_quality\":0.9192,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":22.1,\"max_latency_ms\":198315,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.503, "category": "analysis", "complexity": 0.8237, "consequence": 0.6817, "evidence_need": 0.3658, "family": "performance_tuning", "input_tokens": 4720, "request_id": "synthetic-performance_tuning-000632", "reversibility": 0.7369, "risk": 0.3665, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 13.677, "id": "worker_a", "p95_latency_ms": 59971, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.093, "id": "worker_b", "p95_latency_ms": 78682, "pass_rate": 0.8486, "provider_group": "provider_y", "quality": 0.9118, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.955, "id": "worker_c", "p95_latency_ms": 18224, "pass_rate": 0.467, "provider_group": "provider_x", "quality": 0.6409, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 11.38, "max_latency_ms": 222444, "minimum_quality": 0.7379}, "family": "migration_planning", "id": "synthetic-migration_planning-001264", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.843, "estimated_latency_ms": 76728, "estimated_pass_rate": 0.6547, "estimated_quality": 0.8404, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.38, "max_latency_ms": 222444, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 5, "pass_rate": 0.6547, "quality": 0.8404, "quality_target_met": true, "utility": 66.025735}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":11.38,\"max_latency_ms\":222444,\"minimum_quality\":0.7379},\"task\":{\"ambiguity\":0.7026,\"category\":\"coding\",\"complexity\":0.8275,\"consequence\":0.6669,\"evidence_need\":0.7419,\"family\":\"migration_planning\",\"input_tokens\":6440,\"reversibility\":0.4402,\"risk\":0.6919,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":11.052,\"id\":\"worker_a\",\"p95_latency_ms\":89688,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.111,\"id\":\"worker_b\",\"p95_latency_ms\":24621,\"pass_rate\":0.5495,\"provider_group\":\"provider_x\",\"quality\":0.786,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.732,\"id\":\"worker_c\",\"p95_latency_ms\":52107,\"pass_rate\":0.8727,\"provider_group\":\"provider_y\",\"quality\":0.9088,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.843,\"estimated_latency_ms\":76728,\"estimated_pass_rate\":0.6547,\"estimated_quality\":0.8404,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.38,\"max_latency_ms\":222444,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7026, "category": "coding", "complexity": 0.8275, "consequence": 0.6669, "evidence_need": 0.7419, "family": "migration_planning", "input_tokens": 6440, "request_id": "synthetic-migration_planning-001264", "reversibility": 0.4402, "risk": 0.6919, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 11.052, "id": "worker_a", "p95_latency_ms": 89688, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.111, "id": "worker_b", "p95_latency_ms": 24621, "pass_rate": 0.5495, "provider_group": "provider_x", "quality": 0.786, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.732, "id": "worker_c", "p95_latency_ms": 52107, "pass_rate": 0.8727, "provider_group": "provider_y", "quality": 0.9088, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 15.081, "max_latency_ms": 178107, "minimum_quality": 0.8279}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001131", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.833, "estimated_latency_ms": 93802, "estimated_pass_rate": 0.6249, "estimated_quality": 0.89, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.081, "max_latency_ms": 178107, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.6249, "quality": 0.89, "quality_target_met": true, "utility": 63.705027}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":15.081,\"max_latency_ms\":178107,\"minimum_quality\":0.8279},\"task\":{\"ambiguity\":0.5237,\"category\":\"review\",\"complexity\":0.5167,\"consequence\":0.8752,\"evidence_need\":0.8161,\"family\":\"policy_compliance\",\"input_tokens\":5040,\"reversibility\":0.4557,\"risk\":0.5094,\"summary\":\"Audit a workflow for access control against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":14.004,\"id\":\"worker_a\",\"p95_latency_ms\":70480,\"pass_rate\":0.933,\"provider_group\":\"provider_x\",\"quality\":0.9156,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.829,\"id\":\"worker_b\",\"p95_latency_ms\":23322,\"pass_rate\":0.4508,\"provider_group\":\"provider_z\",\"quality\":0.7962,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.355,\"id\":\"worker_c\",\"p95_latency_ms\":52243,\"pass_rate\":0.7185,\"provider_group\":\"provider_y\",\"quality\":0.8346,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.833,\"estimated_latency_ms\":93802,\"estimated_pass_rate\":0.6249,\"estimated_quality\":0.89,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.081,\"max_latency_ms\":178107,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5237, "category": "review", "complexity": 0.5167, "consequence": 0.8752, "evidence_need": 0.8161, "family": "policy_compliance", "input_tokens": 5040, "request_id": "synthetic-policy_compliance-001131", "reversibility": 0.4557, "risk": 0.5094, "summary": "Audit a workflow for access control against explicit access requirements."}, "workers": [{"credits_per_call": 14.004, "id": "worker_a", "p95_latency_ms": 70480, "pass_rate": 0.933, "provider_group": "provider_x", "quality": 0.9156, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.829, "id": "worker_b", "p95_latency_ms": 23322, "pass_rate": 0.4508, "provider_group": "provider_z", "quality": 0.7962, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.355, "id": "worker_c", "p95_latency_ms": 52243, "pass_rate": 0.7185, "provider_group": "provider_y", "quality": 0.8346, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 3.462, "max_latency_ms": 35277, "minimum_quality": 0.4835}, "family": "migration_planning", "id": "synthetic-migration_planning-000655", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.014, "estimated_latency_ms": 27753, "estimated_pass_rate": 0.2808, "estimated_quality": 0.5366, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.462, "max_latency_ms": 35277, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2808, "quality": 0.5366, "quality_target_met": true, "utility": 34.137116}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":3.462,\"max_latency_ms\":35277,\"minimum_quality\":0.4835},\"task\":{\"ambiguity\":0.5601,\"category\":\"coding\",\"complexity\":0.9153,\"consequence\":0.7348,\"evidence_need\":0.6344,\"family\":\"migration_planning\",\"input_tokens\":6640,\"reversibility\":0.6435,\"risk\":0.509,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":14.024,\"id\":\"worker_a\",\"p95_latency_ms\":83515,\"pass_rate\":0.8899,\"provider_group\":\"provider_z\",\"quality\":0.967,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.258,\"id\":\"worker_b\",\"p95_latency_ms\":77456,\"pass_rate\":0.8299,\"provider_group\":\"provider_y\",\"quality\":0.864,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.014,\"id\":\"worker_c\",\"p95_latency_ms\":27753,\"pass_rate\":0.4682,\"provider_group\":\"provider_x\",\"quality\":0.6964,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.014,\"estimated_latency_ms\":27753,\"estimated_pass_rate\":0.2808,\"estimated_quality\":0.5366,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.462,\"max_latency_ms\":35277,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5601, "category": "coding", "complexity": 0.9153, "consequence": 0.7348, "evidence_need": 0.6344, "family": "migration_planning", "input_tokens": 6640, "request_id": "synthetic-migration_planning-000655", "reversibility": 0.6435, "risk": 0.509, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 14.024, "id": "worker_a", "p95_latency_ms": 83515, "pass_rate": 0.8899, "provider_group": "provider_z", "quality": 0.967, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.258, "id": "worker_b", "p95_latency_ms": 77456, "pass_rate": 0.8299, "provider_group": "provider_y", "quality": 0.864, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.014, "id": "worker_c", "p95_latency_ms": 27753, "pass_rate": 0.4682, "provider_group": "provider_x", "quality": 0.6964, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 5.699, "max_latency_ms": 101209, "minimum_quality": 0.8859}, "family": "migration_planning", "id": "synthetic-migration_planning-001042", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.963, "estimated_latency_ms": 85995, "estimated_pass_rate": 0.548, "estimated_quality": 0.7709, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 5.699, "max_latency_ms": 101209, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.548, "quality": 0.7709, "quality_target_met": false, "utility": 46.577139}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":5.699,\"max_latency_ms\":101209,\"minimum_quality\":0.8859},\"task\":{\"ambiguity\":0.3258,\"category\":\"coding\",\"complexity\":0.6263,\"consequence\":0.5823,\"evidence_need\":0.6028,\"family\":\"migration_planning\",\"input_tokens\":4720,\"reversibility\":0.5126,\"risk\":0.4229,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.218,\"id\":\"worker_a\",\"p95_latency_ms\":29998,\"pass_rate\":0.4773,\"provider_group\":\"provider_x\",\"quality\":0.7181,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.745,\"id\":\"worker_b\",\"p95_latency_ms\":55997,\"pass_rate\":0.7659,\"provider_group\":\"provider_y\",\"quality\":0.8001,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.662,\"id\":\"worker_c\",\"p95_latency_ms\":59199,\"pass_rate\":0.991,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.963,\"estimated_latency_ms\":85995,\"estimated_pass_rate\":0.548,\"estimated_quality\":0.7709,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":5.699,\"max_latency_ms\":101209,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3258, "category": "coding", "complexity": 0.6263, "consequence": 0.5823, "evidence_need": 0.6028, "family": "migration_planning", "input_tokens": 4720, "request_id": "synthetic-migration_planning-001042", "reversibility": 0.5126, "risk": 0.4229, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 1.218, "id": "worker_a", "p95_latency_ms": 29998, "pass_rate": 0.4773, "provider_group": "provider_x", "quality": 0.7181, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.745, "id": "worker_b", "p95_latency_ms": 55997, "pass_rate": 0.7659, "provider_group": "provider_y", "quality": 0.8001, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.662, "id": "worker_c", "p95_latency_ms": 59199, "pass_rate": 0.991, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.248, "max_latency_ms": 152367, "minimum_quality": 0.8994}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000441", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 10.757, "estimated_latency_ms": 84370, "estimated_pass_rate": 0.7184, "estimated_quality": 0.8485, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.248, "max_latency_ms": 152367, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "credit_constrained", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7184, "quality": 0.8485, "quality_target_met": false, "utility": 62.229099}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.248,\"max_latency_ms\":152367,\"minimum_quality\":0.8994},\"task\":{\"ambiguity\":0.608,\"category\":\"review\",\"complexity\":0.6052,\"consequence\":0.7416,\"evidence_need\":0.5677,\"family\":\"policy_compliance\",\"input_tokens\":8000,\"reversibility\":0.6009,\"risk\":0.8263,\"summary\":\"Audit a workflow for private credentials against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":9.962,\"id\":\"worker_a\",\"p95_latency_ms\":56719,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9134,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.019,\"id\":\"worker_b\",\"p95_latency_ms\":59175,\"pass_rate\":0.798,\"provider_group\":\"provider_z\",\"quality\":0.8694,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.795,\"id\":\"worker_c\",\"p95_latency_ms\":27651,\"pass_rate\":0.507,\"provider_group\":\"provider_y\",\"quality\":0.7173,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":10.757,\"estimated_latency_ms\":84370,\"estimated_pass_rate\":0.7184,\"estimated_quality\":0.8485,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.248,\"max_latency_ms\":152367,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"credit_constrained\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.608, "category": "review", "complexity": 0.6052, "consequence": 0.7416, "evidence_need": 0.5677, "family": "policy_compliance", "input_tokens": 8000, "request_id": "synthetic-policy_compliance-000441", "reversibility": 0.6009, "risk": 0.8263, "summary": "Audit a workflow for private credentials against explicit access requirements."}, "workers": [{"credits_per_call": 9.962, "id": "worker_a", "p95_latency_ms": 56719, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9134, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.019, "id": "worker_b", "p95_latency_ms": 59175, "pass_rate": 0.798, "provider_group": "provider_z", "quality": 0.8694, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.795, "id": "worker_c", "p95_latency_ms": 27651, "pass_rate": 0.507, "provider_group": "provider_y", "quality": 0.7173, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.23, "max_latency_ms": 30327, "minimum_quality": 0.5871}, "family": "migration_planning", "id": "synthetic-migration_planning-000643", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.833, "estimated_latency_ms": 26623, "estimated_pass_rate": 0.3399, "estimated_quality": 0.5825, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.23, "max_latency_ms": 30327, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3399, "quality": 0.5825, "quality_target_met": false, "utility": 35.048459}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.23,\"max_latency_ms\":30327,\"minimum_quality\":0.5871},\"task\":{\"ambiguity\":0.4237,\"category\":\"coding\",\"complexity\":0.9493,\"consequence\":0.682,\"evidence_need\":0.6295,\"family\":\"migration_planning\",\"input_tokens\":5720,\"reversibility\":0.6119,\"risk\":0.5927,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.793,\"id\":\"worker_a\",\"p95_latency_ms\":49485,\"pass_rate\":0.9013,\"provider_group\":\"provider_y\",\"quality\":0.9451,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.833,\"id\":\"worker_b\",\"p95_latency_ms\":26623,\"pass_rate\":0.5276,\"provider_group\":\"provider_z\",\"quality\":0.7453,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.326,\"id\":\"worker_c\",\"p95_latency_ms\":45998,\"pass_rate\":0.7513,\"provider_group\":\"provider_x\",\"quality\":0.8018,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.833,\"estimated_latency_ms\":26623,\"estimated_pass_rate\":0.3399,\"estimated_quality\":0.5825,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.23,\"max_latency_ms\":30327,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4237, "category": "coding", "complexity": 0.9493, "consequence": 0.682, "evidence_need": 0.6295, "family": "migration_planning", "input_tokens": 5720, "request_id": "synthetic-migration_planning-000643", "reversibility": 0.6119, "risk": 0.5927, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 13.793, "id": "worker_a", "p95_latency_ms": 49485, "pass_rate": 0.9013, "provider_group": "provider_y", "quality": 0.9451, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.833, "id": "worker_b", "p95_latency_ms": 26623, "pass_rate": 0.5276, "provider_group": "provider_z", "quality": 0.7453, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.326, "id": "worker_c", "p95_latency_ms": 45998, "pass_rate": 0.7513, "provider_group": "provider_x", "quality": 0.8018, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.507, "max_latency_ms": 185803, "minimum_quality": 0.7484}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000764", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.038, "estimated_latency_ms": 123822, "estimated_pass_rate": 0.9315, "estimated_quality": 0.9684, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.507, "max_latency_ms": 185803, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9315, "quality": 0.9684, "quality_target_met": true, "utility": 73.69484}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.507,\"max_latency_ms\":185803,\"minimum_quality\":0.7484},\"task\":{\"ambiguity\":0.4027,\"category\":\"analysis\",\"complexity\":0.4504,\"consequence\":0.5843,\"evidence_need\":0.4821,\"family\":\"performance_tuning\",\"input_tokens\":4040,\"reversibility\":0.7108,\"risk\":0.379,\"summary\":\"Diagnose latency variance in background jobs and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":2.868,\"id\":\"worker_a\",\"p95_latency_ms\":61027,\"pass_rate\":0.9228,\"provider_group\":\"provider_x\",\"quality\":0.926,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.857,\"id\":\"worker_b\",\"p95_latency_ms\":28398,\"pass_rate\":0.4506,\"provider_group\":\"provider_y\",\"quality\":0.6975,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.17,\"id\":\"worker_c\",\"p95_latency_ms\":62795,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.038,\"estimated_latency_ms\":123822,\"estimated_pass_rate\":0.9315,\"estimated_quality\":0.9684,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.507,\"max_latency_ms\":185803,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4027, "category": "analysis", "complexity": 0.4504, "consequence": 0.5843, "evidence_need": 0.4821, "family": "performance_tuning", "input_tokens": 4040, "request_id": "synthetic-performance_tuning-000764", "reversibility": 0.7108, "risk": 0.379, "summary": "Diagnose latency variance in background jobs and rank the next experiments."}, "workers": [{"credits_per_call": 2.868, "id": "worker_a", "p95_latency_ms": 61027, "pass_rate": 0.9228, "provider_group": "provider_x", "quality": 0.926, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.857, "id": "worker_b", "p95_latency_ms": 28398, "pass_rate": 0.4506, "provider_group": "provider_y", "quality": 0.6975, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.17, "id": "worker_c", "p95_latency_ms": 62795, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 26.51, "max_latency_ms": 186541, "minimum_quality": 0.8065}, "family": "migration_planning", "id": "synthetic-migration_planning-000403", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.943, "estimated_latency_ms": 128950, "estimated_pass_rate": 0.7643, "estimated_quality": 0.8718, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 26.51, "max_latency_ms": 186541, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7643, "quality": 0.8718, "quality_target_met": true, "utility": 76.161819}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":26.51,\"max_latency_ms\":186541,\"minimum_quality\":0.8065},\"task\":{\"ambiguity\":0.5333,\"category\":\"coding\",\"complexity\":0.943,\"consequence\":0.7806,\"evidence_need\":0.5719,\"family\":\"migration_planning\",\"input_tokens\":6720,\"reversibility\":0.4597,\"risk\":0.6281,\"summary\":\"Plan a zero-downtime migration of database failover with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.031,\"id\":\"worker_a\",\"p95_latency_ms\":46147,\"pass_rate\":0.8673,\"provider_group\":\"provider_z\",\"quality\":0.8813,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.983,\"id\":\"worker_b\",\"p95_latency_ms\":50017,\"pass_rate\":0.939,\"provider_group\":\"provider_y\",\"quality\":0.9149,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.929,\"id\":\"worker_c\",\"p95_latency_ms\":32786,\"pass_rate\":0.5884,\"provider_group\":\"provider_x\",\"quality\":0.7164,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.943,\"estimated_latency_ms\":128950,\"estimated_pass_rate\":0.7643,\"estimated_quality\":0.8718,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":26.51,\"max_latency_ms\":186541,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5333, "category": "coding", "complexity": 0.943, "consequence": 0.7806, "evidence_need": 0.5719, "family": "migration_planning", "input_tokens": 6720, "request_id": "synthetic-migration_planning-000403", "reversibility": 0.4597, "risk": 0.6281, "summary": "Plan a zero-downtime migration of database failover with rollback checkpoints."}, "workers": [{"credits_per_call": 4.031, "id": "worker_a", "p95_latency_ms": 46147, "pass_rate": 0.8673, "provider_group": "provider_z", "quality": 0.8813, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.983, "id": "worker_b", "p95_latency_ms": 50017, "pass_rate": 0.939, "provider_group": "provider_y", "quality": 0.9149, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.929, "id": "worker_c", "p95_latency_ms": 32786, "pass_rate": 0.5884, "provider_group": "provider_x", "quality": 0.7164, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.534, "max_latency_ms": 213551, "minimum_quality": 0.7752}, "family": "migration_planning", "id": "synthetic-migration_planning-000106", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.992, "estimated_latency_ms": 113516, "estimated_pass_rate": 0.8423, "estimated_quality": 0.9263, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.534, "max_latency_ms": 213551, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8423, "quality": 0.9263, "quality_target_met": true, "utility": 70.658699}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.534,\"max_latency_ms\":213551,\"minimum_quality\":0.7752},\"task\":{\"ambiguity\":0.5446,\"category\":\"coding\",\"complexity\":0.8232,\"consequence\":0.675,\"evidence_need\":0.8457,\"family\":\"migration_planning\",\"input_tokens\":4360,\"reversibility\":0.2706,\"risk\":0.5565,\"summary\":\"Review a staged migration for OAuth callbacks and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":14.312,\"id\":\"worker_a\",\"p95_latency_ms\":56625,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9109,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.133,\"id\":\"worker_b\",\"p95_latency_ms\":29748,\"pass_rate\":0.55,\"provider_group\":\"provider_z\",\"quality\":0.7303,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.68,\"id\":\"worker_c\",\"p95_latency_ms\":56891,\"pass_rate\":0.8657,\"provider_group\":\"provider_y\",\"quality\":0.9168,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.992,\"estimated_latency_ms\":113516,\"estimated_pass_rate\":0.8423,\"estimated_quality\":0.9263,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.534,\"max_latency_ms\":213551,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5446, "category": "coding", "complexity": 0.8232, "consequence": 0.675, "evidence_need": 0.8457, "family": "migration_planning", "input_tokens": 4360, "request_id": "synthetic-migration_planning-000106", "reversibility": 0.2706, "risk": 0.5565, "summary": "Review a staged migration for OAuth callbacks and identify irreversible steps."}, "workers": [{"credits_per_call": 14.312, "id": "worker_a", "p95_latency_ms": 56625, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9109, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.133, "id": "worker_b", "p95_latency_ms": 29748, "pass_rate": 0.55, "provider_group": "provider_z", "quality": 0.7303, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.68, "id": "worker_c", "p95_latency_ms": 56891, "pass_rate": 0.8657, "provider_group": "provider_y", "quality": 0.9168, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.936, "max_latency_ms": 126246, "minimum_quality": 0.7621}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001083", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.53, "estimated_latency_ms": 100655, "estimated_pass_rate": 0.7365, "estimated_quality": 0.8883, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.936, "max_latency_ms": 126246, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7365, "quality": 0.8883, "quality_target_met": true, "utility": 65.611195}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.936,\"max_latency_ms\":126246,\"minimum_quality\":0.7621},\"task\":{\"ambiguity\":0.7957,\"category\":\"review\",\"complexity\":0.5682,\"consequence\":0.5693,\"evidence_need\":0.7439,\"family\":\"policy_compliance\",\"input_tokens\":2000,\"reversibility\":0.3268,\"risk\":0.8237,\"summary\":\"Identify evidence required before approving data retention for production.\"},\"workers\":[{\"credits_per_call\":3.594,\"id\":\"worker_a\",\"p95_latency_ms\":61563,\"pass_rate\":0.7785,\"provider_group\":\"provider_y\",\"quality\":0.8596,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.563,\"id\":\"worker_b\",\"p95_latency_ms\":73216,\"pass_rate\":0.9501,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.967,\"id\":\"worker_c\",\"p95_latency_ms\":27439,\"pass_rate\":0.5241,\"provider_group\":\"provider_z\",\"quality\":0.7348,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.53,\"estimated_latency_ms\":100655,\"estimated_pass_rate\":0.7365,\"estimated_quality\":0.8883,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.936,\"max_latency_ms\":126246,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7957, "category": "review", "complexity": 0.5682, "consequence": 0.5693, "evidence_need": 0.7439, "family": "policy_compliance", "input_tokens": 2000, "request_id": "synthetic-policy_compliance-001083", "reversibility": 0.3268, "risk": 0.8237, "summary": "Identify evidence required before approving data retention for production."}, "workers": [{"credits_per_call": 3.594, "id": "worker_a", "p95_latency_ms": 61563, "pass_rate": 0.7785, "provider_group": "provider_y", "quality": 0.8596, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 11.563, "id": "worker_b", "p95_latency_ms": 73216, "pass_rate": 0.9501, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.967, "id": "worker_c", "p95_latency_ms": 27439, "pass_rate": 0.5241, "provider_group": "provider_z", "quality": 0.7348, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.78, "max_latency_ms": 42064, "minimum_quality": 0.5785}, "family": "migration_planning", "id": "synthetic-migration_planning-000121", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.927, "estimated_latency_ms": 23800, "estimated_pass_rate": 0.3558, "estimated_quality": 0.6271, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.78, "max_latency_ms": 42064, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3558, "quality": 0.6271, "quality_target_met": true, "utility": 40.871974}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.78,\"max_latency_ms\":42064,\"minimum_quality\":0.5785},\"task\":{\"ambiguity\":0.322,\"category\":\"coding\",\"complexity\":0.7828,\"consequence\":0.4391,\"evidence_need\":0.5933,\"family\":\"migration_planning\",\"input_tokens\":5400,\"reversibility\":0.4697,\"risk\":0.3595,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":3.921,\"id\":\"worker_a\",\"p95_latency_ms\":55374,\"pass_rate\":0.7934,\"provider_group\":\"provider_y\",\"quality\":0.9022,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.198,\"id\":\"worker_b\",\"p95_latency_ms\":88211,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.9695,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.927,\"id\":\"worker_c\",\"p95_latency_ms\":23800,\"pass_rate\":0.498,\"provider_group\":\"provider_x\",\"quality\":0.7362,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.927,\"estimated_latency_ms\":23800,\"estimated_pass_rate\":0.3558,\"estimated_quality\":0.6271,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.78,\"max_latency_ms\":42064,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.322, "category": "coding", "complexity": 0.7828, "consequence": 0.4391, "evidence_need": 0.5933, "family": "migration_planning", "input_tokens": 5400, "request_id": "synthetic-migration_planning-000121", "reversibility": 0.4697, "risk": 0.3595, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 3.921, "id": "worker_a", "p95_latency_ms": 55374, "pass_rate": 0.7934, "provider_group": "provider_y", "quality": 0.9022, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.198, "id": "worker_b", "p95_latency_ms": 88211, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.9695, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.927, "id": "worker_c", "p95_latency_ms": 23800, "pass_rate": 0.498, "provider_group": "provider_x", "quality": 0.7362, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.648, "max_latency_ms": 224054, "minimum_quality": 0.8327}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000123", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.853, "estimated_latency_ms": 113379, "estimated_pass_rate": 0.8056, "estimated_quality": 0.9349, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.648, "max_latency_ms": 224054, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8056, "quality": 0.9349, "quality_target_met": true, "utility": 73.750834}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.648,\"max_latency_ms\":224054,\"minimum_quality\":0.8327},\"task\":{\"ambiguity\":0.7773,\"category\":\"review\",\"complexity\":0.7625,\"consequence\":0.7792,\"evidence_need\":0.9128,\"family\":\"policy_compliance\",\"input_tokens\":5680,\"reversibility\":0.5571,\"risk\":0.7043,\"summary\":\"Identify evidence required before approving streaming inference for production.\"},\"workers\":[{\"credits_per_call\":10.778,\"id\":\"worker_a\",\"p95_latency_ms\":63632,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.075,\"id\":\"worker_b\",\"p95_latency_ms\":49747,\"pass_rate\":0.724,\"provider_group\":\"provider_x\",\"quality\":0.8428,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.971,\"id\":\"worker_c\",\"p95_latency_ms\":31995,\"pass_rate\":0.5957,\"provider_group\":\"provider_z\",\"quality\":0.7434,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.853,\"estimated_latency_ms\":113379,\"estimated_pass_rate\":0.8056,\"estimated_quality\":0.9349,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.648,\"max_latency_ms\":224054,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7773, "category": "review", "complexity": 0.7625, "consequence": 0.7792, "evidence_need": 0.9128, "family": "policy_compliance", "input_tokens": 5680, "request_id": "synthetic-policy_compliance-000123", "reversibility": 0.5571, "risk": 0.7043, "summary": "Identify evidence required before approving streaming inference for production."}, "workers": [{"credits_per_call": 10.778, "id": "worker_a", "p95_latency_ms": 63632, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.075, "id": "worker_b", "p95_latency_ms": 49747, "pass_rate": 0.724, "provider_group": "provider_x", "quality": 0.8428, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.971, "id": "worker_c", "p95_latency_ms": 31995, "pass_rate": 0.5957, "provider_group": "provider_z", "quality": 0.7434, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.461, "max_latency_ms": 140226, "minimum_quality": 0.8257}, "family": "migration_planning", "id": "synthetic-migration_planning-001054", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.949, "estimated_latency_ms": 79896, "estimated_pass_rate": 0.5765, "estimated_quality": 0.7654, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.461, "max_latency_ms": 140226, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5765, "quality": 0.7654, "quality_target_met": false, "utility": 55.737822}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.461,\"max_latency_ms\":140226,\"minimum_quality\":0.8257},\"task\":{\"ambiguity\":0.7352,\"category\":\"coding\",\"complexity\":0.8884,\"consequence\":0.5311,\"evidence_need\":0.8039,\"family\":\"migration_planning\",\"input_tokens\":5760,\"reversibility\":0.276,\"risk\":0.7217,\"summary\":\"Design compatibility testing for a database failover schema migration.\"},\"workers\":[{\"credits_per_call\":15.327,\"id\":\"worker_a\",\"p95_latency_ms\":92561,\"pass_rate\":0.971,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.114,\"id\":\"worker_b\",\"p95_latency_ms\":27721,\"pass_rate\":0.5391,\"provider_group\":\"provider_y\",\"quality\":0.7428,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.835,\"id\":\"worker_c\",\"p95_latency_ms\":52175,\"pass_rate\":0.7626,\"provider_group\":\"provider_x\",\"quality\":0.8118,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.949,\"estimated_latency_ms\":79896,\"estimated_pass_rate\":0.5765,\"estimated_quality\":0.7654,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.461,\"max_latency_ms\":140226,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7352, "category": "coding", "complexity": 0.8884, "consequence": 0.5311, "evidence_need": 0.8039, "family": "migration_planning", "input_tokens": 5760, "request_id": "synthetic-migration_planning-001054", "reversibility": 0.276, "risk": 0.7217, "summary": "Design compatibility testing for a database failover schema migration."}, "workers": [{"credits_per_call": 15.327, "id": "worker_a", "p95_latency_ms": 92561, "pass_rate": 0.971, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.114, "id": "worker_b", "p95_latency_ms": 27721, "pass_rate": 0.5391, "provider_group": "provider_y", "quality": 0.7428, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.835, "id": "worker_c", "p95_latency_ms": 52175, "pass_rate": 0.7626, "provider_group": "provider_x", "quality": 0.8118, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.395, "max_latency_ms": 29995, "minimum_quality": 0.7524}, "family": "migration_planning", "id": "synthetic-migration_planning-000784", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.075, "estimated_latency_ms": 27215, "estimated_pass_rate": 0.3455, "estimated_quality": 0.6422, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.395, "max_latency_ms": 29995, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3455, "quality": 0.6422, "quality_target_met": false, "utility": 35.909607}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.395,\"max_latency_ms\":29995,\"minimum_quality\":0.7524},\"task\":{\"ambiguity\":0.3092,\"category\":\"coding\",\"complexity\":0.9455,\"consequence\":0.6125,\"evidence_need\":0.838,\"family\":\"migration_planning\",\"input_tokens\":520,\"reversibility\":0.561,\"risk\":0.35,\"summary\":\"Design compatibility testing for a model routing schema migration.\"},\"workers\":[{\"credits_per_call\":4.197,\"id\":\"worker_a\",\"p95_latency_ms\":78672,\"pass_rate\":0.7793,\"provider_group\":\"provider_z\",\"quality\":0.9076,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.207,\"id\":\"worker_b\",\"p95_latency_ms\":72670,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.075,\"id\":\"worker_c\",\"p95_latency_ms\":27215,\"pass_rate\":0.5044,\"provider_group\":\"provider_x\",\"quality\":0.7791,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.075,\"estimated_latency_ms\":27215,\"estimated_pass_rate\":0.3455,\"estimated_quality\":0.6422,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.395,\"max_latency_ms\":29995,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3092, "category": "coding", "complexity": 0.9455, "consequence": 0.6125, "evidence_need": 0.838, "family": "migration_planning", "input_tokens": 520, "request_id": "synthetic-migration_planning-000784", "reversibility": 0.561, "risk": 0.35, "summary": "Design compatibility testing for a model routing schema migration."}, "workers": [{"credits_per_call": 4.197, "id": "worker_a", "p95_latency_ms": 78672, "pass_rate": 0.7793, "provider_group": "provider_z", "quality": 0.9076, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.207, "id": "worker_b", "p95_latency_ms": 72670, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.075, "id": "worker_c", "p95_latency_ms": 27215, "pass_rate": 0.5044, "provider_group": "provider_x", "quality": 0.7791, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.584, "max_latency_ms": 37820, "minimum_quality": 0.5001}, "family": "migration_planning", "id": "synthetic-migration_planning-000289", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.155, "estimated_latency_ms": 23376, "estimated_pass_rate": 0.2934, "estimated_quality": 0.5741, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.584, "max_latency_ms": 37820, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2934, "quality": 0.5741, "quality_target_met": true, "utility": 34.456848}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.584,\"max_latency_ms\":37820,\"minimum_quality\":0.5001},\"task\":{\"ambiguity\":0.594,\"category\":\"coding\",\"complexity\":0.6131,\"consequence\":0.7135,\"evidence_need\":0.8412,\"family\":\"migration_planning\",\"input_tokens\":7440,\"reversibility\":0.4788,\"risk\":0.5957,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.908,\"id\":\"worker_a\",\"p95_latency_ms\":76991,\"pass_rate\":0.8205,\"provider_group\":\"provider_x\",\"quality\":0.8672,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.474,\"id\":\"worker_b\",\"p95_latency_ms\":66346,\"pass_rate\":0.9534,\"provider_group\":\"provider_z\",\"quality\":0.9572,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.155,\"id\":\"worker_c\",\"p95_latency_ms\":23376,\"pass_rate\":0.4561,\"provider_group\":\"provider_y\",\"quality\":0.6878,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.155,\"estimated_latency_ms\":23376,\"estimated_pass_rate\":0.2934,\"estimated_quality\":0.5741,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.584,\"max_latency_ms\":37820,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.594, "category": "coding", "complexity": 0.6131, "consequence": 0.7135, "evidence_need": 0.8412, "family": "migration_planning", "input_tokens": 7440, "request_id": "synthetic-migration_planning-000289", "reversibility": 0.4788, "risk": 0.5957, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 2.908, "id": "worker_a", "p95_latency_ms": 76991, "pass_rate": 0.8205, "provider_group": "provider_x", "quality": 0.8672, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.474, "id": "worker_b", "p95_latency_ms": 66346, "pass_rate": 0.9534, "provider_group": "provider_z", "quality": 0.9572, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.155, "id": "worker_c", "p95_latency_ms": 23376, "pass_rate": 0.4561, "provider_group": "provider_y", "quality": 0.6878, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 20.911, "max_latency_ms": 208084, "minimum_quality": 0.9116}, "family": "migration_planning", "id": "synthetic-migration_planning-000979", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.167, "estimated_latency_ms": 155280, "estimated_pass_rate": 0.7113, "estimated_quality": 0.8982, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 20.911, "max_latency_ms": 208084, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "high_consequence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7113, "quality": 0.8982, "quality_target_met": false, "utility": 68.850298}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":20.911,\"max_latency_ms\":208084,\"minimum_quality\":0.9116},\"task\":{\"ambiguity\":0.4614,\"category\":\"coding\",\"complexity\":0.7238,\"consequence\":0.8007,\"evidence_need\":0.4286,\"family\":\"migration_planning\",\"input_tokens\":5360,\"reversibility\":0.2517,\"risk\":0.4812,\"summary\":\"Design compatibility testing for a API-key rotation schema migration.\"},\"workers\":[{\"credits_per_call\":3.579,\"id\":\"worker_a\",\"p95_latency_ms\":67981,\"pass_rate\":0.8082,\"provider_group\":\"provider_z\",\"quality\":0.8074,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.098,\"id\":\"worker_b\",\"p95_latency_ms\":20293,\"pass_rate\":0.5208,\"provider_group\":\"provider_x\",\"quality\":0.7643,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.49,\"id\":\"worker_c\",\"p95_latency_ms\":67006,\"pass_rate\":0.9237,\"provider_group\":\"provider_y\",\"quality\":0.9681,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.167,\"estimated_latency_ms\":155280,\"estimated_pass_rate\":0.7113,\"estimated_quality\":0.8982,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":20.911,\"max_latency_ms\":208084,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"high_consequence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4614, "category": "coding", "complexity": 0.7238, "consequence": 0.8007, "evidence_need": 0.4286, "family": "migration_planning", "input_tokens": 5360, "request_id": "synthetic-migration_planning-000979", "reversibility": 0.2517, "risk": 0.4812, "summary": "Design compatibility testing for a API-key rotation schema migration."}, "workers": [{"credits_per_call": 3.579, "id": "worker_a", "p95_latency_ms": 67981, "pass_rate": 0.8082, "provider_group": "provider_z", "quality": 0.8074, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.098, "id": "worker_b", "p95_latency_ms": 20293, "pass_rate": 0.5208, "provider_group": "provider_x", "quality": 0.7643, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.49, "id": "worker_c", "p95_latency_ms": 67006, "pass_rate": 0.9237, "provider_group": "provider_y", "quality": 0.9681, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.505, "max_latency_ms": 176436, "minimum_quality": 0.8347}, "family": "migration_planning", "id": "synthetic-migration_planning-000697", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.103, "estimated_latency_ms": 142586, "estimated_pass_rate": 0.7128, "estimated_quality": 0.918, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.505, "max_latency_ms": 176436, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7128, "quality": 0.918, "quality_target_met": true, "utility": 74.265911}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.505,\"max_latency_ms\":176436,\"minimum_quality\":0.8347},\"task\":{\"ambiguity\":0.4346,\"category\":\"coding\",\"complexity\":0.8949,\"consequence\":0.7329,\"evidence_need\":0.8269,\"family\":\"migration_planning\",\"input_tokens\":7120,\"reversibility\":0.6462,\"risk\":0.4828,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":11.845,\"id\":\"worker_a\",\"p95_latency_ms\":77287,\"pass_rate\":0.9237,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.958,\"id\":\"worker_b\",\"p95_latency_ms\":20774,\"pass_rate\":0.5153,\"provider_group\":\"provider_x\",\"quality\":0.7301,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.3,\"id\":\"worker_c\",\"p95_latency_ms\":44525,\"pass_rate\":0.7771,\"provider_group\":\"provider_y\",\"quality\":0.903,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.103,\"estimated_latency_ms\":142586,\"estimated_pass_rate\":0.7128,\"estimated_quality\":0.918,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.505,\"max_latency_ms\":176436,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4346, "category": "coding", "complexity": 0.8949, "consequence": 0.7329, "evidence_need": 0.8269, "family": "migration_planning", "input_tokens": 7120, "request_id": "synthetic-migration_planning-000697", "reversibility": 0.6462, "risk": 0.4828, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 11.845, "id": "worker_a", "p95_latency_ms": 77287, "pass_rate": 0.9237, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.958, "id": "worker_b", "p95_latency_ms": 20774, "pass_rate": 0.5153, "provider_group": "provider_x", "quality": 0.7301, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.3, "id": "worker_c", "p95_latency_ms": 44525, "pass_rate": 0.7771, "provider_group": "provider_y", "quality": 0.903, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.748, "max_latency_ms": 27545, "minimum_quality": 0.5198}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000447", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.82, "estimated_latency_ms": 24425, "estimated_pass_rate": 0.2662, "estimated_quality": 0.5418, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.748, "max_latency_ms": 27545, "policy": "lean", "reason_codes": ["complex_task", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2662, "quality": 0.5418, "quality_target_met": true, "utility": 33.63512}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.748,\"max_latency_ms\":27545,\"minimum_quality\":0.5198},\"task\":{\"ambiguity\":0.6752,\"category\":\"review\",\"complexity\":0.8028,\"consequence\":0.5109,\"evidence_need\":0.7618,\"family\":\"policy_compliance\",\"input_tokens\":4320,\"reversibility\":0.5064,\"risk\":0.5686,\"summary\":\"Review whether model routing satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.82,\"id\":\"worker_a\",\"p95_latency_ms\":24425,\"pass_rate\":0.4543,\"provider_group\":\"provider_y\",\"quality\":0.6933,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.365,\"id\":\"worker_b\",\"p95_latency_ms\":74565,\"pass_rate\":0.7671,\"provider_group\":\"provider_z\",\"quality\":0.8921,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.88,\"id\":\"worker_c\",\"p95_latency_ms\":70752,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.82,\"estimated_latency_ms\":24425,\"estimated_pass_rate\":0.2662,\"estimated_quality\":0.5418,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.748,\"max_latency_ms\":27545,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6752, "category": "review", "complexity": 0.8028, "consequence": 0.5109, "evidence_need": 0.7618, "family": "policy_compliance", "input_tokens": 4320, "request_id": "synthetic-policy_compliance-000447", "reversibility": 0.5064, "risk": 0.5686, "summary": "Review whether model routing satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.82, "id": "worker_a", "p95_latency_ms": 24425, "pass_rate": 0.4543, "provider_group": "provider_y", "quality": 0.6933, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.365, "id": "worker_b", "p95_latency_ms": 74565, "pass_rate": 0.7671, "provider_group": "provider_z", "quality": 0.8921, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.88, "id": "worker_c", "p95_latency_ms": 70752, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.193, "max_latency_ms": 185692, "minimum_quality": 0.8668}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000384", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.011, "estimated_latency_ms": 175089, "estimated_pass_rate": 0.7692, "estimated_quality": 0.9131, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.193, "max_latency_ms": 185692, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7692, "quality": 0.9131, "quality_target_met": true, "utility": 74.915069}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.193,\"max_latency_ms\":185692,\"minimum_quality\":0.8668},\"task\":{\"ambiguity\":0.7147,\"category\":\"review\",\"complexity\":0.4151,\"consequence\":0.9335,\"evidence_need\":0.8794,\"family\":\"policy_compliance\",\"input_tokens\":1960,\"reversibility\":0.4529,\"risk\":0.5963,\"summary\":\"Audit a workflow for credit reservation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.176,\"id\":\"worker_a\",\"p95_latency_ms\":24068,\"pass_rate\":0.5906,\"provider_group\":\"provider_z\",\"quality\":0.7834,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.731,\"id\":\"worker_b\",\"p95_latency_ms\":68319,\"pass_rate\":0.7323,\"provider_group\":\"provider_x\",\"quality\":0.9219,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.104,\"id\":\"worker_c\",\"p95_latency_ms\":82702,\"pass_rate\":0.9591,\"provider_group\":\"provider_y\",\"quality\":0.9425,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.011,\"estimated_latency_ms\":175089,\"estimated_pass_rate\":0.7692,\"estimated_quality\":0.9131,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.193,\"max_latency_ms\":185692,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7147, "category": "review", "complexity": 0.4151, "consequence": 0.9335, "evidence_need": 0.8794, "family": "policy_compliance", "input_tokens": 1960, "request_id": "synthetic-policy_compliance-000384", "reversibility": 0.4529, "risk": 0.5963, "summary": "Audit a workflow for credit reservation against explicit access requirements."}, "workers": [{"credits_per_call": 1.176, "id": "worker_a", "p95_latency_ms": 24068, "pass_rate": 0.5906, "provider_group": "provider_z", "quality": 0.7834, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.731, "id": "worker_b", "p95_latency_ms": 68319, "pass_rate": 0.7323, "provider_group": "provider_x", "quality": 0.9219, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.104, "id": "worker_c", "p95_latency_ms": 82702, "pass_rate": 0.9591, "provider_group": "provider_y", "quality": 0.9425, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.122, "max_latency_ms": 168495, "minimum_quality": 0.8976}, "family": "migration_planning", "id": "synthetic-migration_planning-000298", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.538, "estimated_latency_ms": 139096, "estimated_pass_rate": 0.8717, "estimated_quality": 0.931, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.122, "max_latency_ms": 168495, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8717, "quality": 0.931, "quality_target_met": true, "utility": 69.529284}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.122,\"max_latency_ms\":168495,\"minimum_quality\":0.8976},\"task\":{\"ambiguity\":0.6302,\"category\":\"coding\",\"complexity\":0.5585,\"consequence\":0.6128,\"evidence_need\":0.6792,\"family\":\"migration_planning\",\"input_tokens\":3800,\"reversibility\":0.4019,\"risk\":0.4811,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":13.791,\"id\":\"worker_a\",\"p95_latency_ms\":74490,\"pass_rate\":0.9415,\"provider_group\":\"provider_x\",\"quality\":0.9648,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.108,\"id\":\"worker_b\",\"p95_latency_ms\":22928,\"pass_rate\":0.5266,\"provider_group\":\"provider_y\",\"quality\":0.7327,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.747,\"id\":\"worker_c\",\"p95_latency_ms\":64606,\"pass_rate\":0.8447,\"provider_group\":\"provider_z\",\"quality\":0.8799,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.538,\"estimated_latency_ms\":139096,\"estimated_pass_rate\":0.8717,\"estimated_quality\":0.931,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.122,\"max_latency_ms\":168495,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6302, "category": "coding", "complexity": 0.5585, "consequence": 0.6128, "evidence_need": 0.6792, "family": "migration_planning", "input_tokens": 3800, "request_id": "synthetic-migration_planning-000298", "reversibility": 0.4019, "risk": 0.4811, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 13.791, "id": "worker_a", "p95_latency_ms": 74490, "pass_rate": 0.9415, "provider_group": "provider_x", "quality": 0.9648, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.108, "id": "worker_b", "p95_latency_ms": 22928, "pass_rate": 0.5266, "provider_group": "provider_y", "quality": 0.7327, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.747, "id": "worker_c", "p95_latency_ms": 64606, "pass_rate": 0.8447, "provider_group": "provider_z", "quality": 0.8799, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.252, "max_latency_ms": 218434, "minimum_quality": 0.8375}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001106", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.233, "estimated_latency_ms": 154798, "estimated_pass_rate": 0.7293, "estimated_quality": 0.923, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.252, "max_latency_ms": 218434, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7293, "quality": 0.923, "quality_target_met": true, "utility": 73.404509}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.252,\"max_latency_ms\":218434,\"minimum_quality\":0.8375},\"task\":{\"ambiguity\":0.7076,\"category\":\"analysis\",\"complexity\":0.813,\"consequence\":0.4462,\"evidence_need\":0.6906,\"family\":\"performance_tuning\",\"input_tokens\":6120,\"reversibility\":0.5821,\"risk\":0.4272,\"summary\":\"Review a proposed optimization for private credentials for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.028,\"id\":\"worker_a\",\"p95_latency_ms\":32317,\"pass_rate\":0.4623,\"provider_group\":\"provider_x\",\"quality\":0.7149,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.284,\"id\":\"worker_b\",\"p95_latency_ms\":68560,\"pass_rate\":0.7977,\"provider_group\":\"provider_z\",\"quality\":0.9832,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.921,\"id\":\"worker_c\",\"p95_latency_ms\":53921,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.233,\"estimated_latency_ms\":154798,\"estimated_pass_rate\":0.7293,\"estimated_quality\":0.923,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.252,\"max_latency_ms\":218434,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7076, "category": "analysis", "complexity": 0.813, "consequence": 0.4462, "evidence_need": 0.6906, "family": "performance_tuning", "input_tokens": 6120, "request_id": "synthetic-performance_tuning-001106", "reversibility": 0.5821, "risk": 0.4272, "summary": "Review a proposed optimization for private credentials for hidden regressions."}, "workers": [{"credits_per_call": 1.028, "id": "worker_a", "p95_latency_ms": 32317, "pass_rate": 0.4623, "provider_group": "provider_x", "quality": 0.7149, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.284, "id": "worker_b", "p95_latency_ms": 68560, "pass_rate": 0.7977, "provider_group": "provider_z", "quality": 0.9832, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.921, "id": "worker_c", "p95_latency_ms": 53921, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.399, "max_latency_ms": 265063, "minimum_quality": 0.6995}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001263", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.477, "estimated_latency_ms": 97752, "estimated_pass_rate": 0.5682, "estimated_quality": 0.8072, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.399, "max_latency_ms": 265063, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5682, "quality": 0.8072, "quality_target_met": true, "utility": 61.816595}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.399,\"max_latency_ms\":265063,\"minimum_quality\":0.6995},\"task\":{\"ambiguity\":0.3584,\"category\":\"review\",\"complexity\":0.6396,\"consequence\":0.7404,\"evidence_need\":0.732,\"family\":\"policy_compliance\",\"input_tokens\":3840,\"reversibility\":0.6609,\"risk\":0.7803,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":0.935,\"id\":\"worker_a\",\"p95_latency_ms\":30606,\"pass_rate\":0.4509,\"provider_group\":\"provider_x\",\"quality\":0.7554,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.31,\"id\":\"worker_b\",\"p95_latency_ms\":90892,\"pass_rate\":0.9425,\"provider_group\":\"provider_z\",\"quality\":0.9919,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.542,\"id\":\"worker_c\",\"p95_latency_ms\":67146,\"pass_rate\":0.7517,\"provider_group\":\"provider_y\",\"quality\":0.839,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.477,\"estimated_latency_ms\":97752,\"estimated_pass_rate\":0.5682,\"estimated_quality\":0.8072,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.399,\"max_latency_ms\":265063,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3584, "category": "review", "complexity": 0.6396, "consequence": 0.7404, "evidence_need": 0.732, "family": "policy_compliance", "input_tokens": 3840, "request_id": "synthetic-policy_compliance-001263", "reversibility": 0.6609, "risk": 0.7803, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 0.935, "id": "worker_a", "p95_latency_ms": 30606, "pass_rate": 0.4509, "provider_group": "provider_x", "quality": 0.7554, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.31, "id": "worker_b", "p95_latency_ms": 90892, "pass_rate": 0.9425, "provider_group": "provider_z", "quality": 0.9919, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.542, "id": "worker_c", "p95_latency_ms": 67146, "pass_rate": 0.7517, "provider_group": "provider_y", "quality": 0.839, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.28, "max_latency_ms": 23743, "minimum_quality": 0.5689}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000246", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.129, "estimated_latency_ms": 20288, "estimated_pass_rate": 0.4344, "estimated_quality": 0.6489, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.28, "max_latency_ms": 23743, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "latency_constrained"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4344, "quality": 0.6489, "quality_target_met": true, "utility": 27.696661}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.28,\"max_latency_ms\":23743,\"minimum_quality\":0.5689},\"task\":{\"ambiguity\":0.7775,\"category\":\"review\",\"complexity\":0.6164,\"consequence\":0.9278,\"evidence_need\":0.9419,\"family\":\"policy_compliance\",\"input_tokens\":6680,\"reversibility\":0.4176,\"risk\":0.6014,\"summary\":\"Audit a workflow for database failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.125,\"id\":\"worker_a\",\"p95_latency_ms\":64730,\"pass_rate\":0.8427,\"provider_group\":\"provider_x\",\"quality\":0.8825,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.129,\"id\":\"worker_b\",\"p95_latency_ms\":20288,\"pass_rate\":0.6118,\"provider_group\":\"provider_z\",\"quality\":0.7759,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.773,\"id\":\"worker_c\",\"p95_latency_ms\":49240,\"pass_rate\":0.944,\"provider_group\":\"provider_y\",\"quality\":0.9473,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.129,\"estimated_latency_ms\":20288,\"estimated_pass_rate\":0.4344,\"estimated_quality\":0.6489,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.28,\"max_latency_ms\":23743,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7775, "category": "review", "complexity": 0.6164, "consequence": 0.9278, "evidence_need": 0.9419, "family": "policy_compliance", "input_tokens": 6680, "request_id": "synthetic-policy_compliance-000246", "reversibility": 0.4176, "risk": 0.6014, "summary": "Audit a workflow for database failover against explicit access requirements."}, "workers": [{"credits_per_call": 3.125, "id": "worker_a", "p95_latency_ms": 64730, "pass_rate": 0.8427, "provider_group": "provider_x", "quality": 0.8825, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.129, "id": "worker_b", "p95_latency_ms": 20288, "pass_rate": 0.6118, "provider_group": "provider_z", "quality": 0.7759, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.773, "id": "worker_c", "p95_latency_ms": 49240, "pass_rate": 0.944, "provider_group": "provider_y", "quality": 0.9473, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 1.708, "max_latency_ms": 36841, "minimum_quality": 0.5267}, "family": "migration_planning", "id": "synthetic-migration_planning-000019", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.21, "estimated_latency_ms": 31905, "estimated_pass_rate": 0.2957, "estimated_quality": 0.6612, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.708, "max_latency_ms": 36841, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2957, "quality": 0.6612, "quality_target_met": true, "utility": 38.939399}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":1.708,\"max_latency_ms\":36841,\"minimum_quality\":0.5267},\"task\":{\"ambiguity\":0.4093,\"category\":\"coding\",\"complexity\":0.822,\"consequence\":0.575,\"evidence_need\":0.5727,\"family\":\"migration_planning\",\"input_tokens\":3360,\"reversibility\":0.4884,\"risk\":0.3869,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":15.164,\"id\":\"worker_a\",\"p95_latency_ms\":57245,\"pass_rate\":0.9718,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.865,\"id\":\"worker_b\",\"p95_latency_ms\":62304,\"pass_rate\":0.8352,\"provider_group\":\"provider_z\",\"quality\":0.9077,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.21,\"id\":\"worker_c\",\"p95_latency_ms\":31905,\"pass_rate\":0.4512,\"provider_group\":\"provider_x\",\"quality\":0.7852,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.21,\"estimated_latency_ms\":31905,\"estimated_pass_rate\":0.2957,\"estimated_quality\":0.6612,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.708,\"max_latency_ms\":36841,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4093, "category": "coding", "complexity": 0.822, "consequence": 0.575, "evidence_need": 0.5727, "family": "migration_planning", "input_tokens": 3360, "request_id": "synthetic-migration_planning-000019", "reversibility": 0.4884, "risk": 0.3869, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 15.164, "id": "worker_a", "p95_latency_ms": 57245, "pass_rate": 0.9718, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.865, "id": "worker_b", "p95_latency_ms": 62304, "pass_rate": 0.8352, "provider_group": "provider_z", "quality": 0.9077, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.21, "id": "worker_c", "p95_latency_ms": 31905, "pass_rate": 0.4512, "provider_group": "provider_x", "quality": 0.7852, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.561, "max_latency_ms": 253191, "minimum_quality": 0.8355}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000117", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.509, "estimated_latency_ms": 184200, "estimated_pass_rate": 0.7837, "estimated_quality": 0.901, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.561, "max_latency_ms": 253191, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7837, "quality": 0.901, "quality_target_met": true, "utility": 73.665545}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.561,\"max_latency_ms\":253191,\"minimum_quality\":0.8355},\"task\":{\"ambiguity\":0.6051,\"category\":\"review\",\"complexity\":0.5442,\"consequence\":0.6151,\"evidence_need\":0.842,\"family\":\"policy_compliance\",\"input_tokens\":1760,\"reversibility\":0.6505,\"risk\":0.5635,\"summary\":\"Identify evidence required before approving provider failover for production.\"},\"workers\":[{\"credits_per_call\":1.011,\"id\":\"worker_a\",\"p95_latency_ms\":23628,\"pass_rate\":0.5321,\"provider_group\":\"provider_z\",\"quality\":0.6938,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.227,\"id\":\"worker_b\",\"p95_latency_ms\":68755,\"pass_rate\":0.8809,\"provider_group\":\"provider_y\",\"quality\":0.9128,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.271,\"id\":\"worker_c\",\"p95_latency_ms\":91817,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9244,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.509,\"estimated_latency_ms\":184200,\"estimated_pass_rate\":0.7837,\"estimated_quality\":0.901,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.561,\"max_latency_ms\":253191,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6051, "category": "review", "complexity": 0.5442, "consequence": 0.6151, "evidence_need": 0.842, "family": "policy_compliance", "input_tokens": 1760, "request_id": "synthetic-policy_compliance-000117", "reversibility": 0.6505, "risk": 0.5635, "summary": "Identify evidence required before approving provider failover for production."}, "workers": [{"credits_per_call": 1.011, "id": "worker_a", "p95_latency_ms": 23628, "pass_rate": 0.5321, "provider_group": "provider_z", "quality": 0.6938, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.227, "id": "worker_b", "p95_latency_ms": 68755, "pass_rate": 0.8809, "provider_group": "provider_y", "quality": 0.9128, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.271, "id": "worker_c", "p95_latency_ms": 91817, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9244, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.791, "max_latency_ms": 219535, "minimum_quality": 0.948}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000069", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.654, "estimated_latency_ms": 173250, "estimated_pass_rate": 0.7118, "estimated_quality": 0.9163, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.791, "max_latency_ms": 219535, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7118, "quality": 0.9163, "quality_target_met": false, "utility": 75.500866}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.791,\"max_latency_ms\":219535,\"minimum_quality\":0.948},\"task\":{\"ambiguity\":0.6766,\"category\":\"review\",\"complexity\":0.823,\"consequence\":0.7412,\"evidence_need\":0.7389,\"family\":\"policy_compliance\",\"input_tokens\":4000,\"reversibility\":0.3331,\"risk\":0.5605,\"summary\":\"Audit a workflow for credit reservation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.404,\"id\":\"worker_a\",\"p95_latency_ms\":85801,\"pass_rate\":0.9674,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.094,\"id\":\"worker_b\",\"p95_latency_ms\":68806,\"pass_rate\":0.83,\"provider_group\":\"provider_z\",\"quality\":0.806,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.156,\"id\":\"worker_c\",\"p95_latency_ms\":18643,\"pass_rate\":0.4868,\"provider_group\":\"provider_x\",\"quality\":0.7476,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.654,\"estimated_latency_ms\":173250,\"estimated_pass_rate\":0.7118,\"estimated_quality\":0.9163,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.791,\"max_latency_ms\":219535,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6766, "category": "review", "complexity": 0.823, "consequence": 0.7412, "evidence_need": 0.7389, "family": "policy_compliance", "input_tokens": 4000, "request_id": "synthetic-policy_compliance-000069", "reversibility": 0.3331, "risk": 0.5605, "summary": "Audit a workflow for credit reservation against explicit access requirements."}, "workers": [{"credits_per_call": 10.404, "id": "worker_a", "p95_latency_ms": 85801, "pass_rate": 0.9674, "provider_group": "provider_y", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.094, "id": "worker_b", "p95_latency_ms": 68806, "pass_rate": 0.83, "provider_group": "provider_z", "quality": 0.806, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.156, "id": "worker_c", "p95_latency_ms": 18643, "pass_rate": 0.4868, "provider_group": "provider_x", "quality": 0.7476, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.753, "max_latency_ms": 192804, "minimum_quality": 0.8137}, "family": "migration_planning", "id": "synthetic-migration_planning-000328", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.11, "estimated_latency_ms": 174621, "estimated_pass_rate": 0.7599, "estimated_quality": 0.9332, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.753, "max_latency_ms": 192804, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7599, "quality": 0.9332, "quality_target_met": true, "utility": 74.771836}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.753,\"max_latency_ms\":192804,\"minimum_quality\":0.8137},\"task\":{\"ambiguity\":0.5632,\"category\":\"coding\",\"complexity\":0.5567,\"consequence\":0.7737,\"evidence_need\":0.7494,\"family\":\"migration_planning\",\"input_tokens\":4920,\"reversibility\":0.2751,\"risk\":0.5618,\"summary\":\"Design compatibility testing for a deployment rollback schema migration.\"},\"workers\":[{\"credits_per_call\":1.118,\"id\":\"worker_a\",\"p95_latency_ms\":31324,\"pass_rate\":0.5879,\"provider_group\":\"provider_y\",\"quality\":0.7756,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.522,\"id\":\"worker_b\",\"p95_latency_ms\":60504,\"pass_rate\":0.842,\"provider_group\":\"provider_z\",\"quality\":0.8786,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.47,\"id\":\"worker_c\",\"p95_latency_ms\":82793,\"pass_rate\":0.9135,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.11,\"estimated_latency_ms\":174621,\"estimated_pass_rate\":0.7599,\"estimated_quality\":0.9332,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.753,\"max_latency_ms\":192804,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5632, "category": "coding", "complexity": 0.5567, "consequence": 0.7737, "evidence_need": 0.7494, "family": "migration_planning", "input_tokens": 4920, "request_id": "synthetic-migration_planning-000328", "reversibility": 0.2751, "risk": 0.5618, "summary": "Design compatibility testing for a deployment rollback schema migration."}, "workers": [{"credits_per_call": 1.118, "id": "worker_a", "p95_latency_ms": 31324, "pass_rate": 0.5879, "provider_group": "provider_y", "quality": 0.7756, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.522, "id": "worker_b", "p95_latency_ms": 60504, "pass_rate": 0.842, "provider_group": "provider_z", "quality": 0.8786, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.47, "id": "worker_c", "p95_latency_ms": 82793, "pass_rate": 0.9135, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.288, "max_latency_ms": 257193, "minimum_quality": 0.8967}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000330", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.782, "estimated_latency_ms": 185198, "estimated_pass_rate": 0.7721, "estimated_quality": 0.9052, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.288, "max_latency_ms": 257193, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7721, "quality": 0.9052, "quality_target_met": true, "utility": 76.67608}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.288,\"max_latency_ms\":257193,\"minimum_quality\":0.8967},\"task\":{\"ambiguity\":0.4201,\"category\":\"review\",\"complexity\":0.7445,\"consequence\":0.8765,\"evidence_need\":0.6739,\"family\":\"policy_compliance\",\"input_tokens\":240,\"reversibility\":0.6287,\"risk\":0.652,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":1.145,\"id\":\"worker_a\",\"p95_latency_ms\":23147,\"pass_rate\":0.6002,\"provider_group\":\"provider_y\",\"quality\":0.7819,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.575,\"id\":\"worker_b\",\"p95_latency_ms\":88958,\"pass_rate\":0.9344,\"provider_group\":\"provider_z\",\"quality\":0.9475,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.062,\"id\":\"worker_c\",\"p95_latency_ms\":73093,\"pass_rate\":0.8163,\"provider_group\":\"provider_x\",\"quality\":0.8956,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.782,\"estimated_latency_ms\":185198,\"estimated_pass_rate\":0.7721,\"estimated_quality\":0.9052,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.288,\"max_latency_ms\":257193,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4201, "category": "review", "complexity": 0.7445, "consequence": 0.8765, "evidence_need": 0.6739, "family": "policy_compliance", "input_tokens": 240, "request_id": "synthetic-policy_compliance-000330", "reversibility": 0.6287, "risk": 0.652, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 1.145, "id": "worker_a", "p95_latency_ms": 23147, "pass_rate": 0.6002, "provider_group": "provider_y", "quality": 0.7819, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.575, "id": "worker_b", "p95_latency_ms": 88958, "pass_rate": 0.9344, "provider_group": "provider_z", "quality": 0.9475, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.062, "id": "worker_c", "p95_latency_ms": 73093, "pass_rate": 0.8163, "provider_group": "provider_x", "quality": 0.8956, "strengths": ["drafting", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.642, "max_latency_ms": 34413, "minimum_quality": 0.529}, "family": "migration_planning", "id": "synthetic-migration_planning-000577", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.847, "estimated_latency_ms": 25201, "estimated_pass_rate": 0.282, "estimated_quality": 0.593, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.642, "max_latency_ms": 34413, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.282, "quality": 0.593, "quality_target_met": true, "utility": 37.71247}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.642,\"max_latency_ms\":34413,\"minimum_quality\":0.529},\"task\":{\"ambiguity\":0.7212,\"category\":\"coding\",\"complexity\":0.7456,\"consequence\":0.5134,\"evidence_need\":0.4697,\"family\":\"migration_planning\",\"input_tokens\":2520,\"reversibility\":0.4364,\"risk\":0.6078,\"summary\":\"Design compatibility testing for a rate limiting schema migration.\"},\"workers\":[{\"credits_per_call\":11.437,\"id\":\"worker_a\",\"p95_latency_ms\":84727,\"pass_rate\":0.9184,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.375,\"id\":\"worker_b\",\"p95_latency_ms\":59412,\"pass_rate\":0.8576,\"provider_group\":\"provider_y\",\"quality\":0.8712,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.847,\"id\":\"worker_c\",\"p95_latency_ms\":25201,\"pass_rate\":0.4703,\"provider_group\":\"provider_x\",\"quality\":0.74,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.847,\"estimated_latency_ms\":25201,\"estimated_pass_rate\":0.282,\"estimated_quality\":0.593,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.642,\"max_latency_ms\":34413,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7212, "category": "coding", "complexity": 0.7456, "consequence": 0.5134, "evidence_need": 0.4697, "family": "migration_planning", "input_tokens": 2520, "request_id": "synthetic-migration_planning-000577", "reversibility": 0.4364, "risk": 0.6078, "summary": "Design compatibility testing for a rate limiting schema migration."}, "workers": [{"credits_per_call": 11.437, "id": "worker_a", "p95_latency_ms": 84727, "pass_rate": 0.9184, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.375, "id": "worker_b", "p95_latency_ms": 59412, "pass_rate": 0.8576, "provider_group": "provider_y", "quality": 0.8712, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.847, "id": "worker_c", "p95_latency_ms": 25201, "pass_rate": 0.4703, "provider_group": "provider_x", "quality": 0.74, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.224, "max_latency_ms": 174312, "minimum_quality": 0.8168}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000927", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.425, "estimated_latency_ms": 98346, "estimated_pass_rate": 0.5359, "estimated_quality": 0.7392, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.224, "max_latency_ms": 174312, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5359, "quality": 0.7392, "quality_target_met": false, "utility": 51.353979}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.224,\"max_latency_ms\":174312,\"minimum_quality\":0.8168},\"task\":{\"ambiguity\":0.7456,\"category\":\"review\",\"complexity\":0.8435,\"consequence\":0.6052,\"evidence_need\":0.5583,\"family\":\"policy_compliance\",\"input_tokens\":1840,\"reversibility\":0.6792,\"risk\":0.5595,\"summary\":\"Identify evidence required before approving schema evolution for production.\"},\"workers\":[{\"credits_per_call\":3.358,\"id\":\"worker_a\",\"p95_latency_ms\":77729,\"pass_rate\":0.7906,\"provider_group\":\"provider_y\",\"quality\":0.8317,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.067,\"id\":\"worker_b\",\"p95_latency_ms\":20617,\"pass_rate\":0.4712,\"provider_group\":\"provider_z\",\"quality\":0.7178,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.093,\"id\":\"worker_c\",\"p95_latency_ms\":52152,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.425,\"estimated_latency_ms\":98346,\"estimated_pass_rate\":0.5359,\"estimated_quality\":0.7392,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.224,\"max_latency_ms\":174312,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7456, "category": "review", "complexity": 0.8435, "consequence": 0.6052, "evidence_need": 0.5583, "family": "policy_compliance", "input_tokens": 1840, "request_id": "synthetic-policy_compliance-000927", "reversibility": 0.6792, "risk": 0.5595, "summary": "Identify evidence required before approving schema evolution for production."}, "workers": [{"credits_per_call": 3.358, "id": "worker_a", "p95_latency_ms": 77729, "pass_rate": 0.7906, "provider_group": "provider_y", "quality": 0.8317, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.067, "id": "worker_b", "p95_latency_ms": 20617, "pass_rate": 0.4712, "provider_group": "provider_z", "quality": 0.7178, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.093, "id": "worker_c", "p95_latency_ms": 52152, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.903, "max_latency_ms": 183078, "minimum_quality": 0.8413}, "family": "migration_planning", "id": "synthetic-migration_planning-000178", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.165, "estimated_latency_ms": 166234, "estimated_pass_rate": 0.7443, "estimated_quality": 0.9299, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.903, "max_latency_ms": 183078, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7443, "quality": 0.9299, "quality_target_met": true, "utility": 74.469791}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.903,\"max_latency_ms\":183078,\"minimum_quality\":0.8413},\"task\":{\"ambiguity\":0.6229,\"category\":\"coding\",\"complexity\":0.5906,\"consequence\":0.7372,\"evidence_need\":0.6979,\"family\":\"migration_planning\",\"input_tokens\":2120,\"reversibility\":0.3544,\"risk\":0.5863,\"summary\":\"Plan a zero-downtime migration of data retention with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":1.074,\"id\":\"worker_a\",\"p95_latency_ms\":25751,\"pass_rate\":0.4921,\"provider_group\":\"provider_z\",\"quality\":0.779,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.057,\"id\":\"worker_b\",\"p95_latency_ms\":77264,\"pass_rate\":0.8198,\"provider_group\":\"provider_y\",\"quality\":0.849,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.034,\"id\":\"worker_c\",\"p95_latency_ms\":63219,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.165,\"estimated_latency_ms\":166234,\"estimated_pass_rate\":0.7443,\"estimated_quality\":0.9299,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.903,\"max_latency_ms\":183078,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6229, "category": "coding", "complexity": 0.5906, "consequence": 0.7372, "evidence_need": 0.6979, "family": "migration_planning", "input_tokens": 2120, "request_id": "synthetic-migration_planning-000178", "reversibility": 0.3544, "risk": 0.5863, "summary": "Plan a zero-downtime migration of data retention with rollback checkpoints."}, "workers": [{"credits_per_call": 1.074, "id": "worker_a", "p95_latency_ms": 25751, "pass_rate": 0.4921, "provider_group": "provider_z", "quality": 0.779, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.057, "id": "worker_b", "p95_latency_ms": 77264, "pass_rate": 0.8198, "provider_group": "provider_y", "quality": 0.849, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.034, "id": "worker_c", "p95_latency_ms": 63219, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.641, "max_latency_ms": 239240, "minimum_quality": 0.8778}, "family": "migration_planning", "id": "synthetic-migration_planning-000214", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.095, "estimated_latency_ms": 181928, "estimated_pass_rate": 0.7622, "estimated_quality": 0.9125, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 16.641, "max_latency_ms": 239240, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7622, "quality": 0.9125, "quality_target_met": true, "utility": 75.399826}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.641,\"max_latency_ms\":239240,\"minimum_quality\":0.8778},\"task\":{\"ambiguity\":0.4971,\"category\":\"coding\",\"complexity\":0.8673,\"consequence\":0.4831,\"evidence_need\":0.8341,\"family\":\"migration_planning\",\"input_tokens\":7240,\"reversibility\":0.3009,\"risk\":0.5875,\"summary\":\"Design compatibility testing for a streaming inference schema migration.\"},\"workers\":[{\"credits_per_call\":0.952,\"id\":\"worker_a\",\"p95_latency_ms\":19433,\"pass_rate\":0.5955,\"provider_group\":\"provider_z\",\"quality\":0.8036,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.673,\"id\":\"worker_b\",\"p95_latency_ms\":74318,\"pass_rate\":0.7488,\"provider_group\":\"provider_y\",\"quality\":0.8459,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.47,\"id\":\"worker_c\",\"p95_latency_ms\":88177,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9654,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.095,\"estimated_latency_ms\":181928,\"estimated_pass_rate\":0.7622,\"estimated_quality\":0.9125,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.641,\"max_latency_ms\":239240,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4971, "category": "coding", "complexity": 0.8673, "consequence": 0.4831, "evidence_need": 0.8341, "family": "migration_planning", "input_tokens": 7240, "request_id": "synthetic-migration_planning-000214", "reversibility": 0.3009, "risk": 0.5875, "summary": "Design compatibility testing for a streaming inference schema migration."}, "workers": [{"credits_per_call": 0.952, "id": "worker_a", "p95_latency_ms": 19433, "pass_rate": 0.5955, "provider_group": "provider_z", "quality": 0.8036, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.673, "id": "worker_b", "p95_latency_ms": 74318, "pass_rate": 0.7488, "provider_group": "provider_y", "quality": 0.8459, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 10.47, "id": "worker_c", "p95_latency_ms": 88177, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9654, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.444, "max_latency_ms": 201324, "minimum_quality": 0.7448}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000506", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.259, "estimated_latency_ms": 126873, "estimated_pass_rate": 0.8888, "estimated_quality": 0.9191, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.444, "max_latency_ms": 201324, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8888, "quality": 0.9191, "quality_target_met": true, "utility": 68.360602}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.444,\"max_latency_ms\":201324,\"minimum_quality\":0.7448},\"task\":{\"ambiguity\":0.6235,\"category\":\"analysis\",\"complexity\":0.8511,\"consequence\":0.3805,\"evidence_need\":0.5142,\"family\":\"performance_tuning\",\"input_tokens\":3040,\"reversibility\":0.7587,\"risk\":0.3448,\"summary\":\"Review a proposed optimization for data retention for hidden regressions.\"},\"workers\":[{\"credits_per_call\":15.112,\"id\":\"worker_a\",\"p95_latency_ms\":74062,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.956,\"id\":\"worker_b\",\"p95_latency_ms\":32170,\"pass_rate\":0.5414,\"provider_group\":\"provider_z\",\"quality\":0.7162,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.147,\"id\":\"worker_c\",\"p95_latency_ms\":52811,\"pass_rate\":0.8403,\"provider_group\":\"provider_y\",\"quality\":0.931,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.259,\"estimated_latency_ms\":126873,\"estimated_pass_rate\":0.8888,\"estimated_quality\":0.9191,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.444,\"max_latency_ms\":201324,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6235, "category": "analysis", "complexity": 0.8511, "consequence": 0.3805, "evidence_need": 0.5142, "family": "performance_tuning", "input_tokens": 3040, "request_id": "synthetic-performance_tuning-000506", "reversibility": 0.7587, "risk": 0.3448, "summary": "Review a proposed optimization for data retention for hidden regressions."}, "workers": [{"credits_per_call": 15.112, "id": "worker_a", "p95_latency_ms": 74062, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.956, "id": "worker_b", "p95_latency_ms": 32170, "pass_rate": 0.5414, "provider_group": "provider_z", "quality": 0.7162, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.147, "id": "worker_c", "p95_latency_ms": 52811, "pass_rate": 0.8403, "provider_group": "provider_y", "quality": 0.931, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.171, "max_latency_ms": 39659, "minimum_quality": 0.6242}, "family": "migration_planning", "id": "synthetic-migration_planning-000091", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.017, "estimated_latency_ms": 21589, "estimated_pass_rate": 0.4122, "estimated_quality": 0.6227, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.171, "max_latency_ms": 39659, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4122, "quality": 0.6227, "quality_target_met": false, "utility": 43.330321}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.171,\"max_latency_ms\":39659,\"minimum_quality\":0.6242},\"task\":{\"ambiguity\":0.3391,\"category\":\"coding\",\"complexity\":0.7999,\"consequence\":0.549,\"evidence_need\":0.4893,\"family\":\"migration_planning\",\"input_tokens\":6800,\"reversibility\":0.6491,\"risk\":0.4908,\"summary\":\"Review a staged migration for usage metering and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.386,\"id\":\"worker_a\",\"p95_latency_ms\":55775,\"pass_rate\":0.9646,\"provider_group\":\"provider_x\",\"quality\":0.94,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.017,\"id\":\"worker_b\",\"p95_latency_ms\":21589,\"pass_rate\":0.5683,\"provider_group\":\"provider_z\",\"quality\":0.7454,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.987,\"id\":\"worker_c\",\"p95_latency_ms\":71903,\"pass_rate\":0.8784,\"provider_group\":\"provider_y\",\"quality\":0.9008,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.017,\"estimated_latency_ms\":21589,\"estimated_pass_rate\":0.4122,\"estimated_quality\":0.6227,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.171,\"max_latency_ms\":39659,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3391, "category": "coding", "complexity": 0.7999, "consequence": 0.549, "evidence_need": 0.4893, "family": "migration_planning", "input_tokens": 6800, "request_id": "synthetic-migration_planning-000091", "reversibility": 0.6491, "risk": 0.4908, "summary": "Review a staged migration for usage metering and identify irreversible steps."}, "workers": [{"credits_per_call": 13.386, "id": "worker_a", "p95_latency_ms": 55775, "pass_rate": 0.9646, "provider_group": "provider_x", "quality": 0.94, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.017, "id": "worker_b", "p95_latency_ms": 21589, "pass_rate": 0.5683, "provider_group": "provider_z", "quality": 0.7454, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.987, "id": "worker_c", "p95_latency_ms": 71903, "pass_rate": 0.8784, "provider_group": "provider_y", "quality": 0.9008, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 2.649, "max_latency_ms": 52259, "minimum_quality": 0.502}, "family": "migration_planning", "id": "synthetic-migration_planning-000109", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.988, "estimated_latency_ms": 30933, "estimated_pass_rate": 0.4207, "estimated_quality": 0.6152, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.649, "max_latency_ms": 52259, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4207, "quality": 0.6152, "quality_target_met": true, "utility": 42.544076}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":2.649,\"max_latency_ms\":52259,\"minimum_quality\":0.502},\"task\":{\"ambiguity\":0.5171,\"category\":\"coding\",\"complexity\":0.6148,\"consequence\":0.7029,\"evidence_need\":0.6888,\"family\":\"migration_planning\",\"input_tokens\":3280,\"reversibility\":0.3774,\"risk\":0.5051,\"summary\":\"Plan a zero-downtime migration of usage metering with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.988,\"id\":\"worker_a\",\"p95_latency_ms\":30933,\"pass_rate\":0.5704,\"provider_group\":\"provider_z\",\"quality\":0.7175,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.189,\"id\":\"worker_b\",\"p95_latency_ms\":73760,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.984,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.49,\"id\":\"worker_c\",\"p95_latency_ms\":81698,\"pass_rate\":0.864,\"provider_group\":\"provider_x\",\"quality\":0.8185,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.988,\"estimated_latency_ms\":30933,\"estimated_pass_rate\":0.4207,\"estimated_quality\":0.6152,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.649,\"max_latency_ms\":52259,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5171, "category": "coding", "complexity": 0.6148, "consequence": 0.7029, "evidence_need": 0.6888, "family": "migration_planning", "input_tokens": 3280, "request_id": "synthetic-migration_planning-000109", "reversibility": 0.3774, "risk": 0.5051, "summary": "Plan a zero-downtime migration of usage metering with rollback checkpoints."}, "workers": [{"credits_per_call": 0.988, "id": "worker_a", "p95_latency_ms": 30933, "pass_rate": 0.5704, "provider_group": "provider_z", "quality": 0.7175, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.189, "id": "worker_b", "p95_latency_ms": 73760, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.984, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.49, "id": "worker_c", "p95_latency_ms": 81698, "pass_rate": 0.864, "provider_group": "provider_x", "quality": 0.8185, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.602, "max_latency_ms": 38749, "minimum_quality": 0.5154}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000452", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.086, "estimated_latency_ms": 21676, "estimated_pass_rate": 0.3949, "estimated_quality": 0.6666, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.602, "max_latency_ms": 38749, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3949, "quality": 0.6666, "quality_target_met": true, "utility": 44.973023}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.602,\"max_latency_ms\":38749,\"minimum_quality\":0.5154},\"task\":{\"ambiguity\":0.5727,\"category\":\"analysis\",\"complexity\":0.509,\"consequence\":0.4081,\"evidence_need\":0.6379,\"family\":\"performance_tuning\",\"input_tokens\":5680,\"reversibility\":0.8,\"risk\":0.3103,\"summary\":\"Diagnose latency variance in access control and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":10.622,\"id\":\"worker_a\",\"p95_latency_ms\":77480,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.086,\"id\":\"worker_b\",\"p95_latency_ms\":21676,\"pass_rate\":0.5211,\"provider_group\":\"provider_y\",\"quality\":0.7395,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":4.126,\"id\":\"worker_c\",\"p95_latency_ms\":82844,\"pass_rate\":0.9007,\"provider_group\":\"provider_z\",\"quality\":0.9563,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.086,\"estimated_latency_ms\":21676,\"estimated_pass_rate\":0.3949,\"estimated_quality\":0.6666,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.602,\"max_latency_ms\":38749,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5727, "category": "analysis", "complexity": 0.509, "consequence": 0.4081, "evidence_need": 0.6379, "family": "performance_tuning", "input_tokens": 5680, "request_id": "synthetic-performance_tuning-000452", "reversibility": 0.8, "risk": 0.3103, "summary": "Diagnose latency variance in access control and rank the next experiments."}, "workers": [{"credits_per_call": 10.622, "id": "worker_a", "p95_latency_ms": 77480, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.086, "id": "worker_b", "p95_latency_ms": 21676, "pass_rate": 0.5211, "provider_group": "provider_y", "quality": 0.7395, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 4.126, "id": "worker_c", "p95_latency_ms": 82844, "pass_rate": 0.9007, "provider_group": "provider_z", "quality": 0.9563, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.274, "max_latency_ms": 28038, "minimum_quality": 0.7042}, "family": "migration_planning", "id": "synthetic-migration_planning-000544", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.861, "estimated_latency_ms": 20875, "estimated_pass_rate": 0.465, "estimated_quality": 0.6457, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.274, "max_latency_ms": 28038, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.465, "quality": 0.6457, "quality_target_met": false, "utility": 35.580169}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.274,\"max_latency_ms\":28038,\"minimum_quality\":0.7042},\"task\":{\"ambiguity\":0.3567,\"category\":\"coding\",\"complexity\":0.7287,\"consequence\":0.641,\"evidence_need\":0.5215,\"family\":\"migration_planning\",\"input_tokens\":1400,\"reversibility\":0.2615,\"risk\":0.5004,\"summary\":\"Plan a zero-downtime migration of private credentials with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":2.856,\"id\":\"worker_a\",\"p95_latency_ms\":61002,\"pass_rate\":0.8597,\"provider_group\":\"provider_z\",\"quality\":0.8779,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.367,\"id\":\"worker_b\",\"p95_latency_ms\":62037,\"pass_rate\":0.9316,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.861,\"id\":\"worker_c\",\"p95_latency_ms\":20875,\"pass_rate\":0.6151,\"provider_group\":\"provider_x\",\"quality\":0.7574,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.861,\"estimated_latency_ms\":20875,\"estimated_pass_rate\":0.465,\"estimated_quality\":0.6457,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.274,\"max_latency_ms\":28038,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3567, "category": "coding", "complexity": 0.7287, "consequence": 0.641, "evidence_need": 0.5215, "family": "migration_planning", "input_tokens": 1400, "request_id": "synthetic-migration_planning-000544", "reversibility": 0.2615, "risk": 0.5004, "summary": "Plan a zero-downtime migration of private credentials with rollback checkpoints."}, "workers": [{"credits_per_call": 2.856, "id": "worker_a", "p95_latency_ms": 61002, "pass_rate": 0.8597, "provider_group": "provider_z", "quality": 0.8779, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 11.367, "id": "worker_b", "p95_latency_ms": 62037, "pass_rate": 0.9316, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.861, "id": "worker_c", "p95_latency_ms": 20875, "pass_rate": 0.6151, "provider_group": "provider_x", "quality": 0.7574, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.427, "max_latency_ms": 253028, "minimum_quality": 0.8611}, "family": "migration_planning", "id": "synthetic-migration_planning-000949", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.092, "estimated_latency_ms": 101890, "estimated_pass_rate": 0.6792, "estimated_quality": 0.8007, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.427, "max_latency_ms": 253028, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6792, "quality": 0.8007, "quality_target_met": false, "utility": 55.895116}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.427,\"max_latency_ms\":253028,\"minimum_quality\":0.8611},\"task\":{\"ambiguity\":0.3623,\"category\":\"coding\",\"complexity\":0.8973,\"consequence\":0.5132,\"evidence_need\":0.6741,\"family\":\"migration_planning\",\"input_tokens\":5880,\"reversibility\":0.3216,\"risk\":0.5396,\"summary\":\"Design compatibility testing for a tenant-scoped caching schema migration.\"},\"workers\":[{\"credits_per_call\":2.892,\"id\":\"worker_a\",\"p95_latency_ms\":71685,\"pass_rate\":0.7494,\"provider_group\":\"provider_x\",\"quality\":0.8666,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.584,\"id\":\"worker_b\",\"p95_latency_ms\":79308,\"pass_rate\":0.9288,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.2,\"id\":\"worker_c\",\"p95_latency_ms\":30205,\"pass_rate\":0.5985,\"provider_group\":\"provider_y\",\"quality\":0.7667,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.092,\"estimated_latency_ms\":101890,\"estimated_pass_rate\":0.6792,\"estimated_quality\":0.8007,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.427,\"max_latency_ms\":253028,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3623, "category": "coding", "complexity": 0.8973, "consequence": 0.5132, "evidence_need": 0.6741, "family": "migration_planning", "input_tokens": 5880, "request_id": "synthetic-migration_planning-000949", "reversibility": 0.3216, "risk": 0.5396, "summary": "Design compatibility testing for a tenant-scoped caching schema migration."}, "workers": [{"credits_per_call": 2.892, "id": "worker_a", "p95_latency_ms": 71685, "pass_rate": 0.7494, "provider_group": "provider_x", "quality": 0.8666, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.584, "id": "worker_b", "p95_latency_ms": 79308, "pass_rate": 0.9288, "provider_group": "provider_z", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.2, "id": "worker_c", "p95_latency_ms": 30205, "pass_rate": 0.5985, "provider_group": "provider_y", "quality": 0.7667, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.099, "max_latency_ms": 202747, "minimum_quality": 0.944}, "family": "migration_planning", "id": "synthetic-migration_planning-001201", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.407, "estimated_latency_ms": 158172, "estimated_pass_rate": 0.8119, "estimated_quality": 0.9062, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 24.099, "max_latency_ms": 202747, "policy": "strict", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8119, "quality": 0.9062, "quality_target_met": false, "utility": 73.85669}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.099,\"max_latency_ms\":202747,\"minimum_quality\":0.944},\"task\":{\"ambiguity\":0.4481,\"category\":\"coding\",\"complexity\":0.6712,\"consequence\":0.559,\"evidence_need\":0.6087,\"family\":\"migration_planning\",\"input_tokens\":3080,\"reversibility\":0.503,\"risk\":0.7373,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.615,\"id\":\"worker_a\",\"p95_latency_ms\":68407,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9606,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.876,\"id\":\"worker_b\",\"p95_latency_ms\":57479,\"pass_rate\":0.8615,\"provider_group\":\"provider_x\",\"quality\":0.8885,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.916,\"id\":\"worker_c\",\"p95_latency_ms\":32286,\"pass_rate\":0.6064,\"provider_group\":\"provider_z\",\"quality\":0.724,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.407,\"estimated_latency_ms\":158172,\"estimated_pass_rate\":0.8119,\"estimated_quality\":0.9062,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.099,\"max_latency_ms\":202747,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4481, "category": "coding", "complexity": 0.6712, "consequence": 0.559, "evidence_need": 0.6087, "family": "migration_planning", "input_tokens": 3080, "request_id": "synthetic-migration_planning-001201", "reversibility": 0.503, "risk": 0.7373, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 13.615, "id": "worker_a", "p95_latency_ms": 68407, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9606, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.876, "id": "worker_b", "p95_latency_ms": 57479, "pass_rate": 0.8615, "provider_group": "provider_x", "quality": 0.8885, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.916, "id": "worker_c", "p95_latency_ms": 32286, "pass_rate": 0.6064, "provider_group": "provider_z", "quality": 0.724, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 19.147, "max_latency_ms": 227110, "minimum_quality": 0.8164}, "family": "migration_planning", "id": "synthetic-migration_planning-001126", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.568, "estimated_latency_ms": 163753, "estimated_pass_rate": 0.7449, "estimated_quality": 0.9268, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 19.147, "max_latency_ms": 227110, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7449, "quality": 0.9268, "quality_target_met": true, "utility": 71.668351}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":19.147,\"max_latency_ms\":227110,\"minimum_quality\":0.8164},\"task\":{\"ambiguity\":0.4364,\"category\":\"coding\",\"complexity\":0.639,\"consequence\":0.7773,\"evidence_need\":0.7715,\"family\":\"migration_planning\",\"input_tokens\":5760,\"reversibility\":0.56,\"risk\":0.4311,\"summary\":\"Design compatibility testing for a schema evolution schema migration.\"},\"workers\":[{\"credits_per_call\":0.841,\"id\":\"worker_a\",\"p95_latency_ms\":19949,\"pass_rate\":0.5437,\"provider_group\":\"provider_z\",\"quality\":0.7622,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.103,\"id\":\"worker_b\",\"p95_latency_ms\":54876,\"pass_rate\":0.7891,\"provider_group\":\"provider_y\",\"quality\":0.8243,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.624,\"id\":\"worker_c\",\"p95_latency_ms\":88928,\"pass_rate\":0.9828,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.568,\"estimated_latency_ms\":163753,\"estimated_pass_rate\":0.7449,\"estimated_quality\":0.9268,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":19.147,\"max_latency_ms\":227110,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4364, "category": "coding", "complexity": 0.639, "consequence": 0.7773, "evidence_need": 0.7715, "family": "migration_planning", "input_tokens": 5760, "request_id": "synthetic-migration_planning-001126", "reversibility": 0.56, "risk": 0.4311, "summary": "Design compatibility testing for a schema evolution schema migration."}, "workers": [{"credits_per_call": 0.841, "id": "worker_a", "p95_latency_ms": 19949, "pass_rate": 0.5437, "provider_group": "provider_z", "quality": 0.7622, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.103, "id": "worker_b", "p95_latency_ms": 54876, "pass_rate": 0.7891, "provider_group": "provider_y", "quality": 0.8243, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.624, "id": "worker_c", "p95_latency_ms": 88928, "pass_rate": 0.9828, "provider_group": "provider_x", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.617, "max_latency_ms": 240115, "minimum_quality": 0.8332}, "family": "migration_planning", "id": "synthetic-migration_planning-001219", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.394, "estimated_latency_ms": 179053, "estimated_pass_rate": 0.7856, "estimated_quality": 0.9027, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 16.617, "max_latency_ms": 240115, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7856, "quality": 0.9027, "quality_target_met": true, "utility": 71.810402}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.617,\"max_latency_ms\":240115,\"minimum_quality\":0.8332},\"task\":{\"ambiguity\":0.3622,\"category\":\"coding\",\"complexity\":0.8442,\"consequence\":0.6344,\"evidence_need\":0.4156,\"family\":\"migration_planning\",\"input_tokens\":4040,\"reversibility\":0.3937,\"risk\":0.551,\"summary\":\"Review a staged migration for tenant-scoped caching and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.025,\"id\":\"worker_a\",\"p95_latency_ms\":28659,\"pass_rate\":0.5945,\"provider_group\":\"provider_x\",\"quality\":0.684,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.633,\"id\":\"worker_b\",\"p95_latency_ms\":89806,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.736,\"id\":\"worker_c\",\"p95_latency_ms\":60588,\"pass_rate\":0.8553,\"provider_group\":\"provider_z\",\"quality\":0.8605,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.394,\"estimated_latency_ms\":179053,\"estimated_pass_rate\":0.7856,\"estimated_quality\":0.9027,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.617,\"max_latency_ms\":240115,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3622, "category": "coding", "complexity": 0.8442, "consequence": 0.6344, "evidence_need": 0.4156, "family": "migration_planning", "input_tokens": 4040, "request_id": "synthetic-migration_planning-001219", "reversibility": 0.3937, "risk": 0.551, "summary": "Review a staged migration for tenant-scoped caching and identify irreversible steps."}, "workers": [{"credits_per_call": 1.025, "id": "worker_a", "p95_latency_ms": 28659, "pass_rate": 0.5945, "provider_group": "provider_x", "quality": 0.684, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.633, "id": "worker_b", "p95_latency_ms": 89806, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.736, "id": "worker_c", "p95_latency_ms": 60588, "pass_rate": 0.8553, "provider_group": "provider_z", "quality": 0.8605, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 6.907, "max_latency_ms": 224038, "minimum_quality": 0.8875}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000582", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.784, "estimated_latency_ms": 101397, "estimated_pass_rate": 0.6173, "estimated_quality": 0.8279, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 6.907, "max_latency_ms": 224038, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6173, "quality": 0.8279, "quality_target_met": false, "utility": 58.961255}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":6.907,\"max_latency_ms\":224038,\"minimum_quality\":0.8875},\"task\":{\"ambiguity\":0.4915,\"category\":\"review\",\"complexity\":0.6212,\"consequence\":0.6376,\"evidence_need\":0.5623,\"family\":\"policy_compliance\",\"input_tokens\":4560,\"reversibility\":0.5889,\"risk\":0.7034,\"summary\":\"Audit a workflow for rate limiting against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":13.413,\"id\":\"worker_a\",\"p95_latency_ms\":75851,\"pass_rate\":0.9565,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.689,\"id\":\"worker_b\",\"p95_latency_ms\":76513,\"pass_rate\":0.7936,\"provider_group\":\"provider_y\",\"quality\":0.9215,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.095,\"id\":\"worker_c\",\"p95_latency_ms\":24884,\"pass_rate\":0.484,\"provider_group\":\"provider_x\",\"quality\":0.7813,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.784,\"estimated_latency_ms\":101397,\"estimated_pass_rate\":0.6173,\"estimated_quality\":0.8279,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":6.907,\"max_latency_ms\":224038,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4915, "category": "review", "complexity": 0.6212, "consequence": 0.6376, "evidence_need": 0.5623, "family": "policy_compliance", "input_tokens": 4560, "request_id": "synthetic-policy_compliance-000582", "reversibility": 0.5889, "risk": 0.7034, "summary": "Audit a workflow for rate limiting against explicit access requirements."}, "workers": [{"credits_per_call": 13.413, "id": "worker_a", "p95_latency_ms": 75851, "pass_rate": 0.9565, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.689, "id": "worker_b", "p95_latency_ms": 76513, "pass_rate": 0.7936, "provider_group": "provider_y", "quality": 0.9215, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.095, "id": "worker_c", "p95_latency_ms": 24884, "pass_rate": 0.484, "provider_group": "provider_x", "quality": 0.7813, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.627, "max_latency_ms": 21900, "minimum_quality": 0.6166}, "family": "migration_planning", "id": "synthetic-migration_planning-000604", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.084, "estimated_latency_ms": 20407, "estimated_pass_rate": 0.4357, "estimated_quality": 0.6371, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.627, "max_latency_ms": 21900, "policy": "lean", "reason_codes": ["complex_task", "latency_constrained"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4357, "quality": 0.6371, "quality_target_met": true, "utility": 42.439615}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.627,\"max_latency_ms\":21900,\"minimum_quality\":0.6166},\"task\":{\"ambiguity\":0.677,\"category\":\"coding\",\"complexity\":0.7289,\"consequence\":0.6069,\"evidence_need\":0.643,\"family\":\"migration_planning\",\"input_tokens\":5520,\"reversibility\":0.4139,\"risk\":0.5938,\"summary\":\"Plan a zero-downtime migration of observability with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":4.039,\"id\":\"worker_a\",\"p95_latency_ms\":50698,\"pass_rate\":0.8307,\"provider_group\":\"provider_z\",\"quality\":0.8211,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.932,\"id\":\"worker_b\",\"p95_latency_ms\":86098,\"pass_rate\":0.9265,\"provider_group\":\"provider_y\",\"quality\":0.9739,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.084,\"id\":\"worker_c\",\"p95_latency_ms\":20407,\"pass_rate\":0.6176,\"provider_group\":\"provider_x\",\"quality\":0.7771,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.084,\"estimated_latency_ms\":20407,\"estimated_pass_rate\":0.4357,\"estimated_quality\":0.6371,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.627,\"max_latency_ms\":21900,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"latency_constrained\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.677, "category": "coding", "complexity": 0.7289, "consequence": 0.6069, "evidence_need": 0.643, "family": "migration_planning", "input_tokens": 5520, "request_id": "synthetic-migration_planning-000604", "reversibility": 0.4139, "risk": 0.5938, "summary": "Plan a zero-downtime migration of observability with rollback checkpoints."}, "workers": [{"credits_per_call": 4.039, "id": "worker_a", "p95_latency_ms": 50698, "pass_rate": 0.8307, "provider_group": "provider_z", "quality": 0.8211, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.932, "id": "worker_b", "p95_latency_ms": 86098, "pass_rate": 0.9265, "provider_group": "provider_y", "quality": 0.9739, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.084, "id": "worker_c", "p95_latency_ms": 20407, "pass_rate": 0.6176, "provider_group": "provider_x", "quality": 0.7771, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.253, "max_latency_ms": 206182, "minimum_quality": 0.8321}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001068", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.235, "estimated_latency_ms": 167551, "estimated_pass_rate": 0.7634, "estimated_quality": 0.9173, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.253, "max_latency_ms": 206182, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7634, "quality": 0.9173, "quality_target_met": true, "utility": 78.02292}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.253,\"max_latency_ms\":206182,\"minimum_quality\":0.8321},\"task\":{\"ambiguity\":0.3632,\"category\":\"review\",\"complexity\":0.8138,\"consequence\":0.761,\"evidence_need\":0.7991,\"family\":\"policy_compliance\",\"input_tokens\":7240,\"reversibility\":0.559,\"risk\":0.7489,\"summary\":\"Review whether database failover satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.146,\"id\":\"worker_a\",\"p95_latency_ms\":58716,\"pass_rate\":0.748,\"provider_group\":\"provider_z\",\"quality\":0.9187,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.004,\"id\":\"worker_b\",\"p95_latency_ms\":24009,\"pass_rate\":0.536,\"provider_group\":\"provider_y\",\"quality\":0.7675,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":15.085,\"id\":\"worker_c\",\"p95_latency_ms\":84826,\"pass_rate\":0.9853,\"provider_group\":\"provider_x\",\"quality\":0.9781,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.235,\"estimated_latency_ms\":167551,\"estimated_pass_rate\":0.7634,\"estimated_quality\":0.9173,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.253,\"max_latency_ms\":206182,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3632, "category": "review", "complexity": 0.8138, "consequence": 0.761, "evidence_need": 0.7991, "family": "policy_compliance", "input_tokens": 7240, "request_id": "synthetic-policy_compliance-001068", "reversibility": 0.559, "risk": 0.7489, "summary": "Review whether database failover satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.146, "id": "worker_a", "p95_latency_ms": 58716, "pass_rate": 0.748, "provider_group": "provider_z", "quality": 0.9187, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.004, "id": "worker_b", "p95_latency_ms": 24009, "pass_rate": 0.536, "provider_group": "provider_y", "quality": 0.7675, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 15.085, "id": "worker_c", "p95_latency_ms": 84826, "pass_rate": 0.9853, "provider_group": "provider_x", "quality": 0.9781, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.389, "max_latency_ms": 57419, "minimum_quality": 0.4917}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000320", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.876, "estimated_latency_ms": 30354, "estimated_pass_rate": 0.3987, "estimated_quality": 0.5389, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.389, "max_latency_ms": 57419, "policy": "lean", "reason_codes": ["single_call_sufficient"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3987, "quality": 0.5389, "quality_target_met": true, "utility": 35.788332}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.389,\"max_latency_ms\":57419,\"minimum_quality\":0.4917},\"task\":{\"ambiguity\":0.3075,\"category\":\"analysis\",\"complexity\":0.6569,\"consequence\":0.4905,\"evidence_need\":0.3547,\"family\":\"performance_tuning\",\"input_tokens\":5320,\"reversibility\":0.6557,\"risk\":0.3289,\"summary\":\"Review a proposed optimization for webhook delivery for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.773,\"id\":\"worker_a\",\"p95_latency_ms\":78187,\"pass_rate\":0.8554,\"provider_group\":\"provider_y\",\"quality\":0.9873,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.888,\"id\":\"worker_b\",\"p95_latency_ms\":51069,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.876,\"id\":\"worker_c\",\"p95_latency_ms\":30354,\"pass_rate\":0.5231,\"provider_group\":\"provider_z\",\"quality\":0.622,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.876,\"estimated_latency_ms\":30354,\"estimated_pass_rate\":0.3987,\"estimated_quality\":0.5389,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.389,\"max_latency_ms\":57419,\"policy\":\"lean\",\"reason_codes\":[\"single_call_sufficient\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3075, "category": "analysis", "complexity": 0.6569, "consequence": 0.4905, "evidence_need": 0.3547, "family": "performance_tuning", "input_tokens": 5320, "request_id": "synthetic-performance_tuning-000320", "reversibility": 0.6557, "risk": 0.3289, "summary": "Review a proposed optimization for webhook delivery for hidden regressions."}, "workers": [{"credits_per_call": 3.773, "id": "worker_a", "p95_latency_ms": 78187, "pass_rate": 0.8554, "provider_group": "provider_y", "quality": 0.9873, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.888, "id": "worker_b", "p95_latency_ms": 51069, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.876, "id": "worker_c", "p95_latency_ms": 30354, "pass_rate": 0.5231, "provider_group": "provider_z", "quality": 0.622, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.342, "max_latency_ms": 203559, "minimum_quality": 0.8356}, "family": "migration_planning", "id": "synthetic-migration_planning-001096", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.777, "estimated_latency_ms": 109146, "estimated_pass_rate": 0.686, "estimated_quality": 0.7923, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.342, "max_latency_ms": 203559, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.686, "quality": 0.7923, "quality_target_met": false, "utility": 55.363258}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.342,\"max_latency_ms\":203559,\"minimum_quality\":0.8356},\"task\":{\"ambiguity\":0.4649,\"category\":\"coding\",\"complexity\":0.8628,\"consequence\":0.6517,\"evidence_need\":0.6622,\"family\":\"migration_planning\",\"input_tokens\":960,\"reversibility\":0.6492,\"risk\":0.3874,\"summary\":\"Review a staged migration for deployment rollback and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.817,\"id\":\"worker_a\",\"p95_latency_ms\":31937,\"pass_rate\":0.5212,\"provider_group\":\"provider_y\",\"quality\":0.7404,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.96,\"id\":\"worker_b\",\"p95_latency_ms\":77209,\"pass_rate\":0.7824,\"provider_group\":\"provider_z\",\"quality\":0.8731,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.539,\"id\":\"worker_c\",\"p95_latency_ms\":91391,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9845,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.777,\"estimated_latency_ms\":109146,\"estimated_pass_rate\":0.686,\"estimated_quality\":0.7923,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.342,\"max_latency_ms\":203559,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4649, "category": "coding", "complexity": 0.8628, "consequence": 0.6517, "evidence_need": 0.6622, "family": "migration_planning", "input_tokens": 960, "request_id": "synthetic-migration_planning-001096", "reversibility": 0.6492, "risk": 0.3874, "summary": "Review a staged migration for deployment rollback and identify irreversible steps."}, "workers": [{"credits_per_call": 0.817, "id": "worker_a", "p95_latency_ms": 31937, "pass_rate": 0.5212, "provider_group": "provider_y", "quality": 0.7404, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.96, "id": "worker_b", "p95_latency_ms": 77209, "pass_rate": 0.7824, "provider_group": "provider_z", "quality": 0.8731, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.539, "id": "worker_c", "p95_latency_ms": 91391, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9845, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.258, "max_latency_ms": 264283, "minimum_quality": 0.7986}, "family": "migration_planning", "id": "synthetic-migration_planning-001174", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.628, "estimated_latency_ms": 109685, "estimated_pass_rate": 0.7418, "estimated_quality": 0.7874, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.258, "max_latency_ms": 264283, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7418, "quality": 0.7874, "quality_target_met": false, "utility": 56.220232}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.258,\"max_latency_ms\":264283,\"minimum_quality\":0.7986},\"task\":{\"ambiguity\":0.448,\"category\":\"coding\",\"complexity\":0.9149,\"consequence\":0.556,\"evidence_need\":0.4091,\"family\":\"migration_planning\",\"input_tokens\":4360,\"reversibility\":0.6101,\"risk\":0.5479,\"summary\":\"Plan a zero-downtime migration of prompt caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.89,\"id\":\"worker_a\",\"p95_latency_ms\":30115,\"pass_rate\":0.4515,\"provider_group\":\"provider_z\",\"quality\":0.7501,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.33,\"id\":\"worker_b\",\"p95_latency_ms\":81306,\"pass_rate\":0.9927,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.738,\"id\":\"worker_c\",\"p95_latency_ms\":79570,\"pass_rate\":0.8719,\"provider_group\":\"provider_x\",\"quality\":0.9053,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.628,\"estimated_latency_ms\":109685,\"estimated_pass_rate\":0.7418,\"estimated_quality\":0.7874,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.258,\"max_latency_ms\":264283,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.448, "category": "coding", "complexity": 0.9149, "consequence": 0.556, "evidence_need": 0.4091, "family": "migration_planning", "input_tokens": 4360, "request_id": "synthetic-migration_planning-001174", "reversibility": 0.6101, "risk": 0.5479, "summary": "Plan a zero-downtime migration of prompt caching with rollback checkpoints."}, "workers": [{"credits_per_call": 0.89, "id": "worker_a", "p95_latency_ms": 30115, "pass_rate": 0.4515, "provider_group": "provider_z", "quality": 0.7501, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.33, "id": "worker_b", "p95_latency_ms": 81306, "pass_rate": 0.9927, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.738, "id": "worker_c", "p95_latency_ms": 79570, "pass_rate": 0.8719, "provider_group": "provider_x", "quality": 0.9053, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.05, "max_latency_ms": 132600, "minimum_quality": 0.685}, "family": "migration_planning", "id": "synthetic-migration_planning-000370", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 13.221, "estimated_latency_ms": 102036, "estimated_pass_rate": 0.837, "estimated_quality": 0.8955, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.05, "max_latency_ms": 132600, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 7, "pass_rate": 0.837, "quality": 0.8955, "quality_target_met": true, "utility": 63.235385}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.05,\"max_latency_ms\":132600,\"minimum_quality\":0.685},\"task\":{\"ambiguity\":0.7013,\"category\":\"coding\",\"complexity\":0.592,\"consequence\":0.708,\"evidence_need\":0.561,\"family\":\"migration_planning\",\"input_tokens\":1040,\"reversibility\":0.2574,\"risk\":0.4441,\"summary\":\"Design compatibility testing for a access control schema migration.\"},\"workers\":[{\"credits_per_call\":12.284,\"id\":\"worker_a\",\"p95_latency_ms\":75392,\"pass_rate\":0.9672,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.937,\"id\":\"worker_b\",\"p95_latency_ms\":26644,\"pass_rate\":0.4874,\"provider_group\":\"provider_y\",\"quality\":0.7191,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.98,\"id\":\"worker_c\",\"p95_latency_ms\":68222,\"pass_rate\":0.7888,\"provider_group\":\"provider_z\",\"quality\":0.8828,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":13.221,\"estimated_latency_ms\":102036,\"estimated_pass_rate\":0.837,\"estimated_quality\":0.8955,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.05,\"max_latency_ms\":132600,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7013, "category": "coding", "complexity": 0.592, "consequence": 0.708, "evidence_need": 0.561, "family": "migration_planning", "input_tokens": 1040, "request_id": "synthetic-migration_planning-000370", "reversibility": 0.2574, "risk": 0.4441, "summary": "Design compatibility testing for a access control schema migration."}, "workers": [{"credits_per_call": 12.284, "id": "worker_a", "p95_latency_ms": 75392, "pass_rate": 0.9672, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.937, "id": "worker_b", "p95_latency_ms": 26644, "pass_rate": 0.4874, "provider_group": "provider_y", "quality": 0.7191, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.98, "id": "worker_c", "p95_latency_ms": 68222, "pass_rate": 0.7888, "provider_group": "provider_z", "quality": 0.8828, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 15.293, "max_latency_ms": 191312, "minimum_quality": 0.8838}, "family": "migration_planning", "id": "synthetic-migration_planning-000199", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.741, "estimated_latency_ms": 106866, "estimated_pass_rate": 0.8594, "estimated_quality": 0.9137, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 15.293, "max_latency_ms": 191312, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8594, "quality": 0.9137, "quality_target_met": true, "utility": 69.198887}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":15.293,\"max_latency_ms\":191312,\"minimum_quality\":0.8838},\"task\":{\"ambiguity\":0.7054,\"category\":\"coding\",\"complexity\":0.8065,\"consequence\":0.7783,\"evidence_need\":0.5543,\"family\":\"migration_planning\",\"input_tokens\":2800,\"reversibility\":0.6203,\"risk\":0.4747,\"summary\":\"Design compatibility testing for a background jobs schema migration.\"},\"workers\":[{\"credits_per_call\":11.75,\"id\":\"worker_a\",\"p95_latency_ms\":57982,\"pass_rate\":0.9593,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.06,\"id\":\"worker_b\",\"p95_latency_ms\":18115,\"pass_rate\":0.4947,\"provider_group\":\"provider_z\",\"quality\":0.7218,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.991,\"id\":\"worker_c\",\"p95_latency_ms\":48884,\"pass_rate\":0.7925,\"provider_group\":\"provider_y\",\"quality\":0.9112,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.741,\"estimated_latency_ms\":106866,\"estimated_pass_rate\":0.8594,\"estimated_quality\":0.9137,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":15.293,\"max_latency_ms\":191312,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7054, "category": "coding", "complexity": 0.8065, "consequence": 0.7783, "evidence_need": 0.5543, "family": "migration_planning", "input_tokens": 2800, "request_id": "synthetic-migration_planning-000199", "reversibility": 0.6203, "risk": 0.4747, "summary": "Design compatibility testing for a background jobs schema migration."}, "workers": [{"credits_per_call": 11.75, "id": "worker_a", "p95_latency_ms": 57982, "pass_rate": 0.9593, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.06, "id": "worker_b", "p95_latency_ms": 18115, "pass_rate": 0.4947, "provider_group": "provider_z", "quality": 0.7218, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.991, "id": "worker_c", "p95_latency_ms": 48884, "pass_rate": 0.7925, "provider_group": "provider_y", "quality": 0.9112, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.09, "max_latency_ms": 221273, "minimum_quality": 0.9429}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001095", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.789, "estimated_latency_ms": 199443, "estimated_pass_rate": 0.7094, "estimated_quality": 0.8765, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 24.09, "max_latency_ms": 221273, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7094, "quality": 0.8765, "quality_target_met": false, "utility": 69.896961}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.09,\"max_latency_ms\":221273,\"minimum_quality\":0.9429},\"task\":{\"ambiguity\":0.7689,\"category\":\"review\",\"complexity\":0.4259,\"consequence\":0.8806,\"evidence_need\":0.6571,\"family\":\"policy_compliance\",\"input_tokens\":6400,\"reversibility\":0.4008,\"risk\":0.7309,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":13.368,\"id\":\"worker_a\",\"p95_latency_ms\":92450,\"pass_rate\":0.9209,\"provider_group\":\"provider_y\",\"quality\":0.9208,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.147,\"id\":\"worker_b\",\"p95_latency_ms\":25767,\"pass_rate\":0.4867,\"provider_group\":\"provider_z\",\"quality\":0.7279,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.274,\"id\":\"worker_c\",\"p95_latency_ms\":81226,\"pass_rate\":0.755,\"provider_group\":\"provider_x\",\"quality\":0.8326,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.789,\"estimated_latency_ms\":199443,\"estimated_pass_rate\":0.7094,\"estimated_quality\":0.8765,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.09,\"max_latency_ms\":221273,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7689, "category": "review", "complexity": 0.4259, "consequence": 0.8806, "evidence_need": 0.6571, "family": "policy_compliance", "input_tokens": 6400, "request_id": "synthetic-policy_compliance-001095", "reversibility": 0.4008, "risk": 0.7309, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 13.368, "id": "worker_a", "p95_latency_ms": 92450, "pass_rate": 0.9209, "provider_group": "provider_y", "quality": 0.9208, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.147, "id": "worker_b", "p95_latency_ms": 25767, "pass_rate": 0.4867, "provider_group": "provider_z", "quality": 0.7279, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.274, "id": "worker_c", "p95_latency_ms": 81226, "pass_rate": 0.755, "provider_group": "provider_x", "quality": 0.8326, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.023, "max_latency_ms": 49020, "minimum_quality": 0.607}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000827", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.844, "estimated_latency_ms": 31491, "estimated_pass_rate": 0.2789, "estimated_quality": 0.513, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.023, "max_latency_ms": 49020, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2789, "quality": 0.513, "quality_target_met": false, "utility": 28.461434}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.023,\"max_latency_ms\":49020,\"minimum_quality\":0.607},\"task\":{\"ambiguity\":0.7356,\"category\":\"analysis\",\"complexity\":0.5814,\"consequence\":0.6409,\"evidence_need\":0.3547,\"family\":\"performance_tuning\",\"input_tokens\":800,\"reversibility\":0.8726,\"risk\":0.4674,\"summary\":\"Review a proposed optimization for prompt caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":3.898,\"id\":\"worker_a\",\"p95_latency_ms\":63826,\"pass_rate\":0.7896,\"provider_group\":\"provider_y\",\"quality\":0.8968,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.789,\"id\":\"worker_b\",\"p95_latency_ms\":67943,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.844,\"id\":\"worker_c\",\"p95_latency_ms\":31491,\"pass_rate\":0.4383,\"provider_group\":\"provider_x\",\"quality\":0.6212,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.844,\"estimated_latency_ms\":31491,\"estimated_pass_rate\":0.2789,\"estimated_quality\":0.513,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.023,\"max_latency_ms\":49020,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7356, "category": "analysis", "complexity": 0.5814, "consequence": 0.6409, "evidence_need": 0.3547, "family": "performance_tuning", "input_tokens": 800, "request_id": "synthetic-performance_tuning-000827", "reversibility": 0.8726, "risk": 0.4674, "summary": "Review a proposed optimization for prompt caching for hidden regressions."}, "workers": [{"credits_per_call": 3.898, "id": "worker_a", "p95_latency_ms": 63826, "pass_rate": 0.7896, "provider_group": "provider_y", "quality": 0.8968, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.789, "id": "worker_b", "p95_latency_ms": 67943, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.844, "id": "worker_c", "p95_latency_ms": 31491, "pass_rate": 0.4383, "provider_group": "provider_x", "quality": 0.6212, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.443, "max_latency_ms": 166247, "minimum_quality": 0.8706}, "family": "migration_planning", "id": "synthetic-migration_planning-000916", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.015, "estimated_latency_ms": 156490, "estimated_pass_rate": 0.757, "estimated_quality": 0.8898, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.443, "max_latency_ms": 166247, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.757, "quality": 0.8898, "quality_target_met": true, "utility": 69.752011}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.443,\"max_latency_ms\":166247,\"minimum_quality\":0.8706},\"task\":{\"ambiguity\":0.7256,\"category\":\"coding\",\"complexity\":0.6712,\"consequence\":0.5385,\"evidence_need\":0.4381,\"family\":\"migration_planning\",\"input_tokens\":3200,\"reversibility\":0.5071,\"risk\":0.5049,\"summary\":\"Plan a zero-downtime migration of access control with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":10.814,\"id\":\"worker_a\",\"p95_latency_ms\":75234,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9285,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.13,\"id\":\"worker_b\",\"p95_latency_ms\":51923,\"pass_rate\":0.8634,\"provider_group\":\"provider_y\",\"quality\":0.8308,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.071,\"id\":\"worker_c\",\"p95_latency_ms\":29333,\"pass_rate\":0.5442,\"provider_group\":\"provider_z\",\"quality\":0.7777,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.015,\"estimated_latency_ms\":156490,\"estimated_pass_rate\":0.757,\"estimated_quality\":0.8898,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.443,\"max_latency_ms\":166247,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7256, "category": "coding", "complexity": 0.6712, "consequence": 0.5385, "evidence_need": 0.4381, "family": "migration_planning", "input_tokens": 3200, "request_id": "synthetic-migration_planning-000916", "reversibility": 0.5071, "risk": 0.5049, "summary": "Plan a zero-downtime migration of access control with rollback checkpoints."}, "workers": [{"credits_per_call": 10.814, "id": "worker_a", "p95_latency_ms": 75234, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9285, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.13, "id": "worker_b", "p95_latency_ms": 51923, "pass_rate": 0.8634, "provider_group": "provider_y", "quality": 0.8308, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.071, "id": "worker_c", "p95_latency_ms": 29333, "pass_rate": 0.5442, "provider_group": "provider_z", "quality": 0.7777, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.115, "max_latency_ms": 184369, "minimum_quality": 0.7368}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000047", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.836, "estimated_latency_ms": 102580, "estimated_pass_rate": 0.7243, "estimated_quality": 0.7887, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.115, "max_latency_ms": 184369, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7243, "quality": 0.7887, "quality_target_met": true, "utility": 60.24885}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.115,\"max_latency_ms\":184369,\"minimum_quality\":0.7368},\"task\":{\"ambiguity\":0.5636,\"category\":\"analysis\",\"complexity\":0.8525,\"consequence\":0.6859,\"evidence_need\":0.5894,\"family\":\"performance_tuning\",\"input_tokens\":7880,\"reversibility\":0.8804,\"risk\":0.546,\"summary\":\"Choose between throughput and tail-latency changes for observability.\"},\"workers\":[{\"credits_per_call\":14.028,\"id\":\"worker_a\",\"p95_latency_ms\":72107,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.934,\"id\":\"worker_b\",\"p95_latency_ms\":73410,\"pass_rate\":0.8525,\"provider_group\":\"provider_y\",\"quality\":0.8987,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.902,\"id\":\"worker_c\",\"p95_latency_ms\":29170,\"pass_rate\":0.5047,\"provider_group\":\"provider_x\",\"quality\":0.6796,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.836,\"estimated_latency_ms\":102580,\"estimated_pass_rate\":0.7243,\"estimated_quality\":0.7887,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.115,\"max_latency_ms\":184369,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5636, "category": "analysis", "complexity": 0.8525, "consequence": 0.6859, "evidence_need": 0.5894, "family": "performance_tuning", "input_tokens": 7880, "request_id": "synthetic-performance_tuning-000047", "reversibility": 0.8804, "risk": 0.546, "summary": "Choose between throughput and tail-latency changes for observability."}, "workers": [{"credits_per_call": 14.028, "id": "worker_a", "p95_latency_ms": 72107, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.934, "id": "worker_b", "p95_latency_ms": 73410, "pass_rate": 0.8525, "provider_group": "provider_y", "quality": 0.8987, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.902, "id": "worker_c", "p95_latency_ms": 29170, "pass_rate": 0.5047, "provider_group": "provider_x", "quality": 0.6796, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 22.51, "max_latency_ms": 229571, "minimum_quality": 0.786}, "family": "migration_planning", "id": "synthetic-migration_planning-001105", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.278, "estimated_latency_ms": 180655, "estimated_pass_rate": 0.7367, "estimated_quality": 0.9046, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 22.51, "max_latency_ms": 229571, "policy": "strict", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7367, "quality": 0.9046, "quality_target_met": true, "utility": 70.964026}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":22.51,\"max_latency_ms\":229571,\"minimum_quality\":0.786},\"task\":{\"ambiguity\":0.5615,\"category\":\"coding\",\"complexity\":0.5607,\"consequence\":0.6663,\"evidence_need\":0.55,\"family\":\"migration_planning\",\"input_tokens\":2120,\"reversibility\":0.5687,\"risk\":0.5854,\"summary\":\"Plan a zero-downtime migration of streaming inference with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.806,\"id\":\"worker_a\",\"p95_latency_ms\":63336,\"pass_rate\":0.8311,\"provider_group\":\"provider_x\",\"quality\":0.8109,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.184,\"id\":\"worker_b\",\"p95_latency_ms\":27101,\"pass_rate\":0.5328,\"provider_group\":\"provider_y\",\"quality\":0.7719,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.288,\"id\":\"worker_c\",\"p95_latency_ms\":90218,\"pass_rate\":0.9395,\"provider_group\":\"provider_z\",\"quality\":0.9588,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.278,\"estimated_latency_ms\":180655,\"estimated_pass_rate\":0.7367,\"estimated_quality\":0.9046,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":22.51,\"max_latency_ms\":229571,\"policy\":\"strict\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.5615, "category": "coding", "complexity": 0.5607, "consequence": 0.6663, "evidence_need": 0.55, "family": "migration_planning", "input_tokens": 2120, "request_id": "synthetic-migration_planning-001105", "reversibility": 0.5687, "risk": 0.5854, "summary": "Plan a zero-downtime migration of streaming inference with rollback checkpoints."}, "workers": [{"credits_per_call": 3.806, "id": "worker_a", "p95_latency_ms": 63336, "pass_rate": 0.8311, "provider_group": "provider_x", "quality": 0.8109, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.184, "id": "worker_b", "p95_latency_ms": 27101, "pass_rate": 0.5328, "provider_group": "provider_y", "quality": 0.7719, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.288, "id": "worker_c", "p95_latency_ms": 90218, "pass_rate": 0.9395, "provider_group": "provider_z", "quality": 0.9588, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.566, "max_latency_ms": 39269, "minimum_quality": 0.6618}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000855", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.943, "estimated_latency_ms": 25339, "estimated_pass_rate": 0.435, "estimated_quality": 0.6666, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.566, "max_latency_ms": 39269, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.435, "quality": 0.6666, "quality_target_met": true, "utility": 45.858011}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.566,\"max_latency_ms\":39269,\"minimum_quality\":0.6618},\"task\":{\"ambiguity\":0.5531,\"category\":\"review\",\"complexity\":0.5002,\"consequence\":0.6397,\"evidence_need\":0.7828,\"family\":\"policy_compliance\",\"input_tokens\":400,\"reversibility\":0.4763,\"risk\":0.6556,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":0.943,\"id\":\"worker_a\",\"p95_latency_ms\":25339,\"pass_rate\":0.5866,\"provider_group\":\"provider_z\",\"quality\":0.7614,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.827,\"id\":\"worker_b\",\"p95_latency_ms\":88379,\"pass_rate\":0.9333,\"provider_group\":\"provider_y\",\"quality\":0.9857,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.509,\"id\":\"worker_c\",\"p95_latency_ms\":78194,\"pass_rate\":0.7261,\"provider_group\":\"provider_x\",\"quality\":0.8842,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.943,\"estimated_latency_ms\":25339,\"estimated_pass_rate\":0.435,\"estimated_quality\":0.6666,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.566,\"max_latency_ms\":39269,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5531, "category": "review", "complexity": 0.5002, "consequence": 0.6397, "evidence_need": 0.7828, "family": "policy_compliance", "input_tokens": 400, "request_id": "synthetic-policy_compliance-000855", "reversibility": 0.4763, "risk": 0.6556, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 0.943, "id": "worker_a", "p95_latency_ms": 25339, "pass_rate": 0.5866, "provider_group": "provider_z", "quality": 0.7614, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.827, "id": "worker_b", "p95_latency_ms": 88379, "pass_rate": 0.9333, "provider_group": "provider_y", "quality": 0.9857, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.509, "id": "worker_c", "p95_latency_ms": 78194, "pass_rate": 0.7261, "provider_group": "provider_x", "quality": 0.8842, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.07, "max_latency_ms": 235502, "minimum_quality": 0.9005}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001053", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.558, "estimated_latency_ms": 168012, "estimated_pass_rate": 0.7971, "estimated_quality": 0.9139, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.07, "max_latency_ms": 235502, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7971, "quality": 0.9139, "quality_target_met": true, "utility": 79.155535}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.07,\"max_latency_ms\":235502,\"minimum_quality\":0.9005},\"task\":{\"ambiguity\":0.5001,\"category\":\"review\",\"complexity\":0.7345,\"consequence\":0.7673,\"evidence_need\":0.5587,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.5137,\"risk\":0.8167,\"summary\":\"Review whether artifact storage satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":3.522,\"id\":\"worker_a\",\"p95_latency_ms\":68501,\"pass_rate\":0.8769,\"provider_group\":\"provider_x\",\"quality\":0.7992,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.966,\"id\":\"worker_b\",\"p95_latency_ms\":79485,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.07,\"id\":\"worker_c\",\"p95_latency_ms\":20026,\"pass_rate\":0.6022,\"provider_group\":\"provider_y\",\"quality\":0.7364,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.558,\"estimated_latency_ms\":168012,\"estimated_pass_rate\":0.7971,\"estimated_quality\":0.9139,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.07,\"max_latency_ms\":235502,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5001, "category": "review", "complexity": 0.7345, "consequence": 0.7673, "evidence_need": 0.5587, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-001053", "reversibility": 0.5137, "risk": 0.8167, "summary": "Review whether artifact storage satisfies a documented retention policy."}, "workers": [{"credits_per_call": 3.522, "id": "worker_a", "p95_latency_ms": 68501, "pass_rate": 0.8769, "provider_group": "provider_x", "quality": 0.7992, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.966, "id": "worker_b", "p95_latency_ms": 79485, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.07, "id": "worker_c", "p95_latency_ms": 20026, "pass_rate": 0.6022, "provider_group": "provider_y", "quality": 0.7364, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 18.726, "max_latency_ms": 222136, "minimum_quality": 0.8815}, "family": "migration_planning", "id": "synthetic-migration_planning-000208", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.265, "estimated_latency_ms": 132640, "estimated_pass_rate": 0.8359, "estimated_quality": 0.8675, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 18.726, "max_latency_ms": 222136, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "credit_constrained", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8359, "quality": 0.8675, "quality_target_met": false, "utility": 65.803782}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":18.726,\"max_latency_ms\":222136,\"minimum_quality\":0.8815},\"task\":{\"ambiguity\":0.4541,\"category\":\"coding\",\"complexity\":0.8506,\"consequence\":0.5773,\"evidence_need\":0.7883,\"family\":\"migration_planning\",\"input_tokens\":3240,\"reversibility\":0.3327,\"risk\":0.6417,\"summary\":\"Design compatibility testing for a webhook delivery schema migration.\"},\"workers\":[{\"credits_per_call\":2.846,\"id\":\"worker_a\",\"p95_latency_ms\":68436,\"pass_rate\":0.7495,\"provider_group\":\"provider_y\",\"quality\":0.8215,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.953,\"id\":\"worker_b\",\"p95_latency_ms\":26417,\"pass_rate\":0.4538,\"provider_group\":\"provider_z\",\"quality\":0.7386,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.419,\"id\":\"worker_c\",\"p95_latency_ms\":64204,\"pass_rate\":0.931,\"provider_group\":\"provider_x\",\"quality\":0.9135,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.265,\"estimated_latency_ms\":132640,\"estimated_pass_rate\":0.8359,\"estimated_quality\":0.8675,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":18.726,\"max_latency_ms\":222136,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"credit_constrained\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4541, "category": "coding", "complexity": 0.8506, "consequence": 0.5773, "evidence_need": 0.7883, "family": "migration_planning", "input_tokens": 3240, "request_id": "synthetic-migration_planning-000208", "reversibility": 0.3327, "risk": 0.6417, "summary": "Design compatibility testing for a webhook delivery schema migration."}, "workers": [{"credits_per_call": 2.846, "id": "worker_a", "p95_latency_ms": 68436, "pass_rate": 0.7495, "provider_group": "provider_y", "quality": 0.8215, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.953, "id": "worker_b", "p95_latency_ms": 26417, "pass_rate": 0.4538, "provider_group": "provider_z", "quality": 0.7386, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.419, "id": "worker_c", "p95_latency_ms": 64204, "pass_rate": 0.931, "provider_group": "provider_x", "quality": 0.9135, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 9.701, "max_latency_ms": 170950, "minimum_quality": 0.7798}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000731", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.218, "estimated_latency_ms": 91913, "estimated_pass_rate": 0.7026, "estimated_quality": 0.856, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.701, "max_latency_ms": 170950, "policy": "balanced", "reason_codes": ["independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7026, "quality": 0.856, "quality_target_met": true, "utility": 63.483522}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":9.701,\"max_latency_ms\":170950,\"minimum_quality\":0.7798},\"task\":{\"ambiguity\":0.6335,\"category\":\"analysis\",\"complexity\":0.6379,\"consequence\":0.6979,\"evidence_need\":0.7385,\"family\":\"performance_tuning\",\"input_tokens\":7760,\"reversibility\":0.7527,\"risk\":0.4842,\"summary\":\"Review a proposed optimization for usage metering for hidden regressions.\"},\"workers\":[{\"credits_per_call\":1.185,\"id\":\"worker_a\",\"p95_latency_ms\":24378,\"pass_rate\":0.4394,\"provider_group\":\"provider_y\",\"quality\":0.6984,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":10.418,\"id\":\"worker_b\",\"p95_latency_ms\":63083,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.033,\"id\":\"worker_c\",\"p95_latency_ms\":67535,\"pass_rate\":0.8136,\"provider_group\":\"provider_z\",\"quality\":0.9535,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.218,\"estimated_latency_ms\":91913,\"estimated_pass_rate\":0.7026,\"estimated_quality\":0.856,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.701,\"max_latency_ms\":170950,\"policy\":\"balanced\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6335, "category": "analysis", "complexity": 0.6379, "consequence": 0.6979, "evidence_need": 0.7385, "family": "performance_tuning", "input_tokens": 7760, "request_id": "synthetic-performance_tuning-000731", "reversibility": 0.7527, "risk": 0.4842, "summary": "Review a proposed optimization for usage metering for hidden regressions."}, "workers": [{"credits_per_call": 1.185, "id": "worker_a", "p95_latency_ms": 24378, "pass_rate": 0.4394, "provider_group": "provider_y", "quality": 0.6984, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 10.418, "id": "worker_b", "p95_latency_ms": 63083, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.033, "id": "worker_c", "p95_latency_ms": 67535, "pass_rate": 0.8136, "provider_group": "provider_z", "quality": 0.9535, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.199, "max_latency_ms": 182246, "minimum_quality": 0.808}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000341", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.654, "estimated_latency_ms": 168177, "estimated_pass_rate": 0.8622, "estimated_quality": 0.9505, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 17.199, "max_latency_ms": 182246, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8622, "quality": 0.9505, "quality_target_met": true, "utility": 70.402734}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.199,\"max_latency_ms\":182246,\"minimum_quality\":0.808},\"task\":{\"ambiguity\":0.6117,\"category\":\"analysis\",\"complexity\":0.6823,\"consequence\":0.293,\"evidence_need\":0.3106,\"family\":\"performance_tuning\",\"input_tokens\":7160,\"reversibility\":0.6623,\"risk\":0.3059,\"summary\":\"Review a proposed optimization for OAuth callbacks for hidden regressions.\"},\"workers\":[{\"credits_per_call\":10.736,\"id\":\"worker_a\",\"p95_latency_ms\":71506,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9926,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.827,\"id\":\"worker_b\",\"p95_latency_ms\":23154,\"pass_rate\":0.5329,\"provider_group\":\"provider_x\",\"quality\":0.7387,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":4.091,\"id\":\"worker_c\",\"p95_latency_ms\":73517,\"pass_rate\":0.8873,\"provider_group\":\"provider_z\",\"quality\":0.9765,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.654,\"estimated_latency_ms\":168177,\"estimated_pass_rate\":0.8622,\"estimated_quality\":0.9505,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.199,\"max_latency_ms\":182246,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6117, "category": "analysis", "complexity": 0.6823, "consequence": 0.293, "evidence_need": 0.3106, "family": "performance_tuning", "input_tokens": 7160, "request_id": "synthetic-performance_tuning-000341", "reversibility": 0.6623, "risk": 0.3059, "summary": "Review a proposed optimization for OAuth callbacks for hidden regressions."}, "workers": [{"credits_per_call": 10.736, "id": "worker_a", "p95_latency_ms": 71506, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9926, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.827, "id": "worker_b", "p95_latency_ms": 23154, "pass_rate": 0.5329, "provider_group": "provider_x", "quality": 0.7387, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 4.091, "id": "worker_c", "p95_latency_ms": 73517, "pass_rate": 0.8873, "provider_group": "provider_z", "quality": 0.9765, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.694, "max_latency_ms": 266612, "minimum_quality": 0.8608}, "family": "migration_planning", "id": "synthetic-migration_planning-000334", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.526, "estimated_latency_ms": 186512, "estimated_pass_rate": 0.7181, "estimated_quality": 0.8935, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 18.694, "max_latency_ms": 266612, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7181, "quality": 0.8935, "quality_target_met": true, "utility": 70.08865}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.694,\"max_latency_ms\":266612,\"minimum_quality\":0.8608},\"task\":{\"ambiguity\":0.4742,\"category\":\"coding\",\"complexity\":0.9285,\"consequence\":0.5567,\"evidence_need\":0.4838,\"family\":\"migration_planning\",\"input_tokens\":1800,\"reversibility\":0.5311,\"risk\":0.4053,\"summary\":\"Review a staged migration for webhook delivery and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.062,\"id\":\"worker_a\",\"p95_latency_ms\":22301,\"pass_rate\":0.5276,\"provider_group\":\"provider_x\",\"quality\":0.7782,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.473,\"id\":\"worker_b\",\"p95_latency_ms\":82004,\"pass_rate\":0.8205,\"provider_group\":\"provider_z\",\"quality\":0.8314,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.991,\"id\":\"worker_c\",\"p95_latency_ms\":82207,\"pass_rate\":0.9557,\"provider_group\":\"provider_y\",\"quality\":0.9512,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.526,\"estimated_latency_ms\":186512,\"estimated_pass_rate\":0.7181,\"estimated_quality\":0.8935,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.694,\"max_latency_ms\":266612,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4742, "category": "coding", "complexity": 0.9285, "consequence": 0.5567, "evidence_need": 0.4838, "family": "migration_planning", "input_tokens": 1800, "request_id": "synthetic-migration_planning-000334", "reversibility": 0.5311, "risk": 0.4053, "summary": "Review a staged migration for webhook delivery and identify irreversible steps."}, "workers": [{"credits_per_call": 1.062, "id": "worker_a", "p95_latency_ms": 22301, "pass_rate": 0.5276, "provider_group": "provider_x", "quality": 0.7782, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.473, "id": "worker_b", "p95_latency_ms": 82004, "pass_rate": 0.8205, "provider_group": "provider_z", "quality": 0.8314, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.991, "id": "worker_c", "p95_latency_ms": 82207, "pass_rate": 0.9557, "provider_group": "provider_y", "quality": 0.9512, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 1.731, "max_latency_ms": 36917, "minimum_quality": 0.5414}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000236", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.025, "estimated_latency_ms": 29951, "estimated_pass_rate": 0.3425, "estimated_quality": 0.543, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 1.731, "max_latency_ms": 36917, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3425, "quality": 0.543, "quality_target_met": true, "utility": 33.747618}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":1.731,\"max_latency_ms\":36917,\"minimum_quality\":0.5414},\"task\":{\"ambiguity\":0.4081,\"category\":\"analysis\",\"complexity\":0.8849,\"consequence\":0.5442,\"evidence_need\":0.4582,\"family\":\"performance_tuning\",\"input_tokens\":7200,\"reversibility\":0.6532,\"risk\":0.261,\"summary\":\"Diagnose latency variance in background jobs and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":3.889,\"id\":\"worker_a\",\"p95_latency_ms\":53135,\"pass_rate\":0.844,\"provider_group\":\"provider_x\",\"quality\":0.8788,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.025,\"id\":\"worker_b\",\"p95_latency_ms\":29951,\"pass_rate\":0.4948,\"provider_group\":\"provider_z\",\"quality\":0.6692,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":11.956,\"id\":\"worker_c\",\"p95_latency_ms\":87491,\"pass_rate\":0.9599,\"provider_group\":\"provider_y\",\"quality\":0.9814,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.025,\"estimated_latency_ms\":29951,\"estimated_pass_rate\":0.3425,\"estimated_quality\":0.543,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":1.731,\"max_latency_ms\":36917,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4081, "category": "analysis", "complexity": 0.8849, "consequence": 0.5442, "evidence_need": 0.4582, "family": "performance_tuning", "input_tokens": 7200, "request_id": "synthetic-performance_tuning-000236", "reversibility": 0.6532, "risk": 0.261, "summary": "Diagnose latency variance in background jobs and rank the next experiments."}, "workers": [{"credits_per_call": 3.889, "id": "worker_a", "p95_latency_ms": 53135, "pass_rate": 0.844, "provider_group": "provider_x", "quality": 0.8788, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.025, "id": "worker_b", "p95_latency_ms": 29951, "pass_rate": 0.4948, "provider_group": "provider_z", "quality": 0.6692, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 11.956, "id": "worker_c", "p95_latency_ms": 87491, "pass_rate": 0.9599, "provider_group": "provider_y", "quality": 0.9814, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.722, "max_latency_ms": 47317, "minimum_quality": 0.7075}, "family": "migration_planning", "id": "synthetic-migration_planning-000709", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.819, "estimated_latency_ms": 30379, "estimated_pass_rate": 0.3024, "estimated_quality": 0.6157, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.722, "max_latency_ms": 47317, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3024, "quality": 0.6157, "quality_target_met": false, "utility": 20.252718}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.722,\"max_latency_ms\":47317,\"minimum_quality\":0.7075},\"task\":{\"ambiguity\":0.4637,\"category\":\"coding\",\"complexity\":0.9213,\"consequence\":0.8476,\"evidence_need\":0.8268,\"family\":\"migration_planning\",\"input_tokens\":1600,\"reversibility\":0.4863,\"risk\":0.7323,\"summary\":\"Review a staged migration for provider failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":0.819,\"id\":\"worker_a\",\"p95_latency_ms\":30379,\"pass_rate\":0.5012,\"provider_group\":\"provider_x\",\"quality\":0.7862,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.473,\"id\":\"worker_b\",\"p95_latency_ms\":55842,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9207,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.219,\"id\":\"worker_c\",\"p95_latency_ms\":57092,\"pass_rate\":0.8739,\"provider_group\":\"provider_z\",\"quality\":0.8423,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.819,\"estimated_latency_ms\":30379,\"estimated_pass_rate\":0.3024,\"estimated_quality\":0.6157,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.722,\"max_latency_ms\":47317,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4637, "category": "coding", "complexity": 0.9213, "consequence": 0.8476, "evidence_need": 0.8268, "family": "migration_planning", "input_tokens": 1600, "request_id": "synthetic-migration_planning-000709", "reversibility": 0.4863, "risk": 0.7323, "summary": "Review a staged migration for provider failover and identify irreversible steps."}, "workers": [{"credits_per_call": 0.819, "id": "worker_a", "p95_latency_ms": 30379, "pass_rate": 0.5012, "provider_group": "provider_x", "quality": 0.7862, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.473, "id": "worker_b", "p95_latency_ms": 55842, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9207, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.219, "id": "worker_c", "p95_latency_ms": 57092, "pass_rate": 0.8739, "provider_group": "provider_z", "quality": 0.8423, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.38, "max_latency_ms": 180947, "minimum_quality": 0.8012}, "family": "migration_planning", "id": "synthetic-migration_planning-000982", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.63, "estimated_latency_ms": 126250, "estimated_pass_rate": 0.7658, "estimated_quality": 0.8879, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 21.38, "max_latency_ms": 180947, "policy": "strict", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7658, "quality": 0.8879, "quality_target_met": true, "utility": 73.293146}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.38,\"max_latency_ms\":180947,\"minimum_quality\":0.8012},\"task\":{\"ambiguity\":0.7145,\"category\":\"coding\",\"complexity\":0.7115,\"consequence\":0.4989,\"evidence_need\":0.7078,\"family\":\"migration_planning\",\"input_tokens\":1080,\"reversibility\":0.4987,\"risk\":0.5614,\"summary\":\"Design compatibility testing for a OAuth callbacks schema migration.\"},\"workers\":[{\"credits_per_call\":13.243,\"id\":\"worker_a\",\"p95_latency_ms\":63021,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9129,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.13,\"id\":\"worker_b\",\"p95_latency_ms\":17915,\"pass_rate\":0.5798,\"provider_group\":\"provider_z\",\"quality\":0.7847,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.257,\"id\":\"worker_c\",\"p95_latency_ms\":45314,\"pass_rate\":0.7662,\"provider_group\":\"provider_x\",\"quality\":0.8931,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.63,\"estimated_latency_ms\":126250,\"estimated_pass_rate\":0.7658,\"estimated_quality\":0.8879,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.38,\"max_latency_ms\":180947,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7145, "category": "coding", "complexity": 0.7115, "consequence": 0.4989, "evidence_need": 0.7078, "family": "migration_planning", "input_tokens": 1080, "request_id": "synthetic-migration_planning-000982", "reversibility": 0.4987, "risk": 0.5614, "summary": "Design compatibility testing for a OAuth callbacks schema migration."}, "workers": [{"credits_per_call": 13.243, "id": "worker_a", "p95_latency_ms": 63021, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9129, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.13, "id": "worker_b", "p95_latency_ms": 17915, "pass_rate": 0.5798, "provider_group": "provider_z", "quality": 0.7847, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.257, "id": "worker_c", "p95_latency_ms": 45314, "pass_rate": 0.7662, "provider_group": "provider_x", "quality": 0.8931, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.568, "max_latency_ms": 61969, "minimum_quality": 0.5689}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000777", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.095, "estimated_latency_ms": 31294, "estimated_pass_rate": 0.3635, "estimated_quality": 0.6427, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.568, "max_latency_ms": 61969, "policy": "lean", "reason_codes": ["complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3635, "quality": 0.6427, "quality_target_met": true, "utility": 43.787163}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.568,\"max_latency_ms\":61969,\"minimum_quality\":0.5689},\"task\":{\"ambiguity\":0.6118,\"category\":\"review\",\"complexity\":0.7216,\"consequence\":0.6728,\"evidence_need\":0.5621,\"family\":\"policy_compliance\",\"input_tokens\":5800,\"reversibility\":0.4677,\"risk\":0.6818,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":10.556,\"id\":\"worker_a\",\"p95_latency_ms\":82002,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9355,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.255,\"id\":\"worker_b\",\"p95_latency_ms\":74639,\"pass_rate\":0.8073,\"provider_group\":\"provider_y\",\"quality\":0.8557,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.095,\"id\":\"worker_c\",\"p95_latency_ms\":31294,\"pass_rate\":0.5468,\"provider_group\":\"provider_z\",\"quality\":0.7834,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.095,\"estimated_latency_ms\":31294,\"estimated_pass_rate\":0.3635,\"estimated_quality\":0.6427,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.568,\"max_latency_ms\":61969,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6118, "category": "review", "complexity": 0.7216, "consequence": 0.6728, "evidence_need": 0.5621, "family": "policy_compliance", "input_tokens": 5800, "request_id": "synthetic-policy_compliance-000777", "reversibility": 0.4677, "risk": 0.6818, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 10.556, "id": "worker_a", "p95_latency_ms": 82002, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9355, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.255, "id": "worker_b", "p95_latency_ms": 74639, "pass_rate": 0.8073, "provider_group": "provider_y", "quality": 0.8557, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.095, "id": "worker_c", "p95_latency_ms": 31294, "pass_rate": 0.5468, "provider_group": "provider_z", "quality": 0.7834, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.675, "max_latency_ms": 47698, "minimum_quality": 0.5054}, "family": "migration_planning", "id": "synthetic-migration_planning-000487", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.888, "estimated_latency_ms": 29426, "estimated_pass_rate": 0.3301, "estimated_quality": 0.5304, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.675, "max_latency_ms": 47698, "policy": "lean", "reason_codes": ["complex_task", "high_consequence"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3301, "quality": 0.5304, "quality_target_met": true, "utility": 35.377208}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.675,\"max_latency_ms\":47698,\"minimum_quality\":0.5054},\"task\":{\"ambiguity\":0.3487,\"category\":\"coding\",\"complexity\":0.9454,\"consequence\":0.8286,\"evidence_need\":0.587,\"family\":\"migration_planning\",\"input_tokens\":6400,\"reversibility\":0.4197,\"risk\":0.581,\"summary\":\"Plan a zero-downtime migration of schema evolution with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.888,\"id\":\"worker_a\",\"p95_latency_ms\":29426,\"pass_rate\":0.5107,\"provider_group\":\"provider_x\",\"quality\":0.6866,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.727,\"id\":\"worker_b\",\"p95_latency_ms\":54658,\"pass_rate\":0.7444,\"provider_group\":\"provider_z\",\"quality\":0.9142,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.671,\"id\":\"worker_c\",\"p95_latency_ms\":85709,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9162,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.888,\"estimated_latency_ms\":29426,\"estimated_pass_rate\":0.3301,\"estimated_quality\":0.5304,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.675,\"max_latency_ms\":47698,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3487, "category": "coding", "complexity": 0.9454, "consequence": 0.8286, "evidence_need": 0.587, "family": "migration_planning", "input_tokens": 6400, "request_id": "synthetic-migration_planning-000487", "reversibility": 0.4197, "risk": 0.581, "summary": "Plan a zero-downtime migration of schema evolution with rollback checkpoints."}, "workers": [{"credits_per_call": 0.888, "id": "worker_a", "p95_latency_ms": 29426, "pass_rate": 0.5107, "provider_group": "provider_x", "quality": 0.6866, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.727, "id": "worker_b", "p95_latency_ms": 54658, "pass_rate": 0.7444, "provider_group": "provider_z", "quality": 0.9142, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 12.671, "id": "worker_c", "p95_latency_ms": 85709, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9162, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 25.327, "max_latency_ms": 207184, "minimum_quality": 0.8481}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001046", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.888, "estimated_latency_ms": 172387, "estimated_pass_rate": 0.7954, "estimated_quality": 0.9279, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 25.327, "max_latency_ms": 207184, "policy": "strict", "reason_codes": ["complex_task", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7954, "quality": 0.9279, "quality_target_met": true, "utility": 74.057064}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":25.327,\"max_latency_ms\":207184,\"minimum_quality\":0.8481},\"task\":{\"ambiguity\":0.5844,\"category\":\"analysis\",\"complexity\":0.7528,\"consequence\":0.3889,\"evidence_need\":0.4687,\"family\":\"performance_tuning\",\"input_tokens\":4440,\"reversibility\":0.7286,\"risk\":0.5415,\"summary\":\"Diagnose latency variance in observability and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.137,\"id\":\"worker_a\",\"p95_latency_ms\":28201,\"pass_rate\":0.534,\"provider_group\":\"provider_x\",\"quality\":0.7372,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.992,\"id\":\"worker_b\",\"p95_latency_ms\":79308,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.759,\"id\":\"worker_c\",\"p95_latency_ms\":64878,\"pass_rate\":0.9246,\"provider_group\":\"provider_z\",\"quality\":0.9717,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.888,\"estimated_latency_ms\":172387,\"estimated_pass_rate\":0.7954,\"estimated_quality\":0.9279,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":25.327,\"max_latency_ms\":207184,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5844, "category": "analysis", "complexity": 0.7528, "consequence": 0.3889, "evidence_need": 0.4687, "family": "performance_tuning", "input_tokens": 4440, "request_id": "synthetic-performance_tuning-001046", "reversibility": 0.7286, "risk": 0.5415, "summary": "Diagnose latency variance in observability and rank the next experiments."}, "workers": [{"credits_per_call": 1.137, "id": "worker_a", "p95_latency_ms": 28201, "pass_rate": 0.534, "provider_group": "provider_x", "quality": 0.7372, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.992, "id": "worker_b", "p95_latency_ms": 79308, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.759, "id": "worker_c", "p95_latency_ms": 64878, "pass_rate": 0.9246, "provider_group": "provider_z", "quality": 0.9717, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.735, "max_latency_ms": 246353, "minimum_quality": 0.9025}, "family": "migration_planning", "id": "synthetic-migration_planning-000919", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.618, "estimated_latency_ms": 175737, "estimated_pass_rate": 0.7685, "estimated_quality": 0.9045, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.735, "max_latency_ms": 246353, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7685, "quality": 0.9045, "quality_target_met": true, "utility": 76.852002}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.735,\"max_latency_ms\":246353,\"minimum_quality\":0.9025},\"task\":{\"ambiguity\":0.3377,\"category\":\"coding\",\"complexity\":0.9494,\"consequence\":0.6698,\"evidence_need\":0.6623,\"family\":\"migration_planning\",\"input_tokens\":3680,\"reversibility\":0.356,\"risk\":0.6766,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.896,\"id\":\"worker_a\",\"p95_latency_ms\":21839,\"pass_rate\":0.5948,\"provider_group\":\"provider_y\",\"quality\":0.7562,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.675,\"id\":\"worker_b\",\"p95_latency_ms\":75354,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9786,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.047,\"id\":\"worker_c\",\"p95_latency_ms\":78544,\"pass_rate\":0.7345,\"provider_group\":\"provider_z\",\"quality\":0.8698,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.618,\"estimated_latency_ms\":175737,\"estimated_pass_rate\":0.7685,\"estimated_quality\":0.9045,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.735,\"max_latency_ms\":246353,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3377, "category": "coding", "complexity": 0.9494, "consequence": 0.6698, "evidence_need": 0.6623, "family": "migration_planning", "input_tokens": 3680, "request_id": "synthetic-migration_planning-000919", "reversibility": 0.356, "risk": 0.6766, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 0.896, "id": "worker_a", "p95_latency_ms": 21839, "pass_rate": 0.5948, "provider_group": "provider_y", "quality": 0.7562, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.675, "id": "worker_b", "p95_latency_ms": 75354, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9786, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.047, "id": "worker_c", "p95_latency_ms": 78544, "pass_rate": 0.7345, "provider_group": "provider_z", "quality": 0.8698, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.635, "max_latency_ms": 225002, "minimum_quality": 0.9586}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000405", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.871, "estimated_latency_ms": 166232, "estimated_pass_rate": 0.7476, "estimated_quality": 0.8848, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.635, "max_latency_ms": 225002, "policy": "strict", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7476, "quality": 0.8848, "quality_target_met": false, "utility": 72.535058}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.635,\"max_latency_ms\":225002,\"minimum_quality\":0.9586},\"task\":{\"ambiguity\":0.6144,\"category\":\"review\",\"complexity\":0.8093,\"consequence\":0.6031,\"evidence_need\":0.8784,\"family\":\"policy_compliance\",\"input_tokens\":2640,\"reversibility\":0.3071,\"risk\":0.5784,\"summary\":\"Audit a workflow for usage metering against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":1.137,\"id\":\"worker_a\",\"p95_latency_ms\":27779,\"pass_rate\":0.5016,\"provider_group\":\"provider_y\",\"quality\":0.6944,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.729,\"id\":\"worker_b\",\"p95_latency_ms\":74296,\"pass_rate\":0.9864,\"provider_group\":\"provider_x\",\"quality\":0.9136,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.005,\"id\":\"worker_c\",\"p95_latency_ms\":64157,\"pass_rate\":0.8207,\"provider_group\":\"provider_z\",\"quality\":0.913,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.871,\"estimated_latency_ms\":166232,\"estimated_pass_rate\":0.7476,\"estimated_quality\":0.8848,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.635,\"max_latency_ms\":225002,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6144, "category": "review", "complexity": 0.8093, "consequence": 0.6031, "evidence_need": 0.8784, "family": "policy_compliance", "input_tokens": 2640, "request_id": "synthetic-policy_compliance-000405", "reversibility": 0.3071, "risk": 0.5784, "summary": "Audit a workflow for usage metering against explicit access requirements."}, "workers": [{"credits_per_call": 1.137, "id": "worker_a", "p95_latency_ms": 27779, "pass_rate": 0.5016, "provider_group": "provider_y", "quality": 0.6944, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.729, "id": "worker_b", "p95_latency_ms": 74296, "pass_rate": 0.9864, "provider_group": "provider_x", "quality": 0.9136, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.005, "id": "worker_c", "p95_latency_ms": 64157, "pass_rate": 0.8207, "provider_group": "provider_z", "quality": 0.913, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.097, "max_latency_ms": 185970, "minimum_quality": 0.899}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001044", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.545, "estimated_latency_ms": 106458, "estimated_pass_rate": 0.7059, "estimated_quality": 0.9141, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 17.097, "max_latency_ms": 185970, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7059, "quality": 0.9141, "quality_target_met": true, "utility": 68.608935}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.097,\"max_latency_ms\":185970,\"minimum_quality\":0.899},\"task\":{\"ambiguity\":0.6092,\"category\":\"review\",\"complexity\":0.5627,\"consequence\":0.6357,\"evidence_need\":0.8825,\"family\":\"policy_compliance\",\"input_tokens\":7120,\"reversibility\":0.3672,\"risk\":0.7931,\"summary\":\"Identify evidence required before approving access control for production.\"},\"workers\":[{\"credits_per_call\":3.571,\"id\":\"worker_a\",\"p95_latency_ms\":48907,\"pass_rate\":0.7377,\"provider_group\":\"provider_x\",\"quality\":0.8337,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.833,\"id\":\"worker_b\",\"p95_latency_ms\":30492,\"pass_rate\":0.4949,\"provider_group\":\"provider_y\",\"quality\":0.8034,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.712,\"id\":\"worker_c\",\"p95_latency_ms\":75966,\"pass_rate\":0.9381,\"provider_group\":\"provider_z\",\"quality\":0.96,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.545,\"estimated_latency_ms\":106458,\"estimated_pass_rate\":0.7059,\"estimated_quality\":0.9141,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":17.097,\"max_latency_ms\":185970,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6092, "category": "review", "complexity": 0.5627, "consequence": 0.6357, "evidence_need": 0.8825, "family": "policy_compliance", "input_tokens": 7120, "request_id": "synthetic-policy_compliance-001044", "reversibility": 0.3672, "risk": 0.7931, "summary": "Identify evidence required before approving access control for production."}, "workers": [{"credits_per_call": 3.571, "id": "worker_a", "p95_latency_ms": 48907, "pass_rate": 0.7377, "provider_group": "provider_x", "quality": 0.8337, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.833, "id": "worker_b", "p95_latency_ms": 30492, "pass_rate": 0.4949, "provider_group": "provider_y", "quality": 0.8034, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.712, "id": "worker_c", "p95_latency_ms": 75966, "pass_rate": 0.9381, "provider_group": "provider_z", "quality": 0.96, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.23, "max_latency_ms": 58799, "minimum_quality": 0.7032}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000647", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.93, "estimated_latency_ms": 29426, "estimated_pass_rate": 0.2854, "estimated_quality": 0.5435, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.23, "max_latency_ms": 58799, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.2854, "quality": 0.5435, "quality_target_met": false, "utility": 29.286292}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.23,\"max_latency_ms\":58799,\"minimum_quality\":0.7032},\"task\":{\"ambiguity\":0.4436,\"category\":\"analysis\",\"complexity\":0.8151,\"consequence\":0.4917,\"evidence_need\":0.5182,\"family\":\"performance_tuning\",\"input_tokens\":3760,\"reversibility\":0.6748,\"risk\":0.4153,\"summary\":\"Diagnose latency variance in rate limiting and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":13.716,\"id\":\"worker_a\",\"p95_latency_ms\":83393,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.93,\"id\":\"worker_b\",\"p95_latency_ms\":29426,\"pass_rate\":0.445,\"provider_group\":\"provider_x\",\"quality\":0.6706,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.115,\"id\":\"worker_c\",\"p95_latency_ms\":61984,\"pass_rate\":0.8897,\"provider_group\":\"provider_y\",\"quality\":0.8726,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.93,\"estimated_latency_ms\":29426,\"estimated_pass_rate\":0.2854,\"estimated_quality\":0.5435,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.23,\"max_latency_ms\":58799,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.4436, "category": "analysis", "complexity": 0.8151, "consequence": 0.4917, "evidence_need": 0.5182, "family": "performance_tuning", "input_tokens": 3760, "request_id": "synthetic-performance_tuning-000647", "reversibility": 0.6748, "risk": 0.4153, "summary": "Diagnose latency variance in rate limiting and rank the next experiments."}, "workers": [{"credits_per_call": 13.716, "id": "worker_a", "p95_latency_ms": 83393, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.93, "id": "worker_b", "p95_latency_ms": 29426, "pass_rate": 0.445, "provider_group": "provider_x", "quality": 0.6706, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.115, "id": "worker_c", "p95_latency_ms": 61984, "pass_rate": 0.8897, "provider_group": "provider_y", "quality": 0.8726, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.987, "max_latency_ms": 203481, "minimum_quality": 0.7875}, "family": "migration_planning", "id": "synthetic-migration_planning-001111", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.381, "estimated_latency_ms": 134892, "estimated_pass_rate": 0.8256, "estimated_quality": 0.9225, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.987, "max_latency_ms": 203481, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8256, "quality": 0.9225, "quality_target_met": true, "utility": 69.1187}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.987,\"max_latency_ms\":203481,\"minimum_quality\":0.7875},\"task\":{\"ambiguity\":0.6161,\"category\":\"coding\",\"complexity\":0.5552,\"consequence\":0.5788,\"evidence_need\":0.6989,\"family\":\"migration_planning\",\"input_tokens\":3880,\"reversibility\":0.413,\"risk\":0.6942,\"summary\":\"Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.925,\"id\":\"worker_a\",\"p95_latency_ms\":29653,\"pass_rate\":0.6037,\"provider_group\":\"provider_x\",\"quality\":0.7678,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.958,\"id\":\"worker_b\",\"p95_latency_ms\":84833,\"pass_rate\":0.9038,\"provider_group\":\"provider_y\",\"quality\":0.92,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.423,\"id\":\"worker_c\",\"p95_latency_ms\":50059,\"pass_rate\":0.8207,\"provider_group\":\"provider_z\",\"quality\":0.9208,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.381,\"estimated_latency_ms\":134892,\"estimated_pass_rate\":0.8256,\"estimated_quality\":0.9225,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.987,\"max_latency_ms\":203481,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6161, "category": "coding", "complexity": 0.5552, "consequence": 0.5788, "evidence_need": 0.6989, "family": "migration_planning", "input_tokens": 3880, "request_id": "synthetic-migration_planning-001111", "reversibility": 0.413, "risk": 0.6942, "summary": "Plan a zero-downtime migration of tenant-scoped caching with rollback checkpoints."}, "workers": [{"credits_per_call": 0.925, "id": "worker_a", "p95_latency_ms": 29653, "pass_rate": 0.6037, "provider_group": "provider_x", "quality": 0.7678, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.958, "id": "worker_b", "p95_latency_ms": 84833, "pass_rate": 0.9038, "provider_group": "provider_y", "quality": 0.92, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.423, "id": "worker_c", "p95_latency_ms": 50059, "pass_rate": 0.8207, "provider_group": "provider_z", "quality": 0.9208, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.942, "max_latency_ms": 34565, "minimum_quality": 0.5743}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000108", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.821, "estimated_latency_ms": 21768, "estimated_pass_rate": 0.3132, "estimated_quality": 0.6032, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.942, "max_latency_ms": 34565, "policy": "lean", "reason_codes": ["complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3132, "quality": 0.6032, "quality_target_met": true, "utility": 26.535555}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.942,\"max_latency_ms\":34565,\"minimum_quality\":0.5743},\"task\":{\"ambiguity\":0.7261,\"category\":\"review\",\"complexity\":0.409,\"consequence\":0.8437,\"evidence_need\":0.9188,\"family\":\"policy_compliance\",\"input_tokens\":2760,\"reversibility\":0.4551,\"risk\":0.6052,\"summary\":\"Identify evidence required before approving observability for production.\"},\"workers\":[{\"credits_per_call\":2.926,\"id\":\"worker_a\",\"p95_latency_ms\":68031,\"pass_rate\":0.8442,\"provider_group\":\"provider_x\",\"quality\":0.8472,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.821,\"id\":\"worker_b\",\"p95_latency_ms\":21768,\"pass_rate\":0.4635,\"provider_group\":\"provider_y\",\"quality\":0.6895,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.344,\"id\":\"worker_c\",\"p95_latency_ms\":61046,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.821,\"estimated_latency_ms\":21768,\"estimated_pass_rate\":0.3132,\"estimated_quality\":0.6032,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.942,\"max_latency_ms\":34565,\"policy\":\"lean\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.7261, "category": "review", "complexity": 0.409, "consequence": 0.8437, "evidence_need": 0.9188, "family": "policy_compliance", "input_tokens": 2760, "request_id": "synthetic-policy_compliance-000108", "reversibility": 0.4551, "risk": 0.6052, "summary": "Identify evidence required before approving observability for production."}, "workers": [{"credits_per_call": 2.926, "id": "worker_a", "p95_latency_ms": 68031, "pass_rate": 0.8442, "provider_group": "provider_x", "quality": 0.8472, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 0.821, "id": "worker_b", "p95_latency_ms": 21768, "pass_rate": 0.4635, "provider_group": "provider_y", "quality": 0.6895, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.344, "id": "worker_c", "p95_latency_ms": 61046, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.245, "max_latency_ms": 185797, "minimum_quality": 0.815}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001254", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.394, "estimated_latency_ms": 95069, "estimated_pass_rate": 0.684, "estimated_quality": 0.8216, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.245, "max_latency_ms": 185797, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.684, "quality": 0.8216, "quality_target_met": true, "utility": 63.792016}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.245,\"max_latency_ms\":185797,\"minimum_quality\":0.815},\"task\":{\"ambiguity\":0.5314,\"category\":\"review\",\"complexity\":0.4514,\"consequence\":0.5609,\"evidence_need\":0.6272,\"family\":\"policy_compliance\",\"input_tokens\":4200,\"reversibility\":0.3705,\"risk\":0.7008,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":10.646,\"id\":\"worker_a\",\"p95_latency_ms\":89719,\"pass_rate\":0.9123,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.299,\"id\":\"worker_b\",\"p95_latency_ms\":76179,\"pass_rate\":0.8535,\"provider_group\":\"provider_y\",\"quality\":0.7935,\"strengths\":[\"drafting\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.095,\"id\":\"worker_c\",\"p95_latency_ms\":18890,\"pass_rate\":0.6184,\"provider_group\":\"provider_z\",\"quality\":0.7559,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.394,\"estimated_latency_ms\":95069,\"estimated_pass_rate\":0.684,\"estimated_quality\":0.8216,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.245,\"max_latency_ms\":185797,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5314, "category": "review", "complexity": 0.4514, "consequence": 0.5609, "evidence_need": 0.6272, "family": "policy_compliance", "input_tokens": 4200, "request_id": "synthetic-policy_compliance-001254", "reversibility": 0.3705, "risk": 0.7008, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 10.646, "id": "worker_a", "p95_latency_ms": 89719, "pass_rate": 0.9123, "provider_group": "provider_x", "quality": 0.995, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.299, "id": "worker_b", "p95_latency_ms": 76179, "pass_rate": 0.8535, "provider_group": "provider_y", "quality": 0.7935, "strengths": ["drafting", "reasoning", "review", "writing"]}, {"credits_per_call": 1.095, "id": "worker_c", "p95_latency_ms": 18890, "pass_rate": 0.6184, "provider_group": "provider_z", "quality": 0.7559, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.674, "max_latency_ms": 193303, "minimum_quality": 0.9353}, "family": "migration_planning", "id": "synthetic-migration_planning-000664", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.863, "estimated_latency_ms": 149549, "estimated_pass_rate": 0.715, "estimated_quality": 0.8855, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 23.674, "max_latency_ms": 193303, "policy": "strict", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.715, "quality": 0.8855, "quality_target_met": false, "utility": 69.20565}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.674,\"max_latency_ms\":193303,\"minimum_quality\":0.9353},\"task\":{\"ambiguity\":0.4046,\"category\":\"coding\",\"complexity\":0.7155,\"consequence\":0.6108,\"evidence_need\":0.7488,\"family\":\"migration_planning\",\"input_tokens\":3280,\"reversibility\":0.3828,\"risk\":0.5779,\"summary\":\"Review a staged migration for private credentials and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":13.257,\"id\":\"worker_a\",\"p95_latency_ms\":74027,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9159,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.677,\"id\":\"worker_b\",\"p95_latency_ms\":47386,\"pass_rate\":0.7412,\"provider_group\":\"provider_z\",\"quality\":0.857,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.929,\"id\":\"worker_c\",\"p95_latency_ms\":28136,\"pass_rate\":0.4533,\"provider_group\":\"provider_y\",\"quality\":0.7659,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.863,\"estimated_latency_ms\":149549,\"estimated_pass_rate\":0.715,\"estimated_quality\":0.8855,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.674,\"max_latency_ms\":193303,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4046, "category": "coding", "complexity": 0.7155, "consequence": 0.6108, "evidence_need": 0.7488, "family": "migration_planning", "input_tokens": 3280, "request_id": "synthetic-migration_planning-000664", "reversibility": 0.3828, "risk": 0.5779, "summary": "Review a staged migration for private credentials and identify irreversible steps."}, "workers": [{"credits_per_call": 13.257, "id": "worker_a", "p95_latency_ms": 74027, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9159, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.677, "id": "worker_b", "p95_latency_ms": 47386, "pass_rate": 0.7412, "provider_group": "provider_z", "quality": 0.857, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.929, "id": "worker_c", "p95_latency_ms": 28136, "pass_rate": 0.4533, "provider_group": "provider_y", "quality": 0.7659, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.687, "max_latency_ms": 173833, "minimum_quality": 0.879}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000197", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.864, "estimated_latency_ms": 140851, "estimated_pass_rate": 0.7843, "estimated_quality": 0.9245, "finalizer": "worker_c", "independent_verification": true, "max_calls": 3, "max_credits": 17.687, "max_latency_ms": 173833, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7843, "quality": 0.9245, "quality_target_met": true, "utility": 73.323623}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.687,\"max_latency_ms\":173833,\"minimum_quality\":0.879},\"task\":{\"ambiguity\":0.5794,\"category\":\"analysis\",\"complexity\":0.8427,\"consequence\":0.4997,\"evidence_need\":0.4396,\"family\":\"performance_tuning\",\"input_tokens\":1400,\"reversibility\":0.6112,\"risk\":0.5103,\"summary\":\"Diagnose latency variance in model routing and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.204,\"id\":\"worker_a\",\"p95_latency_ms\":29315,\"pass_rate\":0.4979,\"provider_group\":\"provider_z\",\"quality\":0.7191,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.182,\"id\":\"worker_b\",\"p95_latency_ms\":57269,\"pass_rate\":0.9508,\"provider_group\":\"provider_x\",\"quality\":0.9757,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":12.478,\"id\":\"worker_c\",\"p95_latency_ms\":54267,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.864,\"estimated_latency_ms\":140851,\"estimated_pass_rate\":0.7843,\"estimated_quality\":0.9245,\"finalizer\":\"worker_c\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.687,\"max_latency_ms\":173833,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5794, "category": "analysis", "complexity": 0.8427, "consequence": 0.4997, "evidence_need": 0.4396, "family": "performance_tuning", "input_tokens": 1400, "request_id": "synthetic-performance_tuning-000197", "reversibility": 0.6112, "risk": 0.5103, "summary": "Diagnose latency variance in model routing and rank the next experiments."}, "workers": [{"credits_per_call": 1.204, "id": "worker_a", "p95_latency_ms": 29315, "pass_rate": 0.4979, "provider_group": "provider_z", "quality": 0.7191, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.182, "id": "worker_b", "p95_latency_ms": 57269, "pass_rate": 0.9508, "provider_group": "provider_x", "quality": 0.9757, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 12.478, "id": "worker_c", "p95_latency_ms": 54267, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.819, "max_latency_ms": 31099, "minimum_quality": 0.657}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000653", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.909, "estimated_latency_ms": 20014, "estimated_pass_rate": 0.262, "estimated_quality": 0.6294, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.819, "max_latency_ms": 31099, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_a"}, "oracle": {"candidate_count": 1, "pass_rate": 0.262, "quality": 0.6294, "quality_target_met": false, "utility": 38.736407}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.819,\"max_latency_ms\":31099,\"minimum_quality\":0.657},\"task\":{\"ambiguity\":0.7491,\"category\":\"analysis\",\"complexity\":0.542,\"consequence\":0.2537,\"evidence_need\":0.6415,\"family\":\"performance_tuning\",\"input_tokens\":3080,\"reversibility\":0.7064,\"risk\":0.2015,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":0.909,\"id\":\"worker_a\",\"p95_latency_ms\":20014,\"pass_rate\":0.3964,\"provider_group\":\"provider_z\",\"quality\":0.7122,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":3.747,\"id\":\"worker_b\",\"p95_latency_ms\":79136,\"pass_rate\":0.8576,\"provider_group\":\"provider_y\",\"quality\":0.8685,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.972,\"id\":\"worker_c\",\"p95_latency_ms\":80469,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.909,\"estimated_latency_ms\":20014,\"estimated_pass_rate\":0.262,\"estimated_quality\":0.6294,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.819,\"max_latency_ms\":31099,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.7491, "category": "analysis", "complexity": 0.542, "consequence": 0.2537, "evidence_need": 0.6415, "family": "performance_tuning", "input_tokens": 3080, "request_id": "synthetic-performance_tuning-000653", "reversibility": 0.7064, "risk": 0.2015, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 0.909, "id": "worker_a", "p95_latency_ms": 20014, "pass_rate": 0.3964, "provider_group": "provider_z", "quality": 0.7122, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 3.747, "id": "worker_b", "p95_latency_ms": 79136, "pass_rate": 0.8576, "provider_group": "provider_y", "quality": 0.8685, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.972, "id": "worker_c", "p95_latency_ms": 80469, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 19.369, "max_latency_ms": 170609, "minimum_quality": 0.7407}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001116", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.529, "estimated_latency_ms": 105660, "estimated_pass_rate": 0.9149, "estimated_quality": 0.9482, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 19.369, "max_latency_ms": 170609, "policy": "balanced", "reason_codes": ["high_consequence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9149, "quality": 0.9482, "quality_target_met": true, "utility": 76.405899}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":19.369,\"max_latency_ms\":170609,\"minimum_quality\":0.7407},\"task\":{\"ambiguity\":0.475,\"category\":\"review\",\"complexity\":0.541,\"consequence\":0.8948,\"evidence_need\":0.556,\"family\":\"policy_compliance\",\"input_tokens\":1840,\"reversibility\":0.6254,\"risk\":0.7114,\"summary\":\"Identify evidence required before approving webhook delivery for production.\"},\"workers\":[{\"credits_per_call\":0.968,\"id\":\"worker_a\",\"p95_latency_ms\":26993,\"pass_rate\":0.5307,\"provider_group\":\"provider_y\",\"quality\":0.7133,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":4.219,\"id\":\"worker_b\",\"p95_latency_ms\":53675,\"pass_rate\":0.8591,\"provider_group\":\"provider_x\",\"quality\":0.9119,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.31,\"id\":\"worker_c\",\"p95_latency_ms\":51985,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.529,\"estimated_latency_ms\":105660,\"estimated_pass_rate\":0.9149,\"estimated_quality\":0.9482,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":19.369,\"max_latency_ms\":170609,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.475, "category": "review", "complexity": 0.541, "consequence": 0.8948, "evidence_need": 0.556, "family": "policy_compliance", "input_tokens": 1840, "request_id": "synthetic-policy_compliance-001116", "reversibility": 0.6254, "risk": 0.7114, "summary": "Identify evidence required before approving webhook delivery for production."}, "workers": [{"credits_per_call": 0.968, "id": "worker_a", "p95_latency_ms": 26993, "pass_rate": 0.5307, "provider_group": "provider_y", "quality": 0.7133, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 4.219, "id": "worker_b", "p95_latency_ms": 53675, "pass_rate": 0.8591, "provider_group": "provider_x", "quality": 0.9119, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.31, "id": "worker_c", "p95_latency_ms": 51985, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.536, "max_latency_ms": 175742, "minimum_quality": 0.8097}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000015", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.771, "estimated_latency_ms": 162495, "estimated_pass_rate": 0.7488, "estimated_quality": 0.9283, "finalizer": "worker_a", "independent_verification": true, "max_calls": 3, "max_credits": 18.536, "max_latency_ms": 175742, "policy": "strict", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7488, "quality": 0.9283, "quality_target_met": true, "utility": 72.03518}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.536,\"max_latency_ms\":175742,\"minimum_quality\":0.8097},\"task\":{\"ambiguity\":0.6902,\"category\":\"review\",\"complexity\":0.5785,\"consequence\":0.5777,\"evidence_need\":0.585,\"family\":\"policy_compliance\",\"input_tokens\":5080,\"reversibility\":0.6091,\"risk\":0.5262,\"summary\":\"Audit a workflow for deployment rollback against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.272,\"id\":\"worker_a\",\"p95_latency_ms\":69671,\"pass_rate\":0.9342,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.411,\"id\":\"worker_b\",\"p95_latency_ms\":69590,\"pass_rate\":0.7638,\"provider_group\":\"provider_z\",\"quality\":0.9212,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.088,\"id\":\"worker_c\",\"p95_latency_ms\":23234,\"pass_rate\":0.5735,\"provider_group\":\"provider_x\",\"quality\":0.7967,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.771,\"estimated_latency_ms\":162495,\"estimated_pass_rate\":0.7488,\"estimated_quality\":0.9283,\"finalizer\":\"worker_a\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.536,\"max_latency_ms\":175742,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6902, "category": "review", "complexity": 0.5785, "consequence": 0.5777, "evidence_need": 0.585, "family": "policy_compliance", "input_tokens": 5080, "request_id": "synthetic-policy_compliance-000015", "reversibility": 0.6091, "risk": 0.5262, "summary": "Audit a workflow for deployment rollback against explicit access requirements."}, "workers": [{"credits_per_call": 11.272, "id": "worker_a", "p95_latency_ms": 69671, "pass_rate": 0.9342, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.411, "id": "worker_b", "p95_latency_ms": 69590, "pass_rate": 0.7638, "provider_group": "provider_z", "quality": 0.9212, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.088, "id": "worker_c", "p95_latency_ms": 23234, "pass_rate": 0.5735, "provider_group": "provider_x", "quality": 0.7967, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.88, "max_latency_ms": 192288, "minimum_quality": 0.8355}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000516", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 15.693, "estimated_latency_ms": 152232, "estimated_pass_rate": 0.7651, "estimated_quality": 0.8743, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 16.88, "max_latency_ms": 192288, "policy": "strict", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7651, "quality": 0.8743, "quality_target_met": true, "utility": 73.926339}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.88,\"max_latency_ms\":192288,\"minimum_quality\":0.8355},\"task\":{\"ambiguity\":0.4978,\"category\":\"review\",\"complexity\":0.7176,\"consequence\":0.7766,\"evidence_need\":0.6627,\"family\":\"policy_compliance\",\"input_tokens\":1280,\"reversibility\":0.3886,\"risk\":0.7455,\"summary\":\"Identify evidence required before approving OAuth callbacks for production.\"},\"workers\":[{\"credits_per_call\":1.034,\"id\":\"worker_a\",\"p95_latency_ms\":18916,\"pass_rate\":0.5627,\"provider_group\":\"provider_z\",\"quality\":0.7312,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.784,\"id\":\"worker_b\",\"p95_latency_ms\":55242,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9171,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.875,\"id\":\"worker_c\",\"p95_latency_ms\":78074,\"pass_rate\":0.7413,\"provider_group\":\"provider_x\",\"quality\":0.8619,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":15.693,\"estimated_latency_ms\":152232,\"estimated_pass_rate\":0.7651,\"estimated_quality\":0.8743,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":16.88,\"max_latency_ms\":192288,\"policy\":\"strict\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4978, "category": "review", "complexity": 0.7176, "consequence": 0.7766, "evidence_need": 0.6627, "family": "policy_compliance", "input_tokens": 1280, "request_id": "synthetic-policy_compliance-000516", "reversibility": 0.3886, "risk": 0.7455, "summary": "Identify evidence required before approving OAuth callbacks for production."}, "workers": [{"credits_per_call": 1.034, "id": "worker_a", "p95_latency_ms": 18916, "pass_rate": 0.5627, "provider_group": "provider_z", "quality": 0.7312, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.784, "id": "worker_b", "p95_latency_ms": 55242, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9171, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.875, "id": "worker_c", "p95_latency_ms": 78074, "pass_rate": 0.7413, "provider_group": "provider_x", "quality": 0.8619, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 21.193, "max_latency_ms": 217520, "minimum_quality": 0.838}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000087", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.757, "estimated_latency_ms": 188027, "estimated_pass_rate": 0.7405, "estimated_quality": 0.9284, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 21.193, "max_latency_ms": 217520, "policy": "strict", "reason_codes": ["independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7405, "quality": 0.9284, "quality_target_met": true, "utility": 73.715752}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":21.193,\"max_latency_ms\":217520,\"minimum_quality\":0.838},\"task\":{\"ambiguity\":0.38,\"category\":\"review\",\"complexity\":0.5528,\"consequence\":0.5965,\"evidence_need\":0.886,\"family\":\"policy_compliance\",\"input_tokens\":1000,\"reversibility\":0.5403,\"risk\":0.6706,\"summary\":\"Audit a workflow for provider failover against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":4.168,\"id\":\"worker_a\",\"p95_latency_ms\":71242,\"pass_rate\":0.7388,\"provider_group\":\"provider_x\",\"quality\":0.8268,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":11.772,\"id\":\"worker_b\",\"p95_latency_ms\":87767,\"pass_rate\":0.9879,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.817,\"id\":\"worker_c\",\"p95_latency_ms\":29018,\"pass_rate\":0.5055,\"provider_group\":\"provider_z\",\"quality\":0.7713,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.757,\"estimated_latency_ms\":188027,\"estimated_pass_rate\":0.7405,\"estimated_quality\":0.9284,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":21.193,\"max_latency_ms\":217520,\"policy\":\"strict\",\"reason_codes\":[\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.38, "category": "review", "complexity": 0.5528, "consequence": 0.5965, "evidence_need": 0.886, "family": "policy_compliance", "input_tokens": 1000, "request_id": "synthetic-policy_compliance-000087", "reversibility": 0.5403, "risk": 0.6706, "summary": "Audit a workflow for provider failover against explicit access requirements."}, "workers": [{"credits_per_call": 4.168, "id": "worker_a", "p95_latency_ms": 71242, "pass_rate": 0.7388, "provider_group": "provider_x", "quality": 0.8268, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 11.772, "id": "worker_b", "p95_latency_ms": 87767, "pass_rate": 0.9879, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.817, "id": "worker_c", "p95_latency_ms": 29018, "pass_rate": 0.5055, "provider_group": "provider_z", "quality": 0.7713, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 12.372, "max_latency_ms": 197954, "minimum_quality": 0.8602}, "family": "migration_planning", "id": "synthetic-migration_planning-000499", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.316, "estimated_latency_ms": 88444, "estimated_pass_rate": 0.6394, "estimated_quality": 0.7852, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 12.372, "max_latency_ms": 197954, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6394, "quality": 0.7852, "quality_target_met": false, "utility": 59.397228}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":12.372,\"max_latency_ms\":197954,\"minimum_quality\":0.8602},\"task\":{\"ambiguity\":0.5382,\"category\":\"coding\",\"complexity\":0.6454,\"consequence\":0.6622,\"evidence_need\":0.7444,\"family\":\"migration_planning\",\"input_tokens\":7120,\"reversibility\":0.5601,\"risk\":0.7268,\"summary\":\"Design compatibility testing for a provider failover schema migration.\"},\"workers\":[{\"credits_per_call\":0.986,\"id\":\"worker_a\",\"p95_latency_ms\":30832,\"pass_rate\":0.5591,\"provider_group\":\"provider_x\",\"quality\":0.6941,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.33,\"id\":\"worker_b\",\"p95_latency_ms\":57612,\"pass_rate\":0.8571,\"provider_group\":\"provider_z\",\"quality\":0.7948,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":15.27,\"id\":\"worker_c\",\"p95_latency_ms\":56037,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9291,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.316,\"estimated_latency_ms\":88444,\"estimated_pass_rate\":0.6394,\"estimated_quality\":0.7852,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":12.372,\"max_latency_ms\":197954,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5382, "category": "coding", "complexity": 0.6454, "consequence": 0.6622, "evidence_need": 0.7444, "family": "migration_planning", "input_tokens": 7120, "request_id": "synthetic-migration_planning-000499", "reversibility": 0.5601, "risk": 0.7268, "summary": "Design compatibility testing for a provider failover schema migration."}, "workers": [{"credits_per_call": 0.986, "id": "worker_a", "p95_latency_ms": 30832, "pass_rate": 0.5591, "provider_group": "provider_x", "quality": 0.6941, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.33, "id": "worker_b", "p95_latency_ms": 57612, "pass_rate": 0.8571, "provider_group": "provider_z", "quality": 0.7948, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 15.27, "id": "worker_c", "p95_latency_ms": 56037, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9291, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 21.67, "max_latency_ms": 151347, "minimum_quality": 0.6941}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001059", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 17.978, "estimated_latency_ms": 132200, "estimated_pass_rate": 0.8597, "estimated_quality": 0.9124, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 21.67, "max_latency_ms": 151347, "policy": "balanced", "reason_codes": ["complex_task", "credit_constrained", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8597, "quality": 0.9124, "quality_target_met": true, "utility": 71.451464}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":21.67,\"max_latency_ms\":151347,\"minimum_quality\":0.6941},\"task\":{\"ambiguity\":0.7887,\"category\":\"review\",\"complexity\":0.7874,\"consequence\":0.7637,\"evidence_need\":0.7677,\"family\":\"policy_compliance\",\"input_tokens\":4280,\"reversibility\":0.4882,\"risk\":0.7735,\"summary\":\"Audit a workflow for schema evolution against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":3.565,\"id\":\"worker_a\",\"p95_latency_ms\":73259,\"pass_rate\":0.8276,\"provider_group\":\"provider_x\",\"quality\":0.8723,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.193,\"id\":\"worker_b\",\"p95_latency_ms\":27996,\"pass_rate\":0.5885,\"provider_group\":\"provider_z\",\"quality\":0.6814,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":14.413,\"id\":\"worker_c\",\"p95_latency_ms\":58941,\"pass_rate\":0.9588,\"provider_group\":\"provider_y\",\"quality\":0.982,\"strengths\":[\"drafting\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":17.978,\"estimated_latency_ms\":132200,\"estimated_pass_rate\":0.8597,\"estimated_quality\":0.9124,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":21.67,\"max_latency_ms\":151347,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7887, "category": "review", "complexity": 0.7874, "consequence": 0.7637, "evidence_need": 0.7677, "family": "policy_compliance", "input_tokens": 4280, "request_id": "synthetic-policy_compliance-001059", "reversibility": 0.4882, "risk": 0.7735, "summary": "Audit a workflow for schema evolution against explicit access requirements."}, "workers": [{"credits_per_call": 3.565, "id": "worker_a", "p95_latency_ms": 73259, "pass_rate": 0.8276, "provider_group": "provider_x", "quality": 0.8723, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.193, "id": "worker_b", "p95_latency_ms": 27996, "pass_rate": 0.5885, "provider_group": "provider_z", "quality": 0.6814, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 14.413, "id": "worker_c", "p95_latency_ms": 58941, "pass_rate": 0.9588, "provider_group": "provider_y", "quality": 0.982, "strengths": ["drafting", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.537, "max_latency_ms": 195076, "minimum_quality": 0.7288}, "family": "migration_planning", "id": "synthetic-migration_planning-000310", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.079, "estimated_latency_ms": 87720, "estimated_pass_rate": 0.7213, "estimated_quality": 0.8082, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.537, "max_latency_ms": 195076, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7213, "quality": 0.8082, "quality_target_met": true, "utility": 60.065713}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.537,\"max_latency_ms\":195076,\"minimum_quality\":0.7288},\"task\":{\"ambiguity\":0.371,\"category\":\"coding\",\"complexity\":0.8533,\"consequence\":0.7438,\"evidence_need\":0.6782,\"family\":\"migration_planning\",\"input_tokens\":1880,\"reversibility\":0.3639,\"risk\":0.4583,\"summary\":\"Plan a zero-downtime migration of artifact storage with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":3.972,\"id\":\"worker_a\",\"p95_latency_ms\":70088,\"pass_rate\":0.8208,\"provider_group\":\"provider_x\",\"quality\":0.8896,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.107,\"id\":\"worker_b\",\"p95_latency_ms\":17632,\"pass_rate\":0.5189,\"provider_group\":\"provider_z\",\"quality\":0.7348,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.781,\"id\":\"worker_c\",\"p95_latency_ms\":62929,\"pass_rate\":0.913,\"provider_group\":\"provider_y\",\"quality\":0.9507,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.079,\"estimated_latency_ms\":87720,\"estimated_pass_rate\":0.7213,\"estimated_quality\":0.8082,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.537,\"max_latency_ms\":195076,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.371, "category": "coding", "complexity": 0.8533, "consequence": 0.7438, "evidence_need": 0.6782, "family": "migration_planning", "input_tokens": 1880, "request_id": "synthetic-migration_planning-000310", "reversibility": 0.3639, "risk": 0.4583, "summary": "Plan a zero-downtime migration of artifact storage with rollback checkpoints."}, "workers": [{"credits_per_call": 3.972, "id": "worker_a", "p95_latency_ms": 70088, "pass_rate": 0.8208, "provider_group": "provider_x", "quality": 0.8896, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.107, "id": "worker_b", "p95_latency_ms": 17632, "pass_rate": 0.5189, "provider_group": "provider_z", "quality": 0.7348, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.781, "id": "worker_c", "p95_latency_ms": 62929, "pass_rate": 0.913, "provider_group": "provider_y", "quality": 0.9507, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 7.23, "max_latency_ms": 99542, "minimum_quality": 0.7629}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001140", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.057, "estimated_latency_ms": 88976, "estimated_pass_rate": 0.6644, "estimated_quality": 0.8361, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 7.23, "max_latency_ms": 99542, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6644, "quality": 0.8361, "quality_target_met": true, "utility": 59.125479}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":7.23,\"max_latency_ms\":99542,\"minimum_quality\":0.7629},\"task\":{\"ambiguity\":0.4326,\"category\":\"review\",\"complexity\":0.8122,\"consequence\":0.5285,\"evidence_need\":0.8811,\"family\":\"policy_compliance\",\"input_tokens\":5480,\"reversibility\":0.5508,\"risk\":0.6591,\"summary\":\"Review whether API-key rotation satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":2.962,\"id\":\"worker_a\",\"p95_latency_ms\":62078,\"pass_rate\":0.7324,\"provider_group\":\"provider_x\",\"quality\":0.9091,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.095,\"id\":\"worker_b\",\"p95_latency_ms\":26898,\"pass_rate\":0.5978,\"provider_group\":\"provider_z\",\"quality\":0.7299,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":11.653,\"id\":\"worker_c\",\"p95_latency_ms\":65913,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9327,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.057,\"estimated_latency_ms\":88976,\"estimated_pass_rate\":0.6644,\"estimated_quality\":0.8361,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":7.23,\"max_latency_ms\":99542,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4326, "category": "review", "complexity": 0.8122, "consequence": 0.5285, "evidence_need": 0.8811, "family": "policy_compliance", "input_tokens": 5480, "request_id": "synthetic-policy_compliance-001140", "reversibility": 0.5508, "risk": 0.6591, "summary": "Review whether API-key rotation satisfies a documented retention policy."}, "workers": [{"credits_per_call": 2.962, "id": "worker_a", "p95_latency_ms": 62078, "pass_rate": 0.7324, "provider_group": "provider_x", "quality": 0.9091, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.095, "id": "worker_b", "p95_latency_ms": 26898, "pass_rate": 0.5978, "provider_group": "provider_z", "quality": 0.7299, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 11.653, "id": "worker_c", "p95_latency_ms": 65913, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9327, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 1, "max_credits": 3.101, "max_latency_ms": 32118, "minimum_quality": 0.5681}, "family": "migration_planning", "id": "synthetic-migration_planning-000493", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.81, "estimated_latency_ms": 20846, "estimated_pass_rate": 0.3597, "estimated_quality": 0.6105, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.101, "max_latency_ms": 32118, "policy": "lean", "reason_codes": ["independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3597, "quality": 0.6105, "quality_target_met": true, "utility": 41.374959}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":3.101,\"max_latency_ms\":32118,\"minimum_quality\":0.5681},\"task\":{\"ambiguity\":0.6069,\"category\":\"coding\",\"complexity\":0.5877,\"consequence\":0.5283,\"evidence_need\":0.8405,\"family\":\"migration_planning\",\"input_tokens\":7560,\"reversibility\":0.5334,\"risk\":0.487,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":10.451,\"id\":\"worker_a\",\"p95_latency_ms\":74475,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.909,\"id\":\"worker_b\",\"p95_latency_ms\":53141,\"pass_rate\":0.826,\"provider_group\":\"provider_y\",\"quality\":0.8186,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.81,\"id\":\"worker_c\",\"p95_latency_ms\":20846,\"pass_rate\":0.5117,\"provider_group\":\"provider_z\",\"quality\":0.7126,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.81,\"estimated_latency_ms\":20846,\"estimated_pass_rate\":0.3597,\"estimated_quality\":0.6105,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.101,\"max_latency_ms\":32118,\"policy\":\"lean\",\"reason_codes\":[\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6069, "category": "coding", "complexity": 0.5877, "consequence": 0.5283, "evidence_need": 0.8405, "family": "migration_planning", "input_tokens": 7560, "request_id": "synthetic-migration_planning-000493", "reversibility": 0.5334, "risk": 0.487, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 10.451, "id": "worker_a", "p95_latency_ms": 74475, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.909, "id": "worker_b", "p95_latency_ms": 53141, "pass_rate": 0.826, "provider_group": "provider_y", "quality": 0.8186, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.81, "id": "worker_c", "p95_latency_ms": 20846, "pass_rate": 0.5117, "provider_group": "provider_z", "quality": 0.7126, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 16.332, "max_latency_ms": 125756, "minimum_quality": 0.8639}, "family": "migration_planning", "id": "synthetic-migration_planning-001066", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.127, "estimated_latency_ms": 112821, "estimated_pass_rate": 0.9053, "estimated_quality": 0.9403, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.332, "max_latency_ms": 125756, "policy": "balanced", "reason_codes": ["credit_constrained", "independent_verification", "latency_constrained"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9053, "quality": 0.9403, "quality_target_met": true, "utility": 70.762652}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":16.332,\"max_latency_ms\":125756,\"minimum_quality\":0.8639},\"task\":{\"ambiguity\":0.3042,\"category\":\"coding\",\"complexity\":0.5925,\"consequence\":0.5028,\"evidence_need\":0.6356,\"family\":\"migration_planning\",\"input_tokens\":7640,\"reversibility\":0.2556,\"risk\":0.6608,\"summary\":\"Review a staged migration for background jobs and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":1.098,\"id\":\"worker_a\",\"p95_latency_ms\":20587,\"pass_rate\":0.5005,\"provider_group\":\"provider_z\",\"quality\":0.7508,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.091,\"id\":\"worker_b\",\"p95_latency_ms\":51068,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.036,\"id\":\"worker_c\",\"p95_latency_ms\":61753,\"pass_rate\":0.7973,\"provider_group\":\"provider_x\",\"quality\":0.804,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.127,\"estimated_latency_ms\":112821,\"estimated_pass_rate\":0.9053,\"estimated_quality\":0.9403,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.332,\"max_latency_ms\":125756,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3042, "category": "coding", "complexity": 0.5925, "consequence": 0.5028, "evidence_need": 0.6356, "family": "migration_planning", "input_tokens": 7640, "request_id": "synthetic-migration_planning-001066", "reversibility": 0.2556, "risk": 0.6608, "summary": "Review a staged migration for background jobs and identify irreversible steps."}, "workers": [{"credits_per_call": 1.098, "id": "worker_a", "p95_latency_ms": 20587, "pass_rate": 0.5005, "provider_group": "provider_z", "quality": 0.7508, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.091, "id": "worker_b", "p95_latency_ms": 51068, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.036, "id": "worker_c", "p95_latency_ms": 61753, "pass_rate": 0.7973, "provider_group": "provider_x", "quality": 0.804, "strengths": ["coding", "reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.578, "max_latency_ms": 37804, "minimum_quality": 0.7462}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000822", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.88, "estimated_latency_ms": 29574, "estimated_pass_rate": 0.3676, "estimated_quality": 0.5716, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.578, "max_latency_ms": 37804, "policy": "lean", "reason_codes": ["best_feasible_below_target", "independent_evidence"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3676, "quality": 0.5716, "quality_target_met": false, "utility": 31.318826}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.578,\"max_latency_ms\":37804,\"minimum_quality\":0.7462},\"task\":{\"ambiguity\":0.3536,\"category\":\"review\",\"complexity\":0.7071,\"consequence\":0.59,\"evidence_need\":0.7823,\"family\":\"policy_compliance\",\"input_tokens\":3240,\"reversibility\":0.3046,\"risk\":0.4541,\"summary\":\"Review whether private credentials satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":13.932,\"id\":\"worker_a\",\"p95_latency_ms\":76649,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9606,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.711,\"id\":\"worker_b\",\"p95_latency_ms\":49183,\"pass_rate\":0.7762,\"provider_group\":\"provider_x\",\"quality\":0.8441,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.88,\"id\":\"worker_c\",\"p95_latency_ms\":29574,\"pass_rate\":0.5113,\"provider_group\":\"provider_z\",\"quality\":0.6759,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.88,\"estimated_latency_ms\":29574,\"estimated_pass_rate\":0.3676,\"estimated_quality\":0.5716,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.578,\"max_latency_ms\":37804,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.3536, "category": "review", "complexity": 0.7071, "consequence": 0.59, "evidence_need": 0.7823, "family": "policy_compliance", "input_tokens": 3240, "request_id": "synthetic-policy_compliance-000822", "reversibility": 0.3046, "risk": 0.4541, "summary": "Review whether private credentials satisfies a documented retention policy."}, "workers": [{"credits_per_call": 13.932, "id": "worker_a", "p95_latency_ms": 76649, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9606, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.711, "id": "worker_b", "p95_latency_ms": 49183, "pass_rate": 0.7762, "provider_group": "provider_x", "quality": 0.8441, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.88, "id": "worker_c", "p95_latency_ms": 29574, "pass_rate": 0.5113, "provider_group": "provider_z", "quality": 0.6759, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 18.211, "max_latency_ms": 237717, "minimum_quality": 0.8617}, "family": "migration_planning", "id": "synthetic-migration_planning-000031", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 14.824, "estimated_latency_ms": 176709, "estimated_pass_rate": 0.7677, "estimated_quality": 0.8915, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 18.211, "max_latency_ms": 237717, "policy": "strict", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7677, "quality": 0.8915, "quality_target_met": true, "utility": 78.445817}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":18.211,\"max_latency_ms\":237717,\"minimum_quality\":0.8617},\"task\":{\"ambiguity\":0.516,\"category\":\"coding\",\"complexity\":0.9123,\"consequence\":0.788,\"evidence_need\":0.6792,\"family\":\"migration_planning\",\"input_tokens\":3760,\"reversibility\":0.6052,\"risk\":0.7228,\"summary\":\"Review a staged migration for API-key rotation and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.38,\"id\":\"worker_a\",\"p95_latency_ms\":82636,\"pass_rate\":0.8187,\"provider_group\":\"provider_z\",\"quality\":0.7961,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.325,\"id\":\"worker_b\",\"p95_latency_ms\":68718,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9588,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.119,\"id\":\"worker_c\",\"p95_latency_ms\":25355,\"pass_rate\":0.5884,\"provider_group\":\"provider_y\",\"quality\":0.7295,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":14.824,\"estimated_latency_ms\":176709,\"estimated_pass_rate\":0.7677,\"estimated_quality\":0.8915,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":18.211,\"max_latency_ms\":237717,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.516, "category": "coding", "complexity": 0.9123, "consequence": 0.788, "evidence_need": 0.6792, "family": "migration_planning", "input_tokens": 3760, "request_id": "synthetic-migration_planning-000031", "reversibility": 0.6052, "risk": 0.7228, "summary": "Review a staged migration for API-key rotation and identify irreversible steps."}, "workers": [{"credits_per_call": 3.38, "id": "worker_a", "p95_latency_ms": 82636, "pass_rate": 0.8187, "provider_group": "provider_z", "quality": 0.7961, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.325, "id": "worker_b", "p95_latency_ms": 68718, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9588, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.119, "id": "worker_c", "p95_latency_ms": 25355, "pass_rate": 0.5884, "provider_group": "provider_y", "quality": 0.7295, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 8.986, "max_latency_ms": 161909, "minimum_quality": 0.6918}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000864", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.157, "estimated_latency_ms": 101370, "estimated_pass_rate": 0.5475, "estimated_quality": 0.7643, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.986, "max_latency_ms": 161909, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5475, "quality": 0.7643, "quality_target_met": true, "utility": 56.45348}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":8.986,\"max_latency_ms\":161909,\"minimum_quality\":0.6918},\"task\":{\"ambiguity\":0.5126,\"category\":\"review\",\"complexity\":0.7685,\"consequence\":0.7653,\"evidence_need\":0.8544,\"family\":\"policy_compliance\",\"input_tokens\":1560,\"reversibility\":0.6153,\"risk\":0.4696,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":12.793,\"id\":\"worker_a\",\"p95_latency_ms\":56065,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9375,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.013,\"id\":\"worker_b\",\"p95_latency_ms\":72975,\"pass_rate\":0.7486,\"provider_group\":\"provider_y\",\"quality\":0.8243,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.144,\"id\":\"worker_c\",\"p95_latency_ms\":28395,\"pass_rate\":0.4992,\"provider_group\":\"provider_z\",\"quality\":0.7021,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.157,\"estimated_latency_ms\":101370,\"estimated_pass_rate\":0.5475,\"estimated_quality\":0.7643,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.986,\"max_latency_ms\":161909,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5126, "category": "review", "complexity": 0.7685, "consequence": 0.7653, "evidence_need": 0.8544, "family": "policy_compliance", "input_tokens": 1560, "request_id": "synthetic-policy_compliance-000864", "reversibility": 0.6153, "risk": 0.4696, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 12.793, "id": "worker_a", "p95_latency_ms": 56065, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9375, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.013, "id": "worker_b", "p95_latency_ms": 72975, "pass_rate": 0.7486, "provider_group": "provider_y", "quality": 0.8243, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.144, "id": "worker_c", "p95_latency_ms": 28395, "pass_rate": 0.4992, "provider_group": "provider_z", "quality": 0.7021, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.087, "max_latency_ms": 224456, "minimum_quality": 0.6842}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001086", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.05, "estimated_latency_ms": 98813, "estimated_pass_rate": 0.617, "estimated_quality": 0.8374, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.087, "max_latency_ms": 224456, "policy": "balanced", "reason_codes": ["high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 5, "pass_rate": 0.617, "quality": 0.8374, "quality_target_met": true, "utility": 66.412755}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.087,\"max_latency_ms\":224456,\"minimum_quality\":0.6842},\"task\":{\"ambiguity\":0.5484,\"category\":\"review\",\"complexity\":0.6027,\"consequence\":0.8439,\"evidence_need\":0.7656,\"family\":\"policy_compliance\",\"input_tokens\":3880,\"reversibility\":0.4728,\"risk\":0.7697,\"summary\":\"Review whether prompt caching satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":0.834,\"id\":\"worker_a\",\"p95_latency_ms\":18457,\"pass_rate\":0.4999,\"provider_group\":\"provider_y\",\"quality\":0.7811,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":10.39,\"id\":\"worker_b\",\"p95_latency_ms\":61782,\"pass_rate\":0.9348,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.216,\"id\":\"worker_c\",\"p95_latency_ms\":80356,\"pass_rate\":0.7614,\"provider_group\":\"provider_x\",\"quality\":0.9103,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.05,\"estimated_latency_ms\":98813,\"estimated_pass_rate\":0.617,\"estimated_quality\":0.8374,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.087,\"max_latency_ms\":224456,\"policy\":\"balanced\",\"reason_codes\":[\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.5484, "category": "review", "complexity": 0.6027, "consequence": 0.8439, "evidence_need": 0.7656, "family": "policy_compliance", "input_tokens": 3880, "request_id": "synthetic-policy_compliance-001086", "reversibility": 0.4728, "risk": 0.7697, "summary": "Review whether prompt caching satisfies a documented retention policy."}, "workers": [{"credits_per_call": 0.834, "id": "worker_a", "p95_latency_ms": 18457, "pass_rate": 0.4999, "provider_group": "provider_y", "quality": 0.7811, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 10.39, "id": "worker_b", "p95_latency_ms": 61782, "pass_rate": 0.9348, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.216, "id": "worker_c", "p95_latency_ms": 80356, "pass_rate": 0.7614, "provider_group": "provider_x", "quality": 0.9103, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 17.783, "max_latency_ms": 237554, "minimum_quality": 0.7851}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000605", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 16.51, "estimated_latency_ms": 172845, "estimated_pass_rate": 0.8364, "estimated_quality": 0.9267, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 17.783, "max_latency_ms": 237554, "policy": "strict", "reason_codes": ["complex_task", "credit_constrained", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.8364, "quality": 0.9267, "quality_target_met": true, "utility": 70.264856}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":17.783,\"max_latency_ms\":237554,\"minimum_quality\":0.7851},\"task\":{\"ambiguity\":0.4147,\"category\":\"analysis\",\"complexity\":0.809,\"consequence\":0.3396,\"evidence_need\":0.5562,\"family\":\"performance_tuning\",\"input_tokens\":6880,\"reversibility\":0.5931,\"risk\":0.3349,\"summary\":\"Choose between throughput and tail-latency changes for background jobs.\"},\"workers\":[{\"credits_per_call\":0.878,\"id\":\"worker_a\",\"p95_latency_ms\":32364,\"pass_rate\":0.4493,\"provider_group\":\"provider_z\",\"quality\":0.714,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":12.39,\"id\":\"worker_b\",\"p95_latency_ms\":80629,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.242,\"id\":\"worker_c\",\"p95_latency_ms\":59852,\"pass_rate\":0.85,\"provider_group\":\"provider_y\",\"quality\":0.895,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":16.51,\"estimated_latency_ms\":172845,\"estimated_pass_rate\":0.8364,\"estimated_quality\":0.9267,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":17.783,\"max_latency_ms\":237554,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"credit_constrained\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.4147, "category": "analysis", "complexity": 0.809, "consequence": 0.3396, "evidence_need": 0.5562, "family": "performance_tuning", "input_tokens": 6880, "request_id": "synthetic-performance_tuning-000605", "reversibility": 0.5931, "risk": 0.3349, "summary": "Choose between throughput and tail-latency changes for background jobs."}, "workers": [{"credits_per_call": 0.878, "id": "worker_a", "p95_latency_ms": 32364, "pass_rate": 0.4493, "provider_group": "provider_z", "quality": 0.714, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 12.39, "id": "worker_b", "p95_latency_ms": 80629, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.242, "id": "worker_c", "p95_latency_ms": 59852, "pass_rate": 0.85, "provider_group": "provider_y", "quality": 0.895, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.038, "max_latency_ms": 50734, "minimum_quality": 0.7595}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000708", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.202, "estimated_latency_ms": 27540, "estimated_pass_rate": 0.4088, "estimated_quality": 0.5853, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.038, "max_latency_ms": 50734, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.4088, "quality": 0.5853, "quality_target_met": false, "utility": 31.803302}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.038,\"max_latency_ms\":50734,\"minimum_quality\":0.7595},\"task\":{\"ambiguity\":0.6549,\"category\":\"review\",\"complexity\":0.8288,\"consequence\":0.6361,\"evidence_need\":0.6949,\"family\":\"policy_compliance\",\"input_tokens\":4520,\"reversibility\":0.4833,\"risk\":0.7574,\"summary\":\"Review whether background jobs satisfies a documented retention policy.\"},\"workers\":[{\"credits_per_call\":10.077,\"id\":\"worker_a\",\"p95_latency_ms\":79390,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9203,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.202,\"id\":\"worker_b\",\"p95_latency_ms\":27540,\"pass_rate\":0.6136,\"provider_group\":\"provider_z\",\"quality\":0.7537,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.958,\"id\":\"worker_c\",\"p95_latency_ms\":77759,\"pass_rate\":0.7604,\"provider_group\":\"provider_x\",\"quality\":0.8963,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.202,\"estimated_latency_ms\":27540,\"estimated_pass_rate\":0.4088,\"estimated_quality\":0.5853,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.038,\"max_latency_ms\":50734,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.6549, "category": "review", "complexity": 0.8288, "consequence": 0.6361, "evidence_need": 0.6949, "family": "policy_compliance", "input_tokens": 4520, "request_id": "synthetic-policy_compliance-000708", "reversibility": 0.4833, "risk": 0.7574, "summary": "Review whether background jobs satisfies a documented retention policy."}, "workers": [{"credits_per_call": 10.077, "id": "worker_a", "p95_latency_ms": 79390, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9203, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.202, "id": "worker_b", "p95_latency_ms": 27540, "pass_rate": 0.6136, "provider_group": "provider_z", "quality": 0.7537, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.958, "id": "worker_c", "p95_latency_ms": 77759, "pass_rate": 0.7604, "provider_group": "provider_x", "quality": 0.8963, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 11.408, "max_latency_ms": 221679, "minimum_quality": 0.889}, "family": "migration_planning", "id": "synthetic-migration_planning-000349", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.918, "estimated_latency_ms": 97657, "estimated_pass_rate": 0.5934, "estimated_quality": 0.7846, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 11.408, "max_latency_ms": 221679, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "high_consequence", "independent_evidence"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.5934, "quality": 0.7846, "quality_target_met": false, "utility": 54.498088}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":11.408,\"max_latency_ms\":221679,\"minimum_quality\":0.889},\"task\":{\"ambiguity\":0.3792,\"category\":\"coding\",\"complexity\":0.9091,\"consequence\":0.788,\"evidence_need\":0.7114,\"family\":\"migration_planning\",\"input_tokens\":6160,\"reversibility\":0.5706,\"risk\":0.3755,\"summary\":\"Review a staged migration for database failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.799,\"id\":\"worker_a\",\"p95_latency_ms\":66642,\"pass_rate\":0.7718,\"provider_group\":\"provider_z\",\"quality\":0.8383,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.119,\"id\":\"worker_b\",\"p95_latency_ms\":31015,\"pass_rate\":0.5693,\"provider_group\":\"provider_x\",\"quality\":0.7653,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.988,\"id\":\"worker_c\",\"p95_latency_ms\":49885,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9596,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.918,\"estimated_latency_ms\":97657,\"estimated_pass_rate\":0.5934,\"estimated_quality\":0.7846,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":11.408,\"max_latency_ms\":221679,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"high_consequence\",\"independent_evidence\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3792, "category": "coding", "complexity": 0.9091, "consequence": 0.788, "evidence_need": 0.7114, "family": "migration_planning", "input_tokens": 6160, "request_id": "synthetic-migration_planning-000349", "reversibility": 0.5706, "risk": 0.3755, "summary": "Review a staged migration for database failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.799, "id": "worker_a", "p95_latency_ms": 66642, "pass_rate": 0.7718, "provider_group": "provider_z", "quality": 0.8383, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.119, "id": "worker_b", "p95_latency_ms": 31015, "pass_rate": 0.5693, "provider_group": "provider_x", "quality": 0.7653, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.988, "id": "worker_c", "p95_latency_ms": 49885, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9596, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 3, "max_credits": 24.505, "max_latency_ms": 175636, "minimum_quality": 0.9506}, "family": "policy_compliance", "id": "synthetic-policy_compliance-001218", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.179, "estimated_latency_ms": 152509, "estimated_pass_rate": 0.7584, "estimated_quality": 0.8853, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 24.505, "max_latency_ms": 175636, "policy": "strict", "reason_codes": ["best_feasible_below_target", "independent_evidence", "independent_verification", "latency_constrained"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7584, "quality": 0.8853, "quality_target_met": false, "utility": 68.794612}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":24.505,\"max_latency_ms\":175636,\"minimum_quality\":0.9506},\"task\":{\"ambiguity\":0.702,\"category\":\"review\",\"complexity\":0.5475,\"consequence\":0.7005,\"evidence_need\":0.795,\"family\":\"policy_compliance\",\"input_tokens\":7160,\"reversibility\":0.5414,\"risk\":0.49,\"summary\":\"Identify evidence required before approving rate limiting for production.\"},\"workers\":[{\"credits_per_call\":4.007,\"id\":\"worker_a\",\"p95_latency_ms\":64093,\"pass_rate\":0.754,\"provider_group\":\"provider_z\",\"quality\":0.8621,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":14.065,\"id\":\"worker_b\",\"p95_latency_ms\":70494,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.9171,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.107,\"id\":\"worker_c\",\"p95_latency_ms\":17922,\"pass_rate\":0.5684,\"provider_group\":\"provider_y\",\"quality\":0.7315,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.179,\"estimated_latency_ms\":152509,\"estimated_pass_rate\":0.7584,\"estimated_quality\":0.8853,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":24.505,\"max_latency_ms\":175636,\"policy\":\"strict\",\"reason_codes\":[\"best_feasible_below_target\",\"independent_evidence\",\"independent_verification\",\"latency_constrained\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.702, "category": "review", "complexity": 0.5475, "consequence": 0.7005, "evidence_need": 0.795, "family": "policy_compliance", "input_tokens": 7160, "request_id": "synthetic-policy_compliance-001218", "reversibility": 0.5414, "risk": 0.49, "summary": "Identify evidence required before approving rate limiting for production."}, "workers": [{"credits_per_call": 4.007, "id": "worker_a", "p95_latency_ms": 64093, "pass_rate": 0.754, "provider_group": "provider_z", "quality": 0.8621, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 14.065, "id": "worker_b", "p95_latency_ms": 70494, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.9171, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.107, "id": "worker_c", "p95_latency_ms": 17922, "pass_rate": 0.5684, "provider_group": "provider_y", "quality": 0.7315, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 1, "max_credits": 2.548, "max_latency_ms": 39089, "minimum_quality": 0.7579}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000662", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.19, "estimated_latency_ms": 29584, "estimated_pass_rate": 0.3418, "estimated_quality": 0.4647, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.548, "max_latency_ms": 39089, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3418, "quality": 0.4647, "quality_target_met": false, "utility": 17.683403}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":1,\"max_credits\":2.548,\"max_latency_ms\":39089,\"minimum_quality\":0.7579},\"task\":{\"ambiguity\":0.3877,\"category\":\"analysis\",\"complexity\":0.888,\"consequence\":0.6922,\"evidence_need\":0.6068,\"family\":\"performance_tuning\",\"input_tokens\":5000,\"reversibility\":0.6725,\"risk\":0.5926,\"summary\":\"Review a proposed optimization for tenant-scoped caching for hidden regressions.\"},\"workers\":[{\"credits_per_call\":11.42,\"id\":\"worker_a\",\"p95_latency_ms\":53045,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.19,\"id\":\"worker_b\",\"p95_latency_ms\":29584,\"pass_rate\":0.5198,\"provider_group\":\"provider_y\",\"quality\":0.614,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":4.227,\"id\":\"worker_c\",\"p95_latency_ms\":79393,\"pass_rate\":0.9155,\"provider_group\":\"provider_z\",\"quality\":0.9615,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.19,\"estimated_latency_ms\":29584,\"estimated_pass_rate\":0.3418,\"estimated_quality\":0.4647,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.548,\"max_latency_ms\":39089,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3877, "category": "analysis", "complexity": 0.888, "consequence": 0.6922, "evidence_need": 0.6068, "family": "performance_tuning", "input_tokens": 5000, "request_id": "synthetic-performance_tuning-000662", "reversibility": 0.6725, "risk": 0.5926, "summary": "Review a proposed optimization for tenant-scoped caching for hidden regressions."}, "workers": [{"credits_per_call": 11.42, "id": "worker_a", "p95_latency_ms": 53045, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.19, "id": "worker_b", "p95_latency_ms": 29584, "pass_rate": 0.5198, "provider_group": "provider_y", "quality": 0.614, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 4.227, "id": "worker_c", "p95_latency_ms": 79393, "pass_rate": 0.9155, "provider_group": "provider_z", "quality": 0.9615, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 9.736, "max_latency_ms": 118927, "minimum_quality": 0.8425}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000003", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.655, "estimated_latency_ms": 72205, "estimated_pass_rate": 0.6148, "estimated_quality": 0.8484, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 9.736, "max_latency_ms": 118927, "policy": "balanced", "reason_codes": ["complex_task", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_c"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6148, "quality": 0.8484, "quality_target_met": true, "utility": 65.054643}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":9.736,\"max_latency_ms\":118927,\"minimum_quality\":0.8425},\"task\":{\"ambiguity\":0.7467,\"category\":\"review\",\"complexity\":0.4192,\"consequence\":0.5566,\"evidence_need\":0.9396,\"family\":\"policy_compliance\",\"input_tokens\":2920,\"reversibility\":0.6554,\"risk\":0.7962,\"summary\":\"Audit a workflow for prompt caching against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":15.249,\"id\":\"worker_a\",\"p95_latency_ms\":76227,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":2.875,\"id\":\"worker_b\",\"p95_latency_ms\":51288,\"pass_rate\":0.8183,\"provider_group\":\"provider_z\",\"quality\":0.8531,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.78,\"id\":\"worker_c\",\"p95_latency_ms\":20917,\"pass_rate\":0.4813,\"provider_group\":\"provider_y\",\"quality\":0.7474,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.655,\"estimated_latency_ms\":72205,\"estimated_pass_rate\":0.6148,\"estimated_quality\":0.8484,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":9.736,\"max_latency_ms\":118927,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.7467, "category": "review", "complexity": 0.4192, "consequence": 0.5566, "evidence_need": 0.9396, "family": "policy_compliance", "input_tokens": 2920, "request_id": "synthetic-policy_compliance-000003", "reversibility": 0.6554, "risk": 0.7962, "summary": "Audit a workflow for prompt caching against explicit access requirements."}, "workers": [{"credits_per_call": 15.249, "id": "worker_a", "p95_latency_ms": 76227, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 2.875, "id": "worker_b", "p95_latency_ms": 51288, "pass_rate": 0.8183, "provider_group": "provider_z", "quality": 0.8531, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.78, "id": "worker_c", "p95_latency_ms": 20917, "pass_rate": 0.4813, "provider_group": "provider_y", "quality": 0.7474, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 3, "max_credits": 23.708, "max_latency_ms": 201923, "minimum_quality": 0.8963}, "family": "performance_tuning", "id": "synthetic-performance_tuning-001133", "label": {"access": {"finalizer": ["writer", "reviewer"], "reviewer": ["writer"], "writer": []}, "estimated_credits": 19.008, "estimated_latency_ms": 158695, "estimated_pass_rate": 0.7612, "estimated_quality": 0.9201, "finalizer": "worker_b", "independent_verification": true, "max_calls": 3, "max_credits": 23.708, "max_latency_ms": 201923, "policy": "strict", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 15, "pass_rate": 0.7612, "quality": 0.9201, "quality_target_met": true, "utility": 72.283428}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":23.708,\"max_latency_ms\":201923,\"minimum_quality\":0.8963},\"task\":{\"ambiguity\":0.4756,\"category\":\"analysis\",\"complexity\":0.7655,\"consequence\":0.6183,\"evidence_need\":0.5121,\"family\":\"performance_tuning\",\"input_tokens\":4240,\"reversibility\":0.7876,\"risk\":0.4513,\"summary\":\"Diagnose latency variance in artifact storage and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":1.046,\"id\":\"worker_a\",\"p95_latency_ms\":17816,\"pass_rate\":0.4763,\"provider_group\":\"provider_y\",\"quality\":0.6907,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]},{\"credits_per_call\":14.771,\"id\":\"worker_b\",\"p95_latency_ms\":64793,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.191,\"id\":\"worker_c\",\"p95_latency_ms\":76086,\"pass_rate\":0.9258,\"provider_group\":\"provider_z\",\"quality\":0.927,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[\"writer\",\"reviewer\"],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":19.008,\"estimated_latency_ms\":158695,\"estimated_pass_rate\":0.7612,\"estimated_quality\":0.9201,\"finalizer\":\"worker_b\",\"independent_verification\":true,\"max_calls\":3,\"max_credits\":23.708,\"max_latency_ms\":201923,\"policy\":\"strict\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.4756, "category": "analysis", "complexity": 0.7655, "consequence": 0.6183, "evidence_need": 0.5121, "family": "performance_tuning", "input_tokens": 4240, "request_id": "synthetic-performance_tuning-001133", "reversibility": 0.7876, "risk": 0.4513, "summary": "Diagnose latency variance in artifact storage and rank the next experiments."}, "workers": [{"credits_per_call": 1.046, "id": "worker_a", "p95_latency_ms": 17816, "pass_rate": 0.4763, "provider_group": "provider_y", "quality": 0.6907, "strengths": ["coding", "drafting", "reasoning", "speed"]}, {"credits_per_call": 14.771, "id": "worker_b", "p95_latency_ms": 64793, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.191, "id": "worker_c", "p95_latency_ms": 76086, "pass_rate": 0.9258, "provider_group": "provider_z", "quality": 0.927, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 2, "max_credits": 3.608, "max_latency_ms": 59223, "minimum_quality": 0.6541}, "family": "migration_planning", "id": "synthetic-migration_planning-000430", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 1.115, "estimated_latency_ms": 31386, "estimated_pass_rate": 0.3952, "estimated_quality": 0.5672, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 3.608, "max_latency_ms": 59223, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_b"}, "oracle": {"candidate_count": 1, "pass_rate": 0.3952, "quality": 0.5672, "quality_target_met": false, "utility": 28.558118}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":3.608,\"max_latency_ms\":59223,\"minimum_quality\":0.6541},\"task\":{\"ambiguity\":0.3305,\"category\":\"coding\",\"complexity\":0.7275,\"consequence\":0.4241,\"evidence_need\":0.4549,\"family\":\"migration_planning\",\"input_tokens\":2720,\"reversibility\":0.2559,\"risk\":0.5026,\"summary\":\"Review a staged migration for database failover and identify irreversible steps.\"},\"workers\":[{\"credits_per_call\":3.434,\"id\":\"worker_a\",\"p95_latency_ms\":70167,\"pass_rate\":0.7915,\"provider_group\":\"provider_x\",\"quality\":0.827,\"strengths\":[\"coding\",\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.115,\"id\":\"worker_b\",\"p95_latency_ms\":31386,\"pass_rate\":0.5434,\"provider_group\":\"provider_z\",\"quality\":0.6771,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":12.545,\"id\":\"worker_c\",\"p95_latency_ms\":75567,\"pass_rate\":0.9367,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"coding\",\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":1.115,\"estimated_latency_ms\":31386,\"estimated_pass_rate\":0.3952,\"estimated_quality\":0.5672,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":3.608,\"max_latency_ms\":59223,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3305, "category": "coding", "complexity": 0.7275, "consequence": 0.4241, "evidence_need": 0.4549, "family": "migration_planning", "input_tokens": 2720, "request_id": "synthetic-migration_planning-000430", "reversibility": 0.2559, "risk": 0.5026, "summary": "Review a staged migration for database failover and identify irreversible steps."}, "workers": [{"credits_per_call": 3.434, "id": "worker_a", "p95_latency_ms": 70167, "pass_rate": 0.7915, "provider_group": "provider_x", "quality": 0.827, "strengths": ["coding", "reasoning", "review", "writing"]}, {"credits_per_call": 1.115, "id": "worker_b", "p95_latency_ms": 31386, "pass_rate": 0.5434, "provider_group": "provider_z", "quality": 0.6771, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 12.545, "id": "worker_c", "p95_latency_ms": 75567, "pass_rate": 0.9367, "provider_group": "provider_y", "quality": 0.995, "strengths": ["coding", "implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 16.495, "max_latency_ms": 290648, "minimum_quality": 0.8579}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000359", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.878, "estimated_latency_ms": 165108, "estimated_pass_rate": 0.8657, "estimated_quality": 0.9168, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 16.495, "max_latency_ms": 290648, "policy": "balanced", "reason_codes": ["complex_task", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 9, "pass_rate": 0.8657, "quality": 0.9168, "quality_target_met": true, "utility": 70.17977}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":16.495,\"max_latency_ms\":290648,\"minimum_quality\":0.8579},\"task\":{\"ambiguity\":0.6911,\"category\":\"analysis\",\"complexity\":0.8224,\"consequence\":0.6316,\"evidence_need\":0.6035,\"family\":\"performance_tuning\",\"input_tokens\":3840,\"reversibility\":0.752,\"risk\":0.4735,\"summary\":\"Choose between throughput and tail-latency changes for streaming inference.\"},\"workers\":[{\"credits_per_call\":2.926,\"id\":\"worker_a\",\"p95_latency_ms\":80568,\"pass_rate\":0.7935,\"provider_group\":\"provider_x\",\"quality\":0.9301,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.096,\"id\":\"worker_b\",\"p95_latency_ms\":23823,\"pass_rate\":0.4064,\"provider_group\":\"provider_z\",\"quality\":0.7144,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":9.952,\"id\":\"worker_c\",\"p95_latency_ms\":84540,\"pass_rate\":0.9669,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.878,\"estimated_latency_ms\":165108,\"estimated_pass_rate\":0.8657,\"estimated_quality\":0.9168,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":16.495,\"max_latency_ms\":290648,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6911, "category": "analysis", "complexity": 0.8224, "consequence": 0.6316, "evidence_need": 0.6035, "family": "performance_tuning", "input_tokens": 3840, "request_id": "synthetic-performance_tuning-000359", "reversibility": 0.752, "risk": 0.4735, "summary": "Choose between throughput and tail-latency changes for streaming inference."}, "workers": [{"credits_per_call": 2.926, "id": "worker_a", "p95_latency_ms": 80568, "pass_rate": 0.7935, "provider_group": "provider_x", "quality": 0.9301, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.096, "id": "worker_b", "p95_latency_ms": 23823, "pass_rate": 0.4064, "provider_group": "provider_z", "quality": 0.7144, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 9.952, "id": "worker_c", "p95_latency_ms": 84540, "pass_rate": 0.9669, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 8.375, "max_latency_ms": 248033, "minimum_quality": 0.7398}, "family": "migration_planning", "id": "synthetic-migration_planning-000766", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 3.705, "estimated_latency_ms": 76786, "estimated_pass_rate": 0.7938, "estimated_quality": 0.8322, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 8.375, "max_latency_ms": 248033, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_a", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7938, "quality": 0.8322, "quality_target_met": true, "utility": 63.098278}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":8.375,\"max_latency_ms\":248033,\"minimum_quality\":0.7398},\"task\":{\"ambiguity\":0.3979,\"category\":\"coding\",\"complexity\":0.603,\"consequence\":0.539,\"evidence_need\":0.4394,\"family\":\"migration_planning\",\"input_tokens\":7200,\"reversibility\":0.335,\"risk\":0.3586,\"summary\":\"Plan a zero-downtime migration of model routing with rollback checkpoints.\"},\"workers\":[{\"credits_per_call\":0.874,\"id\":\"worker_a\",\"p95_latency_ms\":29020,\"pass_rate\":0.5885,\"provider_group\":\"provider_y\",\"quality\":0.7901,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":2.831,\"id\":\"worker_b\",\"p95_latency_ms\":47766,\"pass_rate\":0.8663,\"provider_group\":\"provider_z\",\"quality\":0.8658,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":10.663,\"id\":\"worker_c\",\"p95_latency_ms\":89304,\"pass_rate\":0.995,\"provider_group\":\"provider_x\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":3.705,\"estimated_latency_ms\":76786,\"estimated_pass_rate\":0.7938,\"estimated_quality\":0.8322,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":8.375,\"max_latency_ms\":248033,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3979, "category": "coding", "complexity": 0.603, "consequence": 0.539, "evidence_need": 0.4394, "family": "migration_planning", "input_tokens": 7200, "request_id": "synthetic-migration_planning-000766", "reversibility": 0.335, "risk": 0.3586, "summary": "Plan a zero-downtime migration of model routing with rollback checkpoints."}, "workers": [{"credits_per_call": 0.874, "id": "worker_a", "p95_latency_ms": 29020, "pass_rate": 0.5885, "provider_group": "provider_y", "quality": 0.7901, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 2.831, "id": "worker_b", "p95_latency_ms": 47766, "pass_rate": 0.8663, "provider_group": "provider_z", "quality": 0.8658, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 10.663, "id": "worker_c", "p95_latency_ms": 89304, "pass_rate": 0.995, "provider_group": "provider_x", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 13.902, "max_latency_ms": 225327, "minimum_quality": 0.8336}, "family": "policy_compliance", "id": "synthetic-policy_compliance-000255", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 12.806, "estimated_latency_ms": 110097, "estimated_pass_rate": 0.7256, "estimated_quality": 0.8611, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 13.902, "max_latency_ms": 225327, "policy": "balanced", "reason_codes": ["credit_constrained", "high_consequence", "independent_evidence", "independent_verification"], "reviewer": "worker_a", "writer": "worker_c"}, "oracle": {"candidate_count": 7, "pass_rate": 0.7256, "quality": 0.8611, "quality_target_met": true, "utility": 66.803604}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":13.902,\"max_latency_ms\":225327,\"minimum_quality\":0.8336},\"task\":{\"ambiguity\":0.5416,\"category\":\"review\",\"complexity\":0.5498,\"consequence\":0.7595,\"evidence_need\":0.7765,\"family\":\"policy_compliance\",\"input_tokens\":4240,\"reversibility\":0.399,\"risk\":0.8385,\"summary\":\"Audit a workflow for API-key rotation against explicit access requirements.\"},\"workers\":[{\"credits_per_call\":11.728,\"id\":\"worker_a\",\"p95_latency_ms\":82389,\"pass_rate\":0.9582,\"provider_group\":\"provider_z\",\"quality\":0.919,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":4.075,\"id\":\"worker_b\",\"p95_latency_ms\":64041,\"pass_rate\":0.8666,\"provider_group\":\"provider_y\",\"quality\":0.8215,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.078,\"id\":\"worker_c\",\"p95_latency_ms\":27708,\"pass_rate\":0.5362,\"provider_group\":\"provider_x\",\"quality\":0.6753,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":12.806,\"estimated_latency_ms\":110097,\"estimated_pass_rate\":0.7256,\"estimated_quality\":0.8611,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":13.902,\"max_latency_ms\":225327,\"policy\":\"balanced\",\"reason_codes\":[\"credit_constrained\",\"high_consequence\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_a\",\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.5416, "category": "review", "complexity": 0.5498, "consequence": 0.7595, "evidence_need": 0.7765, "family": "policy_compliance", "input_tokens": 4240, "request_id": "synthetic-policy_compliance-000255", "reversibility": 0.399, "risk": 0.8385, "summary": "Audit a workflow for API-key rotation against explicit access requirements."}, "workers": [{"credits_per_call": 11.728, "id": "worker_a", "p95_latency_ms": 82389, "pass_rate": 0.9582, "provider_group": "provider_z", "quality": 0.919, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 4.075, "id": "worker_b", "p95_latency_ms": 64041, "pass_rate": 0.8666, "provider_group": "provider_y", "quality": 0.8215, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.078, "id": "worker_c", "p95_latency_ms": 27708, "pass_rate": 0.5362, "provider_group": "provider_x", "quality": 0.6753, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 10.884, "max_latency_ms": 150536, "minimum_quality": 0.7014}, "family": "migration_planning", "id": "synthetic-migration_planning-000937", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 5.275, "estimated_latency_ms": 94137, "estimated_pass_rate": 0.6635, "estimated_quality": 0.7906, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.884, "max_latency_ms": 150536, "policy": "balanced", "reason_codes": ["complex_task", "independent_evidence", "independent_verification"], "reviewer": "worker_b", "writer": "worker_a"}, "oracle": {"candidate_count": 4, "pass_rate": 0.6635, "quality": 0.7906, "quality_target_met": true, "utility": 56.0375}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":10.884,\"max_latency_ms\":150536,\"minimum_quality\":0.7014},\"task\":{\"ambiguity\":0.6083,\"category\":\"coding\",\"complexity\":0.8131,\"consequence\":0.6622,\"evidence_need\":0.6546,\"family\":\"migration_planning\",\"input_tokens\":2320,\"reversibility\":0.449,\"risk\":0.4702,\"summary\":\"Design compatibility testing for a private credentials schema migration.\"},\"workers\":[{\"credits_per_call\":4.219,\"id\":\"worker_a\",\"p95_latency_ms\":69203,\"pass_rate\":0.7738,\"provider_group\":\"provider_x\",\"quality\":0.89,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":1.056,\"id\":\"worker_b\",\"p95_latency_ms\":24934,\"pass_rate\":0.4809,\"provider_group\":\"provider_y\",\"quality\":0.7127,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":13.666,\"id\":\"worker_c\",\"p95_latency_ms\":75120,\"pass_rate\":0.995,\"provider_group\":\"provider_z\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":5.275,\"estimated_latency_ms\":94137,\"estimated_pass_rate\":0.6635,\"estimated_quality\":0.7906,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.884,\"max_latency_ms\":150536,\"policy\":\"balanced\",\"reason_codes\":[\"complex_task\",\"independent_evidence\",\"independent_verification\"],\"reviewer\":\"worker_b\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.6083, "category": "coding", "complexity": 0.8131, "consequence": 0.6622, "evidence_need": 0.6546, "family": "migration_planning", "input_tokens": 2320, "request_id": "synthetic-migration_planning-000937", "reversibility": 0.449, "risk": 0.4702, "summary": "Design compatibility testing for a private credentials schema migration."}, "workers": [{"credits_per_call": 4.219, "id": "worker_a", "p95_latency_ms": 69203, "pass_rate": 0.7738, "provider_group": "provider_x", "quality": 0.89, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 1.056, "id": "worker_b", "p95_latency_ms": 24934, "pass_rate": 0.4809, "provider_group": "provider_y", "quality": 0.7127, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 13.666, "id": "worker_c", "p95_latency_ms": 75120, "pass_rate": 0.995, "provider_group": "provider_z", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}]}
{"constraints": {"max_calls": 2, "max_credits": 2.247, "max_latency_ms": 34443, "minimum_quality": 0.5816}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000740", "label": {"access": {"finalizer": [], "reviewer": [], "writer": []}, "estimated_credits": 0.789, "estimated_latency_ms": 18725, "estimated_pass_rate": 0.278, "estimated_quality": 0.4906, "finalizer": null, "independent_verification": false, "max_calls": 1, "max_credits": 2.247, "max_latency_ms": 34443, "policy": "lean", "reason_codes": ["best_feasible_below_target", "complex_task"], "reviewer": null, "writer": "worker_c"}, "oracle": {"candidate_count": 1, "pass_rate": 0.278, "quality": 0.4906, "quality_target_met": false, "utility": 28.084823}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":2.247,\"max_latency_ms\":34443,\"minimum_quality\":0.5816},\"task\":{\"ambiguity\":0.6241,\"category\":\"analysis\",\"complexity\":0.7973,\"consequence\":0.5626,\"evidence_need\":0.4758,\"family\":\"performance_tuning\",\"input_tokens\":2200,\"reversibility\":0.6045,\"risk\":0.5608,\"summary\":\"Choose between throughput and tail-latency changes for artifact storage.\"},\"workers\":[{\"credits_per_call\":3.076,\"id\":\"worker_a\",\"p95_latency_ms\":67780,\"pass_rate\":0.7893,\"provider_group\":\"provider_x\",\"quality\":0.9337,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":13.185,\"id\":\"worker_b\",\"p95_latency_ms\":78011,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":0.789,\"id\":\"worker_c\",\"p95_latency_ms\":18725,\"pass_rate\":0.461,\"provider_group\":\"provider_z\",\"quality\":0.6371,\"strengths\":[\"coding\",\"drafting\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":0.789,\"estimated_latency_ms\":18725,\"estimated_pass_rate\":0.278,\"estimated_quality\":0.4906,\"finalizer\":null,\"independent_verification\":false,\"max_calls\":1,\"max_credits\":2.247,\"max_latency_ms\":34443,\"policy\":\"lean\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\"],\"reviewer\":null,\"writer\":\"worker_c\"}", "task": {"ambiguity": 0.6241, "category": "analysis", "complexity": 0.7973, "consequence": 0.5626, "evidence_need": 0.4758, "family": "performance_tuning", "input_tokens": 2200, "request_id": "synthetic-performance_tuning-000740", "reversibility": 0.6045, "risk": 0.5608, "summary": "Choose between throughput and tail-latency changes for artifact storage."}, "workers": [{"credits_per_call": 3.076, "id": "worker_a", "p95_latency_ms": 67780, "pass_rate": 0.7893, "provider_group": "provider_x", "quality": 0.9337, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 13.185, "id": "worker_b", "p95_latency_ms": 78011, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 0.789, "id": "worker_c", "p95_latency_ms": 18725, "pass_rate": 0.461, "provider_group": "provider_z", "quality": 0.6371, "strengths": ["coding", "drafting", "speed"]}]}
{"constraints": {"max_calls": 2, "max_credits": 24.214, "max_latency_ms": 215289, "minimum_quality": 0.7887}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000164", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 18.887, "estimated_latency_ms": 136888, "estimated_pass_rate": 0.9188, "estimated_quality": 0.9463, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 24.214, "max_latency_ms": 215289, "policy": "balanced", "reason_codes": ["independent_verification"], "reviewer": "worker_c", "writer": "worker_a"}, "oracle": {"candidate_count": 9, "pass_rate": 0.9188, "quality": 0.9463, "quality_target_met": true, "utility": 71.53267}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":2,\"max_credits\":24.214,\"max_latency_ms\":215289,\"minimum_quality\":0.7887},\"task\":{\"ambiguity\":0.3307,\"category\":\"analysis\",\"complexity\":0.6466,\"consequence\":0.3513,\"evidence_need\":0.4077,\"family\":\"performance_tuning\",\"input_tokens\":7600,\"reversibility\":0.6065,\"risk\":0.5016,\"summary\":\"Choose between throughput and tail-latency changes for API-key rotation.\"},\"workers\":[{\"credits_per_call\":14.994,\"id\":\"worker_a\",\"p95_latency_ms\":60912,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.995,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":1.047,\"id\":\"worker_b\",\"p95_latency_ms\":25829,\"pass_rate\":0.4482,\"provider_group\":\"provider_x\",\"quality\":0.7234,\"strengths\":[\"coding\",\"drafting\",\"speed\"]},{\"credits_per_call\":3.893,\"id\":\"worker_c\",\"p95_latency_ms\":75976,\"pass_rate\":0.9469,\"provider_group\":\"provider_z\",\"quality\":0.8716,\"strengths\":[\"reasoning\",\"review\",\"writing\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":18.887,\"estimated_latency_ms\":136888,\"estimated_pass_rate\":0.9188,\"estimated_quality\":0.9463,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":24.214,\"max_latency_ms\":215289,\"policy\":\"balanced\",\"reason_codes\":[\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_a\"}", "task": {"ambiguity": 0.3307, "category": "analysis", "complexity": 0.6466, "consequence": 0.3513, "evidence_need": 0.4077, "family": "performance_tuning", "input_tokens": 7600, "request_id": "synthetic-performance_tuning-000164", "reversibility": 0.6065, "risk": 0.5016, "summary": "Choose between throughput and tail-latency changes for API-key rotation."}, "workers": [{"credits_per_call": 14.994, "id": "worker_a", "p95_latency_ms": 60912, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.995, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 1.047, "id": "worker_b", "p95_latency_ms": 25829, "pass_rate": 0.4482, "provider_group": "provider_x", "quality": 0.7234, "strengths": ["coding", "drafting", "speed"]}, {"credits_per_call": 3.893, "id": "worker_c", "p95_latency_ms": 75976, "pass_rate": 0.9469, "provider_group": "provider_z", "quality": 0.8716, "strengths": ["reasoning", "review", "writing"]}]}
{"constraints": {"max_calls": 3, "max_credits": 10.605, "max_latency_ms": 222644, "minimum_quality": 0.8716}, "family": "performance_tuning", "id": "synthetic-performance_tuning-000923", "label": {"access": {"finalizer": [], "reviewer": ["writer"], "writer": []}, "estimated_credits": 4.66, "estimated_latency_ms": 79632, "estimated_pass_rate": 0.7033, "estimated_quality": 0.8017, "finalizer": null, "independent_verification": true, "max_calls": 2, "max_credits": 10.605, "max_latency_ms": 222644, "policy": "balanced", "reason_codes": ["best_feasible_below_target", "complex_task", "independent_verification"], "reviewer": "worker_c", "writer": "worker_b"}, "oracle": {"candidate_count": 4, "pass_rate": 0.7033, "quality": 0.8017, "quality_target_met": false, "utility": 55.831354}, "prompt": "You are the Dot Loom conductor. Select the highest-utility execution plan that obeys all hard call, credit, and latency limits. Use cheap workers when sufficient. Require an independent provider when consequences and verification needs justify it. Return exactly one compact JSON object and no prose.\nINPUT={\"constraints\":{\"max_calls\":3,\"max_credits\":10.605,\"max_latency_ms\":222644,\"minimum_quality\":0.8716},\"task\":{\"ambiguity\":0.3731,\"category\":\"analysis\",\"complexity\":0.8979,\"consequence\":0.4605,\"evidence_need\":0.3933,\"family\":\"performance_tuning\",\"input_tokens\":1560,\"reversibility\":0.7712,\"risk\":0.3438,\"summary\":\"Diagnose latency variance in background jobs and rank the next experiments.\"},\"workers\":[{\"credits_per_call\":12.733,\"id\":\"worker_a\",\"p95_latency_ms\":87786,\"pass_rate\":0.995,\"provider_group\":\"provider_y\",\"quality\":0.9931,\"strengths\":[\"implementation\",\"reasoning\",\"synthesis\"]},{\"credits_per_call\":3.733,\"id\":\"worker_b\",\"p95_latency_ms\":46550,\"pass_rate\":0.8028,\"provider_group\":\"provider_x\",\"quality\":0.8993,\"strengths\":[\"reasoning\",\"review\",\"writing\"]},{\"credits_per_call\":0.927,\"id\":\"worker_c\",\"p95_latency_ms\":33082,\"pass_rate\":0.5348,\"provider_group\":\"provider_z\",\"quality\":0.6994,\"strengths\":[\"coding\",\"drafting\",\"reasoning\",\"speed\"]}]}\nOUTPUT_SCHEMA={\"access\":{\"finalizer\":[],\"reviewer\":[],\"writer\":[]},\"estimated_credits\":\"number\",\"estimated_latency_ms\":\"integer\",\"estimated_pass_rate\":\"number\",\"estimated_quality\":\"number\",\"finalizer\":\"worker id|null\",\"independent_verification\":\"boolean\",\"max_calls\":\"integer\",\"max_credits\":\"number\",\"max_latency_ms\":\"integer\",\"policy\":\"lean|balanced|strict\",\"reason_codes\":[\"bounded machine-readable codes\"],\"reviewer\":\"worker id|null\",\"writer\":\"worker id\"}\nOUTPUT=", "schema_version": "loom-conductor-v2", "source": "synthetic-empirical-v1", "split": "validation", "target": "{\"access\":{\"finalizer\":[],\"reviewer\":[\"writer\"],\"writer\":[]},\"estimated_credits\":4.66,\"estimated_latency_ms\":79632,\"estimated_pass_rate\":0.7033,\"estimated_quality\":0.8017,\"finalizer\":null,\"independent_verification\":true,\"max_calls\":2,\"max_credits\":10.605,\"max_latency_ms\":222644,\"policy\":\"balanced\",\"reason_codes\":[\"best_feasible_below_target\",\"complex_task\",\"independent_verification\"],\"reviewer\":\"worker_c\",\"writer\":\"worker_b\"}", "task": {"ambiguity": 0.3731, "category": "analysis", "complexity": 0.8979, "consequence": 0.4605, "evidence_need": 0.3933, "family": "performance_tuning", "input_tokens": 1560, "request_id": "synthetic-performance_tuning-000923", "reversibility": 0.7712, "risk": 0.3438, "summary": "Diagnose latency variance in background jobs and rank the next experiments."}, "workers": [{"credits_per_call": 12.733, "id": "worker_a", "p95_latency_ms": 87786, "pass_rate": 0.995, "provider_group": "provider_y", "quality": 0.9931, "strengths": ["implementation", "reasoning", "synthesis"]}, {"credits_per_call": 3.733, "id": "worker_b", "p95_latency_ms": 46550, "pass_rate": 0.8028, "provider_group": "provider_x", "quality": 0.8993, "strengths": ["reasoning", "review", "writing"]}, {"credits_per_call": 0.927, "id": "worker_c", "p95_latency_ms": 33082, "pass_rate": 0.5348, "provider_group": "provider_z", "quality": 0.6994, "strengths": ["coding", "drafting", "reasoning", "speed"]}]}