# Security policy

Do not include credentials, private prompts, hidden benchmark material, or
personal data in a public report.

Report suspected malicious files, credential exposure, unsafe serialization,
or other security issues privately through the contact channel at
https://ulam.ai/. Include the model revision or immutable Hub commit, affected
filename, reproduction steps, and impact. Please do not open a public
discussion containing secrets or restricted evaluation content.

The canonical release uses safetensors weights and ordinary Transformers model
classes. Loading the model must not require `trust_remote_code=True`.