Skip to content

Security: FrancoisChastel/secure-inference

Security

SECURITY.md

Security policy

Veil is research software and is not recommended for protecting production data. The default demos use an insecure trusted preprocessing dealer; read docs/THREAT_MODEL.md before evaluating any security claim.

Reporting a vulnerability

Please use GitHub's private Security advisories → Report a vulnerability flow for this repository. Do not include real prompts, keys, masks, integrity sketches, or proprietary model artifacts in a public issue.

For ordinary correctness bugs and documentation issues, open a regular GitHub issue.

Supported version

Only the latest commit on main is supported during the proof-of-concept phase.

There aren't any published security advisories