Skip to content

Security: Synter-Media-AI/agent-shield

Security

SECURITY.md

Security Policy

Supported Versions

Security fixes are supported for the latest published release.

Because this project is still small and evolving, older releases may not receive backported fixes.

Reporting A Vulnerability

Do not open a public GitHub issue for suspected vulnerabilities.

Email hello@syntermedia.ai with:

  • a description of the issue;
  • the affected version or commit;
  • reproduction steps or a proof of concept;
  • any proposed mitigation or fix, if available.

We will acknowledge receipt as soon as possible and coordinate a fix and disclosure timeline.

Scope Notes

When reporting issues, keep the project’s scope in mind:

  • HMAC signing here provides authenticity only for parties that share the same secret.
  • Manifest verification here proves file-content consistency against a signed manifest, not artifact provenance.
  • Prompt sanitization is heuristic and may have false positives or false negatives.
  • Execution policy helpers return decisions, but the host runtime remains responsible for enforcement.

There aren't any published security advisories