Skip to content

Security: mmlqm/ai-code-security-review

Security

SECURITY.md

Security Policy

Supported Scope

This project performs offline defensive code review and CI delivery gating.

Supported reports include:

  • False negatives or false positives in local source scanning.
  • Secret redaction issues in reports.
  • SARIF or CI output problems.
  • Unsafe behavior in the scanner itself.
  • Packaging or skill metadata issues.

Out of scope:

  • Live target scanning.
  • Exploit development.
  • Credential attacks.
  • Bypass generation.
  • Reconnaissance workflows.

Reporting

Do not open a public issue with real secrets, private code, or sensitive reproduction details.

If a finding involves sensitive material, report only:

  • A short description of the issue.
  • The affected file or feature area.
  • A minimal synthetic reproduction when possible.
  • Whether the issue can expose secrets in output.

If no private channel is configured for this repository, create a minimal public issue without sensitive data and request a maintainer contact path.

Secret Handling

If a real token, password, private key, or cloud credential was committed or pasted into a workflow, rotate it immediately. Removing it from a later commit is not enough.

The scanner redacts common secret assignments in reports, but it cannot guarantee that every organization-specific token format is masked.

There aren't any published security advisories