Skip to content

Security: Keesan12/martin-loop

Security

SECURITY.md

Security Policy

MartinLoop governs autonomous AI coding agent runs. Security issues may include unsafe command execution, secret leakage, policy bypasses, incorrect budget enforcement, or unsafe file-scope handling.

Reporting a vulnerability

Please report security issues privately by emailing:

keesan@martinloop.com

Please include:

  • Affected version
  • Reproduction steps
  • Expected behavior
  • Actual behavior
  • Potential impact

Please do not

  • Publish exploit details before we have reviewed the issue
  • Include live secrets, tokens, or credentials in reports
  • Test against systems you do not own

Scope

Security-sensitive areas include:

  • Verifier command execution
  • Policy checks
  • File allow/deny scope handling
  • Rollback behavior
  • Persistence and run records
  • Secret-like value detection
  • Budget enforcement logic

Supported versions

The latest public release is the supported version.

Older versions may not receive fixes unless the issue is severe.

There aren't any published security advisories