Skip to content

Security: andrepontesmelo/archloop

Security

SECURITY.md

Security Policy

Scope

archloop is a local automation driver: it sequences model sessions that read and write code in an isolated worktree, then merges reviewed results. It does not hold credentials — model keys live in the session runner's config, not here.

Supported versions

Only the latest commit on main receives security fixes.

Reporting a vulnerability

Email the owner via the contact on the GitHub profile (andrepontesmelo) rather than opening a public issue. Include: affected commit, the driver script and stage involved, and expected vs actual behavior. You will get an acknowledgement within 7 days and a fix or a documented mitigation for anything confirmed.

What is NOT a vulnerability

  • The loop executing agent-written code in a worktree — that is the product's purpose. Review configs before unattended overnight runs; the loop never touches the canonical tree but it does run what the sessions wrote.
  • A SHIPped change that turns out to be wrong — that is a review-quality issue, not a security issue. Read gate-<item>.md and the morning diff; a bad run costs one git reset --hard plus branch delete.
  • Driver logs containing repo paths or model output — artifacts by design, never credentials. Do not paste unredacted logs into public issues.

There aren't any published security advisories