Skip to content

Security: nathancrtr/gateline

SECURITY.md

Security Policy

Reporting a vulnerability

Please report suspected vulnerabilities privately through GitHub's security advisory form on this repository ("Report a vulnerability" under the Security tab). You will get an acknowledgment within a week; fixes for confirmed issues in the framework's tooling are prioritized ahead of all other work.

Please do not open public issues for suspected vulnerabilities.

Scope

The security-relevant surfaces of gateline are:

  • packages/framework (gateline init|validate|fork|render) — this runs on operator machines against operator repos, and the workflow it writes into a host checks out a pinned framework ref in that host's CI. Anything that could make it write outside the target repo, execute unexpected code, pin a ref the lock does not record, or misrepresent provenance is in scope.
  • Gatehouse (packages/) — the server binds to localhost by design; anything that widens that exposure, bypasses the single write path, or lets a non-human author a gate entry is in scope.
  • Role and contract text — prompt-injection vectors that would cause a pipeline agent to exceed its contract are in scope; reports should include the artifact that carried the injection.

Supported versions

Pre-1.0: fixes land on main only. Once tagged releases exist, this policy will name the supported tags.

What gateline does not do

gateline ships no hosted service and collects no telemetry. The network calls it does make are to your own repository: git push/fetch to your own origin, and, when you configure a token, GitHub API and webhook traffic (gateline sync, PR-review import, the server's /api/webhooks/github endpoint, which is HMAC-verified) scoped to that one repository. The agents you dispatch use whatever runtime and credentials you already have. Your model vendors' security posture is governed by your agreements with them, not by this project.

There aren't any published security advisories