Individual skill

github-ci-advisor

Trigger: CI advisor, build pipeline, test pipeline, GitHub Actions, CI audit, security check, CI evolution. Propose, and only after approval apply, least-privilege CI for an existing repository.

skills/github-ci-advisor/SKILL.md

Read pinned source

Source commit 23d623077b0764d17c90a43fe0dc1f320fa7e181

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
Yes
Repository scan complete
No
Skill scan complete
Yes
Observed behavior
Network or broad shell
Last checked
Nov 5, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings)
23
Repository last updated
Oct 13, 2026
License
MIT

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files
Read excerpt

# GitHub CI Advisor

## Activation Contract

Use for CI advice, build/test pipelines, audits, security checks, or CI evolution (add, modify, retire). Do not activate automatically during repository bootstrap, and invent no fixed CI manifest or unknown schema fields. It proposes, applies only after explicit approval, and never commits, pushes, releases, deploys, or enables checks or protection.

## Hard Rules

- Discover read-only: read manifests, lockfiles, and script definitions; inspect workflows, triggers, action pins, permissions, and required checks. Use the API for required checks only with authorized, valid access. Never run project scripts, install hooks, or execute downloaded code as discovery.
- Propose from observed commands only: build, typecheck, and tests when they exist, otherwise a smoke check. Ask the human for unset requirements, skip unwanted modules, and never guess a universal CI from an assumed npm project.
- Pin every action to an immutable SHA with human-verified provenance; without verified provenance, stop with unresolved pins and invent no SHA, version, dependency, or code.
- Least privilege: `contents: read`, no secrets to forks, no privileged checkout of a pull-request target. Run untrusted pull-request code only sandboxed without secrets or a writable token; containers reduce risk but are not an absolute sandbox.
- Baseline with a pinned secrets scanner (e.g. a verified-ref Gitleaks action), a secrets-permissions review, and a human read-only repository-safety review. Never claim an action detects prompt injection. Run deterministic scanners only when installed and approved, fetch external evidence only when safe, invent nothing.
- Retire a check only when proven redundant or obsolete; preserve required check names or explicitly warn that branch rules depend on them and need a separate manual change, since branch protection does not manage checks.