agentleFS
Sign inStart free
jMertajMerta/codex-skills/ci-fix/SKILL.md

ci-fix

Skill130 starsChanged 11 months agoMITView on GitHub

Use when the user asks to diagnose or fix failing GitHub Actions checks on a pull request, branch, or workflow run, or wants CI log and root-cause analysis. Prefers GitHub connector context, uses gh for Actions logs, and verifies authorized fixes locally.

We split this skill into the rules it gives an agent

7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer the GitHub connector for repository, pull request, patch, review, and check metadata.

    line 10
  2. Very specific:

    Use gh for current-branch PR discovery and GitHub Actions runs, jobs, artifacts, and logs.

    line 11
  3. Distinctive:

    Do not claim the connector can retrieve Actions logs when that capability is unavailable.

    line 12
  4. Distinctive:

    Treat non-GitHub Actions checks as external: report their status and URL unless the user asks for a separate provider investigation.

    line 13
  5. Very specific:

    Use gh pr checks only to discover the failing check and its run or job identifiers.

    line 19
  6. Very specific:

    Treat workflow triggers, token permissions, secrets, pull_request_target, and execution of fork code as security-sensitive.

    line 23
  7. Distinctive:

    Run the closest local equivalent.

    line 24

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

CI fix

Tool routing

  • Prefer the GitHub connector for repository, pull request, patch, review, and check metadata.
  • Use gh for current-branch PR discovery and GitHub Actions runs, jobs, artifacts, and logs.
  • Do not claim the connector can retrieve Actions logs when that capability is unavailable.
  • Treat non-GitHub Actions checks as external: report their status and URL unless the user asks for a separate provider investigation.

Workflow

  1. Resolve the repository and failing PR, branch, or run from the request and local Git context. Ask only if the target remains ambiguous.
  2. Inspect the failing checks and record the check name, run URL, job, failing step, head SHA, and conclusion.
  3. Use gh pr checks only to discover the failing check and its run or job identifiers. Pull the smallest useful log slice with gh run view <run-id> --log-failed, gh run view <run-id> --job <job-id> --log, or the job-logs endpoint through gh api. Never expose secrets from logs.
  4. Compare the failure with the local diff and a known baseline. Distinguish a regression from flaky infrastructure, external service failure, missing credentials, or an unrelated baseline failure.
  5. State the evidence-backed root cause and focused fix. Stop after diagnosis when the request does not authorize code or workflow changes.
  6. Apply the smallest authorized fix. Read references/ci-failure-playbook.md only when its failure class is relevant.
  7. Treat workflow triggers, token permissions, secrets, pull_request_target, and execution of fork code as security-sensitive. Do not broaden access merely to make a job pass.
  8. Run the closest local equivalent. Push, rerun, or dispatch workflows only when the request authorizes that external change.
  9. Recheck the exact failing check and report residual failures, unavailable logs, and unverified assumptions.

Deliverable

Report the failing run and step, root cause, focused diff, local checks, resulting run status, and remaining risk. Do not report CI as green without a completed successful run.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free