agentleFS
Sign inStart free
kaeawckaeawc/auto-mobile/skills/check-ci/SKILL.md

check-ci

Skill56 starsChanged 35 days agoApache-2.0View on GitHub

Use this workflow skill for PR CI triage: inspect pending and failed checks for the exact PR head, read live or completed job logs and artifacts, classify causality, reproduce likely failures locally, and summarize the safe next step.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this for PR health checks and failure triage.

    line 8
  2. Very specific:

    If the log is truncated or lacks the failure body, list the run's artifacts and read the relevant one.

    line 30
  3. Distinctive:

    After any push, restart at step 1 — every check, thread, and comment was scoped to the previous SHA.

    line 38
  4. Distinctive:

    Run/job ledger, including pending jobs and duplicate or stale checks

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Check CI

Use this for PR health checks and failure triage.

  • This is narrower than pr-analysis: prefer check-ci when the main question is why checks are failing or blocked.
  • This is read-only unless the user also asks to fix the problem.

Workflow

  1. Resolve the PR number and snapshot headRefOid, baseRefOid, mergeability, and merge state. Use that SHA for every subsequent query. An empty legacy combined status is inconclusive, not green.
  2. Collect gh pr checks <pr> --json name,state,bucket,link,workflow plus paginated repos/<owner>/<repo>/commits/<head-sha>/check-runs. Branch on bucket: only fail is a failure — skipping and cancel are not. Flag duplicate check names with different conclusions on the same head (a cancelled older run beside a successful newer one); the stale one is what parks automerge.
  3. Enumerate runs for that head SHA, then each run's jobs. Ignore runs for older SHAs. Queued and in-progress jobs stay pending — they are never implicitly green.
  4. For a completed failed job: gh run view <run-id> --job <job-id> --log-failed, or gh api repos/<owner>/<repo>/actions/jobs/<job-id>/logs — the latter works as soon as that job finishes, even while sibling jobs still run, so it does not wait on the slowest leg. For a job still executing, gh api .../actions/jobs/<job-id> returns per-step status naming the running or failed step. gh api --follow does not exist; use --paginate.
  5. If the log is truncated or lacks the failure body, list the run's artifacts and read the relevant one. If gh run download is unauthorized, state that exact blocker — do not infer a root cause from partial evidence.
  6. Classify every red result as PR-caused, pre-existing on main, transient/infrastructure, or unproven, grounded in the job log, the changed paths, and current origin/main. Re-run a job only when the evidence supports transience. Never change code for an unrelated or unproven failure.
  7. Reproduce a PR-caused failure locally with the narrowest authoritative command.
  8. After any push, restart at step 1 — every check, thread, and comment was scoped to the previous SHA.

For PR discussion and review-thread triage, use github-pr-feedback; for the underlying gh and GraphQL mechanics, github-cli.

Repo Validation Mapping

  • TypeScript lint or build: bun run lint, bun run build
  • Tests: bun test or bun test <file>
  • Fast repo checks: bash scripts/all_fast_validate_checks.sh
  • Android: (cd android && ./gradlew <task>)
  • iOS: bash scripts/ios/swift-build.sh or bash scripts/ios/swift-test.sh

Output Structure

  • Current CI state, scoped to the head SHA
  • Run/job ledger, including pending jobs and duplicate or stale checks
  • Failing checks, the evidence, and the causal classification
  • Mergeability or conflict status
  • Artifact or authorization limitations
  • Recommended next action

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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