agentleFS
Sign inStart free
ulises-jeremiasulises-jeremias/agent-toolkit/skills/forge/gh-fix-ci/SKILL.md

gh-fix-ci

Skill19 starsChanged 2 months agoMITView on GitHub

Diagnose failing GitHub Actions checks on a PR via gh, summarize the failure context, propose

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 30 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Inspect checks as source of truth: Use gh pr checks as the source of truth for overall PR CI state (it includes all PR-attached checks, …

    line 40
  2. Very specific:

    Add --json for machine-friendly output suitable for further processing.

    line 56
  3. Very specific:

    Add --max-lines <n> / --context <n> to widen the snippet window.

    line 57
  4. Distinctive:

    Read-only on CI; never rerun jobs or cancel workflows from this skill.

    line 84
  5. Very specific:

    Do not push commits — delegate to github-cli-workflow.

    line 85
  6. Distinctive:

    Do not implement before explicit user approval of the plan.

    line 86
  7. Distinctive:

    If the failure is clearly unrelated to the PR and appears fixed on main, merge latest main instead of bloating the PR with unrelated fixes.

    line 88

“Common” is measured: kept by at least 6 in 10 of the 30 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

Fix Failing PR Checks

Inspect failing GitHub Actions checks on a PR with gh, fetch run logs, extract a focused failure snippet, propose a fix plan before making any change, and iterate until green.

When to use

  • A PR has red checks and the user asks to "fix CI", "debug the failing job", or "why is the build red".
  • You need a fast triage of which jobs failed and the relevant log slice.
  • Branch or PR CI is failing and needs a fast, iterative path to green checks.

Prerequisites

  • gh installed and authenticated (gh auth status exits 0). doctor reports this under Integrations.
  • python3 for the bundled analyzer.
  • The current directory is inside the target Git repository (or pass --repo).

Workflow

  1. Verify auth. gh auth status. If unauthenticated, ask the user to run gh auth login (repo + workflow scopes typically required).
  2. Resolve the PR. Defaults to the current branch's PR, or pass --pr.

  3. Inspect checks as source of truth. Use gh pr checks as the source of truth for overall PR CI state (it includes all PR-attached checks, while gh run list only covers GitHub Actions):

    gh pr checks --json name,bucket,state,workflow,link

    For pending checks, gh pr checks --watch --fail-fast.

  4. Run the analyzer:

    python3 ~/.local/share/agent-toolkit/skills/gh-fix-ci/scripts/inspect_pr_checks.py \
  --repo . --pr <number-or-url>
    • Add --json for machine-friendly output suitable for further processing.
    • Add --max-lines <n> / --context <n> to widen the snippet window.

    The script handles gh pr checks field drift, falls back to job-level logs when run-level logs are not yet available, and exits non-zero when failures remain (so it composes in CI / scripts).

  5. Scope external providers. Any check whose detailsUrl is not a GitHub Actions run is reported as external with only the URL — do not attempt to fetch logs from Buildkite/CircleCI/etc. Use the check link to identify the failing command or service when logs are unavailable.
  6. Summarize. For each failing check report: name, run URL, conclusion, workflow, branch/SHA, and the failure snippet (or "logs pending" if the run is still in progress). Identify the primary failing job and root error.
  7. Draft a plan. Propose the fix as a short, numbered plan. Do not edit files yet. If the project has a planning skill (e.g. planning), delegate the plan structure to it.
  8. Implement after approval — one failure at a time. Once the user approves the plan, apply the smallest safe fix. Prefer minimal, low-risk changes before broader refactors. Summarize the diff and delegate the push/PR-update step to github-cli-workflow.
  9. Recheck and iterate until green. Re-run gh pr checks after each push; the check set can change. If checks already failed, diagnose those first; if pending, watch with --watch --fail-fast. Repeat steps 4-8 for the next actionable failure until all required checks are green. Suggest re-running relevant tests locally between iterations.

Boundaries

  • Read-only on CI; never rerun jobs or cancel workflows from this skill.
  • Do not push commits — delegate to github-cli-workflow.
  • Do not implement before explicit user approval of the plan.
  • Fix one actionable failure at a time.
  • If the failure is clearly unrelated to the PR and appears fixed on main, merge latest main instead of bloating the PR with unrelated fixes.

Bundled resources

  • scripts/inspect_pr_checks.py — failure analyzer (stdlib only). Returns text by default; --json for machine-readable output.

Output

  • Primary failing job and root error
  • Fixes applied in iteration order
  • Current CI status and next action

Validation

  • skills check reports the gh and python3 requirements.
  • doctor shows gh auth status under Integrations.

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