agentleFS
NumericalEarthNumericalEarth/Breeze.jl/.claude/skills/babysit-ci/SKILL.md

babysit-ci

Skill91 starsChanged 4 days agoApache-2.0View on GitHub

Use when asked to watch or fix CI on a PR or branch. Fixes mechanical failures, retriggers infrastructure flakes, and stops to report anything that needs judgment.

  • Commits and pushes

8 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use addskill, apply-firmware and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • ryanportfolioSkill

    addskill

    Create, import, update, or install repo or personal skills: runtime ownership, resources, discovery checks.

    Also in 6 projects that use this

  • ryanportfolioSkill

    apply-firmware

    Bring current folder (plain folder, existing repo, old or partial firmware copy) up to full Harness Firmware in place. Mirror external repo → adopt-repo.

    Also in 6 projects that use this

  • +10 more from ryanportfolio

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 5

    Distinctive

    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
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Check CI status. For each failed job:

    line 26
  2. Distinctive:

    After fixing, commit with a descriptive message and push.

    line 45
  3. Distinctive:

    Stop and explain the problem to the user for:

    line 70
  4. Very specific:

    Always check gh run view <RUN_ID> --log-failed before acting — don't guess from job names

    line 90
  5. Distinctive:

    After pushing a fix, wait for CI to pick it up before checking again

    line 91
  6. Distinctive:

    Never force-push or rewrite history to fix CI — always add new commits

    line 92

“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

Babysit CI

Monitor GitHub Actions CI for the current branch/PR. Auto-fix small issues (whitespace, missing imports, formatting). Pause and describe anything that requires judgment. Retrigger flaky jobs.

Step 1: Find the CI Run

# Get the current branch
git branch --show-current

# Find the latest CI run for this branch
gh run list --branch $(git branch --show-current) --limit 5

# Or for a specific PR
gh pr checks <PR_NUMBER>

Step 2: Monitor Loop

Check CI status. For each failed job:

gh run view <RUN_ID>
gh run view <RUN_ID> --log-failed

Step 3: Triage Each Failure

Auto-fix (commit and push without asking)

FailureHow to fix
Whitespace checkRemove trailing whitespace, ensure final newline, no trailing blank lines
Missing explicit importAdd the missing import to the appropriate file
Doctest output mismatchUpdate the expected output only if the PR intentionally changed what is printed; otherwise pause
Typo in error message or docstringFix the typo
Unused import warningRemove the unused import

After fixing, commit with a descriptive message and push. Continue monitoring.

Retrigger (likely flaky)

"Unrelated to the changed files" is not enough on its own: show the log line that points to infrastructure, or the same failure on main.

SignalAction
Test passed locally but fails in CIRetrigger
Timeout with no test failureRetrigger
Network/download errorRetrigger
Pkg.instantiate failureRetrigger
CI infrastructure errorRetrigger
The same job fails the same way on mainRetrigger, and say it also fails on main
gh run rerun <RUN_ID> --failed
# Or push an empty commit to retrigger all jobs
git commit --allow-empty -m "Retrigger CI"
git push

Pause and describe (needs judgment)

Stop and explain the problem to the user for:

  • Test logic failure: assertion fails that isn't a doctest output update
  • Type instability or GPU error: "dynamic invocation" or similar
  • Regression test failure: numerical results don't match
  • Build failure: package won't precompile or load
  • Multiple interrelated failures: several jobs with a common root cause

Report: which job(s) failed, relevant error output, likely cause, suggested fix options.

Step 4: Confirm All Green

gh pr checks <PR_NUMBER>

Report the final status to the user.

Notes

  • Always check gh run view <RUN_ID> --log-failed before acting — don't guess from job names
  • After pushing a fix, wait for CI to pick it up before checking again
  • Never force-push or rewrite history to fix CI — always add new commits
  • GPU jobs run on self-hosted aws-linux-nvidia-gpu-* runners (tests in CI.yml, docs in Documentation.yml); Validation.yml runs the BOMEX example

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.