agentleFS
GavinGreenwoodGavinGreenwood/agentic-workflow-template/.agents/skills/nightly-check/SKILL.md

nightly-check

Skill5 starsChanged 44 days agoMITView on GitHub

User-invoked only. Triage scheduled CI runs: flake vs regression vs config vs infra

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 9 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Use the gh CLI to fetch scheduled workflow runs for this repo.

    line 15
  2. Distinctive:

    Run ID and workflow name

    line 23
  3. Distinctive:

    Read the logs carefully and identify the root cause.

    line 33
  4. Distinctive:

    If all nightly runs are passing, output: "✅ All nightly runs passing — nothing to action." and stop.

    line 52
  5. Distinctive:

    Read the relevant source files to understand the failure.

    line 59
  6. Distinctive:

    Ask: "Should I create a Jira ticket and pick it up for [failure description]?"

    line 61
  7. Very specific:

    If yes → read and follow .agents/skills/capture/SKILL.md to log it, then read and follow .agents/skills/pickup/SKILL.md with the new ticket …

    line 62
  8. Distinctive:

    If no → note it and move on.

    line 63
  9. Distinctive:

    Check if the flake has occurred more than twice in the last 7 days.

    line 67
  10. Distinctive:

    If yes → it is chronic and needs fixing.

    line 68
  11. Distinctive:

    If no → log it as a known flake and move on.

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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 nightly CI runs and triage failures

Review the most recent nightly workflow runs for this repo, diagnose any failures, and action them.


Step 1 — Fetch recent scheduled runs

Use the gh CLI to fetch scheduled workflow runs for this repo. Look at the last 7 days of nightly (schedule-triggered) runs.

gh run list --event schedule --limit 30 --json databaseId,displayTitle,status,conclusion,createdAt,headSha,workflowName

Filter to runs from the last 7 days. For each run, note:

  • Run ID and workflow name
  • Triggered at (date/time)
  • Overall status (success / failure / in_progress)

Step 2 — Identify failures

List every failed nightly run. For each failure:

  1. Fetch the jobs: gh run view <run-id> --json jobs
  2. For each failed job, fetch the full logs: gh run view <run-id> --log-failed
  3. Read the logs carefully and identify the root cause.

Categorise each failure as one of:

  • Flake — non-deterministic failure (network timeout, race condition, transient infra)
  • Regression — a code change broke something
  • Config — GitHub Actions workflow config, environment variable, or secret issue
  • Infra — underlying infrastructure problem (disk, memory, runner crash)
  • Unknown — cannot determine from logs alone

Step 3 — Summarise findings

Print a table:

| Pipeline | Date | Status | Job | Category | Root cause (1 line) |
|----------|------|--------|-----|----------|---------------------|

If all nightly runs are passing, output: "✅ All nightly runs passing — nothing to action." and stop.

Step 4 — For each Regression or Config failure

These are actionable. For each one:

  1. Identify the affected files, tests, or config.
  2. Read the relevant source files to understand the failure.
  3. Propose a concrete fix plan (what to change, where, why).
  4. Ask: "Should I create a Jira ticket and pick it up for [failure description]?"
    • If yes → read and follow .agents/skills/capture/SKILL.md to log it, then read and follow .agents/skills/pickup/SKILL.md with the new ticket key to start work.
    • If no → note it and move on.

Step 5 — For each Flake failure

  1. Check if the flake has occurred more than twice in the last 7 days.
  2. If yes → it is chronic and needs fixing. Propose a fix plan and ask whether to create a ticket.
  3. If no → log it as a known flake and move on.

Step 6 — For Infra failures

Flag to the user. These are not fixable via code — they need infra investigation. Note the pipeline number and affected jobs so the user can escalate.

Step 7 — Final report

Summarise what was found and what was actioned:

  • Pipelines reviewed
  • Failures found (by category)
  • Tickets created (if any)
  • Recommended next steps

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.