github-ci-failure-triage
Skill6 starsChanged 4 months agoApache-2.0View on GitHub
Use when a GitHub PR shows failing checks and a clean failure readout is needed before changing any code.
In this repository116 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- skills/graphify/references/CLAUDE.mdCLAUDE.md
- agent-config-security-auditSkill
- agent-harness-designSkill
- agent-readable-docsSkill
- agent-runtime-patternsSkill
- agent-session-forensicsSkill
- ai-regression-testingSkill
- ai-slop-cleanerSkill
- api-doc-reviewSkill
- architecture-decision-recordsSkill
- architecture-simplificationSkill
- ask-claudeSkill
- ask-codexSkill
- ask-geminiSkill
- autonomous-delivery-pipelineSkill
- baseline-comparison-reviewSkill
- bigquery-cost-auditSkill
- brainstormingSkill
- browser-automation-safetySkill
- bug-huntSkill
- build-fix-minimal-diffSkill
- canonical-advisor-routingSkill
- change-impact-checkSkill
- changelog-generatorSkill
- ci-failure-readoutSkill
- codebase-migration-batchesSkill
- context-compressionSkill
- context-docsSkill
- context-minimizationSkill
- context-retrieval-loopSkill
- data-qualitySkill
- dataset-split-reviewSkill
- dbt-incremental-strategy-auditSkill
- dbt-preflightSkill
- dbt-strategySkill
- dead-code-removalSkill
- defuddleSkill
- dependency-cleanupSkill
- deploy-pipeline-runbookSkill
- dispatching-parallel-agentsSkill
- doc-consistency-checkSkill
- environment-doctorSkill
- experiment-tracking-reviewSkill
- failure-triageSkill
- filesystem-mcp-guardrailsSkill
- finishing-a-development-branchSkill
- flaky-test-investigationSkill
- gh-address-commentsSkill
- graphifySkill
- handoff-disciplineSkill
- hard-won-skill-extractorSkill
- hook-safety-reviewSkill
- incident-readoutSkill
- incremental-migrationSkill
- json-canvasSkill
- local-rag-mcpSkill
- log-driven-diagnosisSkill
- managed-codebase-contextSkill
- The first 60 shown.
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
- Very specific: line 25
Prefer GitHub-native data (gh or a GitHub connector) over screenshots or copied snippets
- Distinctive: line 28
Do not auto-commit, auto-push, or auto-rerun remote writes without approval
- Distinctive: line 48
If a check is external (not GitHub Actions), report its URL and keep it out of local log parsing.
- Distinctive: line 61
Implement only when the user wants the fix: Keep changes traceable to the failing signal.
≈ “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.
Files
Purpose
Produce a clean, GitHub-native readout of failing PR checks before any code is edited. Triage first: summarize the breakage and propose a local fix plan; implement only after approval when the task is framed as triage-first.
When to Use
- A PR is red and the cause is unknown
- The user asks why CI is failing before asking for a fix
- A triage-first readout must precede any code edits
Inputs
- The PR number or URL (or the current branch to resolve it from)
ghCLI authenticated for the repository, or a GitHub connector
Core Rules
- Resolve the PR first
- Inspect failing checks before guessing
- Prefer GitHub-native data (
ghor a GitHub connector) over screenshots or copied snippets - Summarize the failure before proposing code changes
- Implement only after approval when the task is framed as triage-first
- Do not auto-commit, auto-push, or auto-rerun remote writes without approval
Workflow
Resolve the PR:
gh pr view <number-or-url> --json number,title,headRefName,url gh pr view --json number,title,headRefName,url # from the current branchInspect failing checks. List checks, isolate failing GitHub Actions jobs, and pull only the failing logs:
gh pr checks <number> gh run list --branch <head-branch> --limit 10 gh run view <run-id> --json jobs # job-level status gh run view <run-id> --log-failed # logs from failing steps onlyIf a check is external (not GitHub Actions), report its URL and keep it out of local log parsing.
- Summarize the failure:
- failing job
- likely root cause
- relevant log snippet
- affected files or subsystem
- Propose the fix plan:
- what to change locally
- what to validate locally first
- what remains remote-only
- Implement only when the user wants the fix. Keep changes traceable to the failing signal; rerun the narrowest meaningful local validation first.
Bundled Helper
scripts/inspect_pr_checks.py— use it when you want a structured check summary instead of rawghoutput.
Output
- Failing job(s) with a root-cause hypothesis and the minimal relevant log snippet
- Affected files or subsystem
- A local fix plan with validation steps, separated from remote-only actions
Verification
- PR resolved explicitly (number/URL or current branch)
- Failure evidence pulled from GitHub-native data, not screenshots or pasted snippets
- Log excerpt limited to the failing step, not the whole run log
- Summary written before any code change was proposed
- No remote writes (commit/push/rerun) performed without approval
Failure Modes
- Guess-first triage — proposing fixes before reading the failing job's log.
- Log dumping — pasting entire run logs instead of
--log-failedoutput for the failing step. - Unapproved remote writes — rerunning workflows or pushing while the task was triage-first.
- Parsing external checks locally — non-Actions checks only expose a URL; report it instead.
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
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.

