Skip to content
Get packs

CI Failure Patrol

Sorts stale pull-request CI failures into rerun, comment, or no-action decisions

Source Download skill
Repository
QwenLM/qwen-code
Stars
28,880
Licence
Apache-2.0
Files
1 file

What it does

This skill teaches an AI agent to review a batch of recently failed continuous-integration checks on open pull requests and decide what to do with each one. It decides whether a failure looks like a transient infrastructure or flaky-test problem (rerun), clearly caused by the PR's own changed files (comment), or ambiguous (no action). It does not call GitHub itself; a separate JavaScript driver handles reads, validation, state and writes, and the agent only classifies the candidates and writes a decisions file. It also treats log text as untrusted data and ignores any instructions embedded in it.

Great for

  • Maintainers of large open-source repositories who want help triaging flaky CI failures on stale pull requests
  • Teams building automated CI triage loops that use a driver script and a JSON handoff
  • Developers studying how to structure safe, bounded agent tasks with explicit output schemas

Skip it if

  • Individual users wanting a general CI debugging helper: it only classifies batches of candidates in a driver-based workflow and does not fix code or explain failures interactively
  • Main-branch failures: the skill explicitly excludes them
  • Repositories without a driver script that performs GitHub reads and writes: the skill assumes that infrastructure exists and will not work correctly on its own

Try asking

  • “Classify the stale PR CI failures in ci-flaky-input.json and write ci-flaky-decisions.json”
  • “Decide whether this batch of failed runs is a transient runner timeout that should be rerun”
  • “Identify genuinely nondeterministic tests in these failure logs so a deflake fix can be opened”
Step-by-step guide: how to use CI Failure PatrolInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    ci-flaky-patrol.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip ci-flaky-patrol.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip ci-flaky-patrol.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Only run it through the driver workflow that supplies ci-flaky-input.json; the skill cannot operate on GitHub by itself
  • Review the decisions file before the driver applies actions, especially comment decisions, since they post to real pull requests
  • Check flakyTest entries against the log to confirm the test name is verbatim, because malformed entries are silently ignored

What's inside

  • SKILL.md

Redistributed unmodified from QwenLM/qwen-code under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.