ci-watcher

Monitor CI/CD pipelines and diagnose failing checks until terminal state.

7|Updated Dec 5, 2023
One-click install
npx skills add https://github.com/mParticle/aquarium --skill ci-watcher-mparticle
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ci-watcher
Source: https://github.com/mParticle/aquarium/tree/main/.claude/skills/ci-watcher
Command: npx skills add https://github.com/mParticle/aquarium --skill ci-watcher-mparticle

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

This Skill reduces manual CI triage by continuously monitoring CI/CD pipelines until they reach a terminal state, diagnosing failures, and attempting safe recovery actions so maintainers spend less time debugging flaky or unrelated pipeline failures.

Core Features & Use Cases

  • Real-time Monitoring: Polls PRs, branches, or run IDs and reports check status, duration, and a summary table on each cycle.
  • Failure Classification: Distinguishes RELATED, UNRELATED, and EXTERNAL failures by comparing changed files and extracting failing job logs.
  • Self-Healing & Retriggers: Attempts minimal surgical fixes for related failures, retriggers flaky E2E checks with limits, and escalates when automated recovery fails.
  • Use Case: Run this Skill after a push to a feature branch to watch GitHub Actions runs, auto-retrigger flaky tests, and surface actionable diagnostics for maintainers.

Quick Start

Use the ci-watcher to monitor PR 123 on GitHub Actions, diagnose failing checks, and attempt safe retriggers or fixes until the pipeline becomes green.

Frequently Asked Questions about ci-watcher

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I monitor GitHub Actions runs and automatically fix failing CI checks?

To monitor GitHub Actions runs and auto-fix failing CI checks, use ci-watcher to poll your pipelines, classify failures by comparing changed files to logs, and attempt minimal code fixes until a terminal state is reached.

Can I automatically retrigger flaky E2E tests in Buildkite without manual intervention?

Yes, you can automatically retrigger flaky E2E tests in Buildkite using ci-watcher, which identifies flaky checks and limits retriggers to safe thresholds while escalating failures that automated recovery cannot resolve.

How does CI failure classification distinguish between related and unrelated pipeline failures?

CI failure classification distinguishes related from unrelated pipeline failures by retrieving failing job logs via the gh CLI and comparing the changed files in a pull request against the failing tests to categorize the root cause.

Does ci-watcher work with stalled external checks on pull requests?

Yes, ci-watcher handles stalled external checks on pull requests by performing periodic polling, reporting check status and duration, and surfacing actionable diagnostics for maintainers until the pipeline reaches a terminal state.

What is the best way to reduce manual CI triage for flaky tests on feature branches?

The best way to reduce manual CI triage for flaky tests on feature branches is to run ci-watcher after a push, which continuously monitors pipelines, diagnoses failures, and attempts safe recovery actions like surgical fixes and retriggers.