ci-fix

Diagnose and fix failing CI/CD pipelines by classifying failures and applying targeted playbooks.

15|3|Updated Jul 9, 2026
One-click install
npx skills add https://github.com/kiurakku/cursor-kit-for-ai --skill ci-fix-kiurakku
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ci-fix
Source: https://github.com/kiurakku/cursor-kit-for-ai/tree/main/plugins/devops/skills/ci-fix
Command: npx skills add https://github.com/kiurakku/cursor-kit-for-ai --skill ci-fix-kiurakku

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? CI failures waste hours when developers treat environment problems as code problems or blindly re-run flaky tests. This Skill provides a systematic workflow to read the actual failure logs, classify the root cause, and apply the matching fix playbook. ## Core Features & Use Cases - Failure Classification: Sort failures into code, environment mismatch, flaky, infrastructure, or config categories using concrete signals from the logs. - Targeted Playbooks: Step-by-step fixes for version mismatches, missing services, secrets, filesystem issues, flaky tests, and dependency/registry problems. - Pipeline Hardening: Add timeouts, concurrency groups, caching, and pinned action versions to prevent recurrence. - Use Case: A test suite passes locally but fails in GitHub Actions. Use this Skill to diff local vs CI environments, identify a missing Postgres service, and add a services block with health checks to the workflow. ## Quick Start Ask the AI to debug why the GitHub Actions workflow is failing even though all tests pass locally, and have it classify the failure and propose a fix.

Frequently Asked Questions about ci-fix

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

FAQPage Schema
How do I debug a failing GitHub Actions workflow?▼

Start by reading the actual failure log with 'gh run view <run-id> --log-failed' to find the first error, since later errors cascade from it. Then classify the failure as a code, environment, flaky, infrastructure, or config problem before applying a fix.

Why do tests pass locally but fail in CI?▼

Environment mismatches are the usual cause: different language versions, missing services like Postgres or Redis, unset environment variables, case-sensitive filesystems on Linux runners, or limited CI resources. Reproduce the failure by running the same commands inside the same container image locally.

How do I fix flaky tests in CI pipelines?▼

Confirm flakiness by running the test repeatedly, then fix the root cause: shared state between parallel tests, time-based assertions, or external network calls. Isolate state with unique fixtures, freeze time, and mock network calls rather than re-running until green.

Should I re-run a failed CI job to make it pass?▼

No. Re-running until green hides real race conditions and flaky tests that indicate genuine bugs. Confirm whether the failure is flaky, fix the underlying cause, or quarantine the test with a tracking issue if it cannot be fixed immediately.

How do I fix CI failures caused by dependency or registry errors?▼

Pin action versions like actions/checkout@v4, add package caching to reduce network exposure, and bump the cache key if a corrupted cache causes stale failures. For registry rate limits or timeouts, add retries to the specific step or mirror dependencies.