Automatically identifies patterns in unreliable tests and applies deterministic fixes to improve CI stability.
Flaky Test Repair is a specialized skill for Claude Code designed to eliminate intermittent test failures that plague modern CI/CD pipelines. Based on research demonstrating high success rates for LLM-driven test repair, this skill analyzes flaky test behaviors—including timing issues, race conditions, and shared state—to suggest or auto-apply robust code transformations. It bridges the gap between detecting unreliability and achieving deterministic, green builds by implementing industry best practices like time mocking, proper async handling, and test isolation, significantly reducing the manual effort required for test maintenance.
Características Principales
01Standardized repair reports with original code, suggested fixes, and change summaries
02Deterministic mocking for random values, UUIDs, and external network calls
03Automated application of fixes for common issues like Date.now() and missing awaits
04Pattern-based root cause analysis for timing, async, and environmental dependencies
05Integration with stability metrics to track and improve long-term CI reliability
0667 GitHub stars
Casos de Uso
01Repairing tests that fail intermittently in CI/CD pipelines without clear manual causes
02Automating the migration from real-time assertions to mocked system time for deterministic outcomes
03Refactoring legacy test suites that rely on hardcoded delays or shared global state
