What problem does it solve? Random fixes and guess-and-check patching waste time, mask underlying issues, and introduce new bugs. This Skill enforces a disciplined root-cause-first debugging process so every fix addresses the actual problem instead of its symptoms. ## Core Features & Use Cases - Four-Phase Process: Guides investigation through root cause analysis, pattern comparison, hypothesis testing, and verified implementation, with mandatory checkpoints between phases. - Supporting Techniques: Includes reference guides for bisection (code and git bisect), root-cause tracing through call stacks, rubber duck debugging, defense-in-depth validation, time-boxed escalation, and post-fix documentation. - Agent Pitfall Guardrails: Catalogs common AI agent mistakes such as hallucinating root causes, overconfidence in first hypotheses, and proposing oversized refactors for simple bugs. - Use Case: A test suite fails after a dependency update. Instead of tweaking mocks repeatedly, the Skill walks you through reproducing the failure, bisecting recent commits with git bisect, forming a single testable hypothesis, and verifying the fix with a regression test. ## Quick Start Ask the AI to systematically debug the failing test or error you are seeing, starting with root cause investigation before proposing any fix.
