What problem does it solve? When users report errors, crashes, or broken functionality, ad-hoc guessing wastes time and often introduces new defects. This Skill enforces a systematic debugging discipline: collect evidence first, form ranked hypotheses, fix one issue at a time, and verify with compile and regression checks. ## Core Features & Use Cases - Evidence-Based Diagnosis: Gathers full stack traces, reproduction steps, environment info, and recent changes before forming any conclusion. - Hypothesis-Driven Fixing: Maintains at most 3 ranked hypotheses with explicit verification methods, fixing one logical point at a time and rolling back on failure. - Verified Repair Workflow: Every fix passes TypeScript compilation checks (tsc --noEmit), functional verification, and regression testing, with a stop-and-reassess rule after 3 consecutive failures. - Use Case: A user reports "the login page throws a 500 error after the latest deploy." The Skill collects logs and recent commits, identifies the most likely cause, applies a single targeted fix, verifies compilation and regression, then suggests a commit with a fix: prefix. ## Quick Start Tell the agent "this feature is broken" or paste the error message and stack trace, and it will systematically diagnose the root cause and fix it.
