What problem does it solve? When a bug, failed test, regression, or flaky behavior appears, it is tempting to patch the symptom or guess at fixes. This Skill enforces a disciplined root-cause workflow so failures are reproduced, traced, and fixed at the source instead of being masked by retries or fallbacks. ## Core Features & Use Cases - Structured debugging workflow: Reproduce the failure, capture evidence, trace backward to the first incorrect assumption, test one hypothesis at a time, and apply the smallest root-cause fix. - Boundary tracing for integrations: Record values at each layer of a stack (UI, API routes, agent runtime, domain tools, external services like Supabase, Stripe, or Maps) to localize which boundary failed. - Flaky test and pollution tooling: Use condition-based waiting patterns instead of arbitrary sleeps, defense-in-depth validation across layers, and a bisection script that finds which test pollutes shared state. - Use Case: A Playwright test fails one run in ten. Instead of adding a sleep or rerunning until green, the Skill guides you to wait on actual conditions, isolate state pollution with the find-polluter script, and lock the fix with a regression test. ## Quick Start Use the systematic-debugging skill to find the root cause of this failing integration test before changing any code.
