What problem does it solve? Bug fixes without a root cause become band-aids that recur months later under a different name. This Skill enforces a disciplined debugging workflow so every fix is backed by a reproducible failing test and a documented root cause. ## Core Features & Use Cases - Four-phase methodology: Investigate (gather evidence), Analyze (narrow with 5 Whys), Hypothesize (testable theory), Implement (failing test first, then fix). - Iron Law enforcement: No fix is accepted without both an identified root cause and a failing test that reproduces the bug, with the test and fix committed separately. - Diagnostic aids: Pattern lookup table for common bug classes (race conditions, nil propagation, state corruption), a 3-strike rule for escalating to structural review, and a blast-radius gate for wide fixes. - Use Case: A production API returns intermittent 500 errors. You reproduce it locally, narrow it via commit bisection and 5 Whys to a missing seed in a migration, write a failing test, commit it, then implement and commit the fix with a root-cause section in the PR body. ## Quick Start Debug this failing checkout test by finding the root cause and writing a reproducing test before any fix.
