What problem does it solve?
Prevents wasted time, hidden regressions, and repeated failures by forcing a disciplined workflow that isolates the true root cause before any code changes are proposed.
Core Features & Use Cases
- Four-phase debugging workflow: Root-cause investigation, pattern analysis, hypothesis testing, and root-cause implementation.
- Evidence-first diagnostics: Read and interpret error messages, reproduce reliably, review recent changes, gather data, and trace data flow across components.
- Decision guardrails: Explicitly stops fixes until root cause is understood, and escalates to architecture review after 3+ failed fix attempts.
- Use cases: Test failures, production bugs, unexpected behavior, performance issues, build/integration failures, and multi-component system issues (CI/build/deploy, API/service/db).
Quick Start
Use the skill to debug a failing test by reproducing it, reading the full error output, tracing the data flow to where the bad value originates, and only then planning the smallest single root-cause change.