What problem does it solve?
This Skill eliminates guesswork when tests or runtime behavior fail by enforcing a reproduce → investigate → prove discipline that finds, documents, and verifies root causes with concrete evidence.
Core Features & Use Cases
- Reproduce reliably: run the failing test/command multiple times to distinguish deterministic failures from flaky ones.
- Investigation techniques: apply stack trace analysis, git bisect (binary search), state inspection, minimal reproduction, and dependency isolation with a strict 5-minute-per-technique timeout.
- Prove and document: require file:line plus actual vs expected values, perform at least three "why" steps, log findings to .godmode/debug-findings.tsv, and hand off non-trivial fixes to /godmode:fix.
- Failure safety: rules for skipping stuck bugs, rollback on catastrophic failure, and classification of skipped bugs for future lessons.
Quick Start
Run the debug loop on the failing test command, reproduce the failure three times, pick an investigation technique (stack trace, git bisect, or state inspection), prove the root cause with file:line and actual vs expected values, and either apply a one-line fix or hand off to /godmode:fix.