ai-debuglisted

Diagnoses broken behavior systematically with a 4-phase root-cause loop: test failures, runtime errors, crashes, regressions. Never patches symptoms. Trigger for 'it is not working', 'something broke', 'this used to work', 'I am getting an error', 'CI is failing', 'why is X happening'. Not for adding tests; use /ai-test instead. Not for security findings; use /ai-security instead.
arcasilesgroup/ai-engineering · ★ 49 · AI & Automation · score 84

Install: claude install-skill arcasilesgroup/ai-engineering

# Debug ## Purpose Systematic debugging skill. Four phases, always in order. NEVER fix symptoms -- always find and fix the root cause. After 2 failed fix attempts, escalate to the user. ## When to Use - Test failures (expected vs actual mismatch) - Runtime errors (exceptions, crashes, hangs) - Regressions (worked before, broken now) - Unexpected behavior (no error, but wrong result) ## Process Step 0 (load contexts): read `.ai-engineering/manifest.yml` `providers.stacks`; load `.ai-engineering/overrides/<stack>/conventions.md` for each stack and `.ai-engineering/overrides/_shared/conventions.md`; load `.ai-engineering/team/*.md` for team conventions. ### Phase 1: Symptom Analysis (WHAT, WHEN, WHERE) Gather facts before forming hypotheses: 1. **WHAT**: exact error message, stack trace, log output 2. **WHEN**: always? intermittent? after a specific change? under load? 3. **WHERE**: which file, function, line? which test? which environment? 4. **SINCE WHEN**: `git log --oneline -20` -- what changed recently? Output: symptom report with all facts classified as KNOWN or SUSPECTED. ### Phase 2: Reproduction (MINIMAL REPRO) Make the bug reproducible with the smallest possible case: 1. Run the failing test or reproduce the error 2. If not reproducible: document exact conditions and STOP (cannot debug what cannot be reproduced) 3. Strip to minimal repro: remove unrelated code, simplify inputs, isolate the component 4. Confirm: the minimal repro fails consistently Output: