Diagnose bugs and performance regressions through a structured reproduce-hypothesise-fix workflow.
npx skills add https://github.com/unnxt30/skills --skill diagnose-unnxt30
Please help me install this Agent Skill. Skill: diagnose Source: https://github.com/unnxt30/skills/tree/main/diagnose Command: npx skills add https://github.com/unnxt30/skills --skill diagnose-unnxt30
[DEBUG-...] prefix so cleanup is a single grep, plus a post-mortem checklist covering regression tests and architectural findings. - Use Case: A user reports that an export button intermittently throws an error. The Skill guides building a deterministic repro loop (or a HITL bash script via scripts/hitl-loop.template.sh if human clicks are required), ranking hypotheses, instrumenting the code path, and landing a fix with a regression test. ## Quick Start Ask the assistant to diagnose the bug you are experiencing and describe the exact symptom, error message, or performance regression you observed.High-intent search queries and answers about installing and using this skill.
For non-deterministic bugs, raise the reproduction rate instead of chasing a clean repro: loop the trigger 100 times, parallelise, add stress, narrow timing windows, or inject sleeps. A 50%-flake bug is debuggable; keep increasing the rate until it is.
Use a human-in-the-loop bash script based on scripts/hitl-loop.template.sh. The script prompts the human through each step, captures answers as KEY=VALUE output, and feeds the results back so the debugging loop stays structured.
First build a fast deterministic feedback loop that reproduces the exact symptom, then generate 3-5 ranked falsifiable hypotheses with explicit predictions. Only instrument and fix after the loop reliably demonstrates the failure the user described.
Skip the regression test when no correct seam exists — meaning no test location exercises the real bug pattern as it occurs at the call site. A shallow test gives false confidence; instead document the missing seam as an architectural finding.
For performance regressions, avoid log-based probing. Establish a baseline measurement with a timing harness, profiler, or query plan, then bisect changes against that measurement. Measure first, fix second.
santosomarLocalize the exact file, function, or line introducing a bug from stack traces, tests, logs, and reports.
janpeterdLocate shallow modules and propose deepened interfaces for codebase architecture.
Productfculty-aipmDesign a continuous improvement flywheel for AI product quality.
Euda1mon1aReview code changes for bugs, security issues, and best practices.