debug

Diagnose failing tests, builds, and crashes to identify root causes without editing code.

Updated Mar 2, 2026
One-click install
npx skills add https://github.com/LeroyAdonis/kitfix-2.0 --skill debug-leroyadonis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: debug
Source: https://github.com/LeroyAdonis/kitfix-2.0/tree/main/.agents/skills/debug
Command: npx skills add https://github.com/LeroyAdonis/kitfix-2.0 --skill debug-leroyadonis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a test fails, a build breaks, or an app crashes, developers often jump straight to patching code without understanding the real cause. This Skill enforces a disciplined diagnosis-first workflow: reproduce the symptom safely, localize the failure, test competing hypotheses, and only then hand off a confirmed root cause for repair. ## Core Features & Use Cases - Safe Reproduction: Runs the smallest existing command to reproduce the symptom without installing dependencies, editing code, or mutating project state. - Root Cause Analysis: Traces stack frames, failing tests, git diffs, and data flow to separate facts from hypotheses, delivering a Confirmed, Likely, or Blocked verdict. - Structured Handoff: Produces a concise debug report with evidence and routes the diagnosis to /fix, /implement, or /feature as the next action. - Use Case: A developer runs "/debug npm test fails in cart-total.test.js" and receives a reproduced failure, the responsible code path, supporting evidence, and a ready-to-use /fix command. ## Quick Start Ask the AI to use the debug skill to investigate why your failing test or broken build is happening and report the root cause without changing any code.

Frequently Asked Questions about debug

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find the root cause of a failing test?▼

Run the smallest failing test to reproduce the symptom, capture the exact error and stack trace, then trace from the first relevant application frame to the responsible code. Test competing hypotheses with read-only commands before concluding a cause.

How to debug a broken build without changing code?▼

Reproduce the build failure with the documented command, capture the exact error output, and inspect configuration, imports, and recent git changes. Use read-only inspection, existing logs, and git diff or blame rather than adding probes or editing files.

Can I diagnose a bug without modifying source code?▼

Yes. Diagnosis relies on existing tests, stack traces, logs, git history, and one-off read-only commands. Editing code to add logs or probes is avoided so the working tree and evidence stay intact.

What should I do when a bug cannot be reproduced?▼

Record what was attempted and what evidence is missing, then continue with static investigation only if it yields a clearly labeled hypothesis. Report a Blocked verdict and recommend the next diagnostic evidence rather than a speculative repair.

When should I debug before fixing a regression?▼

Always diagnose first when the cause is unclear. Use git diff, log, and blame to identify when the regression appeared, confirm the trigger and responsible code, then hand off to a fix workflow with the confirmed cause.