How do I install the Debugging skill?

Run npx skillmds@latest add hi-donwi/debugging in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Debugging safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Debugging?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Debugging free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Debugging?

hi-donwi (@hi-donwi) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Debugging by hi-donwi — Agent Skill

Name: Debugging
Rating: 4.2 (5 reviews)
Author: hi-donwi

Debugging

Systematically diagnose and fix failures instead of guessing. Use when facing a bug, stack trace, failing test, crash, or unexpected behavior, or when a fix attempt did not work. Do not use for greenfield design decisions with no defect to chase.

hi-donwi Updated Sep 22, 2026

File contents

Debugging & Error Recovery

Overview

Find the root cause before changing code. Random edits hide symptoms and create new bugs.

When to use

Any reproducible (or intermittent) defect, crash, or failing check.
After a fix "didn't work" — reset and re-diagnose rather than pile on changes.

Process

Reproduce reliably — get a minimal, deterministic repro. If flaky, find what varies.
Read the actual error — full message + stack trace; identify the first failing frame in your code, not the framework.
Form one hypothesis about the root cause and a cheap way to test it (log, breakpoint, bisect, or a failing test via test-driven-development).
Test the hypothesis — confirm or reject before editing. Change one variable at a time.
Fix the cause, not the symptom. Add a regression test that fails without the fix.
Verify and clean up — remove debug logging; confirm the original repro is gone and nothing nearby regressed.
If the first fix failed — do not stack patches. Revert to diagnosis; use references/playbook.md for bisect, isolation, and domain-specific recovery loops.

Red flags

Editing code before reproducing or reading the error.
Multiple simultaneous changes so you can't tell what fixed it.
"Fixes" that suppress the error (swallowed exceptions, broad try/catch) without understanding.
No regression test after the fix.

Verification

The original repro now passes; a new test guards against regression.
You can explain the root cause in one sentence.

Reference Index

references/playbook.md — flexible recovery when fixes fail, bisect tactics, scroll-video issues.
references/expanded-guidance.md — deeper triage checklist, localization tactics, and recurrence prevention.

---name: debuggingdescription: Systematically diagnose and fix failures instead of guessing. Use when facing a bug, stack trace, failing test, crash, or unexpected behavior, or when a fix attempt did not work. Do not use for greenfield design decisions with no defect to chase.---# Debugging & Error Recovery## OverviewFind the root cause before changing code. Random edits hide symptoms and create new bugs.## When to use- Any reproducible (or intermittent) defect, crash, or failing check.- After a fix "didn't work" — reset and re-diagnose rather than pile on changes.## Process1. **Reproduce reliably** — get a minimal, deterministic repro. If flaky, find what varies.2. **Read the actual error** — full message + stack trace; identify the first failing frame in your code, not the framework.3. **Form one hypothesis** about the root cause and a cheap way to test it (log, breakpoint, bisect, or a failing test via `test-driven-development`).4. **Test the hypothesis** — confirm or reject before editing. Change one variable at a time.5. **Fix the cause, not the symptom.** Add a regression test that fails without the fix.6. **Verify and clean up** — remove debug logging; confirm the original repro is gone and nothing nearby regressed.7. **If the first fix failed** — do not stack patches. Revert to diagnosis; use `references/playbook.md` for bisect, isolation, and domain-specific recovery loops.## Red flags- Editing code before reproducing or reading the error.- Multiple simultaneous changes so you can't tell what fixed it.- "Fixes" that suppress the error (swallowed exceptions, broad try/catch) without understanding.- No regression test after the fix.## Verification- The original repro now passes; a new test guards against regression.- You can explain the root cause in one sentence.## Reference Index- `references/playbook.md` — flexible recovery when fixes fail, bisect tactics, scroll-video issues.- `references/expanded-guidance.md` — deeper triage checklist, localization tactics, and recurrence prevention.

Frequently asked questions