bug-fix

Diagnose and fix production bugs through reproduction, root cause analysis, and regression verification.

Updated Apr 13, 2026
One-click install
npx skills add https://github.com/scoots31/engineering-playbook --skill bug-fix-scoots31
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bug-fix
Source: https://github.com/scoots31/engineering-playbook/tree/main/skills/bug-fix
Command: npx skills add https://github.com/scoots31/engineering-playbook --skill bug-fix-scoots31

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Production bugs get fixed sloppily — symptoms patched instead of root causes, fixes bundled with unrelated changes, and regressions discovered only after deploy. This Skill enforces a disciplined five-step workflow so live-system issues are reproduced, isolated, minimally fixed, regression-checked, and safely deployed. ## Core Features & Use Cases - Reproduction First: Requires a written reproduction recipe before any code is touched, ensuring the bug is actually understood. - Root Cause Isolation: Classifies failures into logic errors, data assumptions, state problems, integration failures, or config drift before writing the fix. - Smallest-Fix Discipline: Prevents scope creep by forbidding refactoring or adjacent improvements during the fix, with a path to surface larger cleanups separately. - Regression & Deploy Verification: Walks adjacent code paths, checks edge cases, deploys the fix alone, and confirms the bug is gone in the live environment. - Use Case: Users report that checkout fails intermittently in production. Use this Skill to reproduce the failure, trace it to a stale-cache state problem, apply a two-line fix, verify the payment and cart flows still work, and deploy the isolated commit. ## Quick Start Use the bug-fix skill to reproduce and fix the production issue where checkout fails for returning customers.

Frequently Asked Questions about bug-fix

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

FAQPage Schema
How do I fix a production bug without breaking other features?

Reproduce the bug first, isolate the root cause, then make the smallest possible fix touching only the broken code. Before deploying, run existing tests on the affected path and manually walk any feature that shares code with the fix.

How to find the root cause of a bug instead of fixing the symptom?

Read the code where the symptom appears and ask why the code behaves that way. Classify the cause as a logic error, data assumption, state problem, integration failure, or config drift, then state the root cause explicitly before writing any fix.

What should I do when I cannot reproduce a reported bug?

Treat it as a report, not a confirmed bug. Document what was tried and ask one specific isolating question, such as which environment, account type, or data shape triggers the issue, before proceeding with any code changes.

Should I refactor code while fixing a production bug?

No. Refactoring during a fix expands the blast radius and makes regressions hard to isolate. Note the refactor opportunity, ship the minimal fix alone, and handle cleanup as a separate piece of work.

When should this bug-fix workflow not be used?

It is designed for production issues in live systems, not development-time errors caught before deploy. Pre-deploy issues should surface through QA and phase testing workflows instead.