debugging-and-error-recovery

Diagnose test failures, build errors, and runtime bugs through structured root-cause triage.

9|Updated Jul 8, 2026
One-click install
npx skills add https://github.com/jerrylin96/dotgemini --skill debugging-and-error-recovery-jerrylin96
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debugging-and-error-recovery
Source: https://github.com/jerrylin96/dotgemini/tree/main/skills/debugging-and-error-recovery
Command: npx skills add https://github.com/jerrylin96/dotgemini --skill debugging-and-error-recovery-jerrylin96

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When tests fail, builds break, or runtime behavior diverges from expectations, developers often guess at fixes or patch symptoms instead of causes. This Skill enforces a disciplined stop-the-line workflow that preserves evidence, localizes the failure, and fixes the actual root cause. ## Core Features & Use Cases - Stop-the-Line Rule: Halts feature work when something breaks, preserving error output, logs, and reproduction steps before any changes continue. - Structured Triage Checklist: Walks through reproduce, localize, reduce, and fix steps in order so no diagnostic stage is skipped. - Subagent Delegation: Offloads log reading, diagnostic shell commands, and repro script writing to subagents to avoid context window bloat. - Use Case: A test suite fails after a refactor. Instead of tweaking assertions, you follow the checklist to reproduce the failure, isolate it to a faulty SQL JOIN, and fix the query rather than filtering duplicates in UI code. ## Quick Start Ask the agent to debug the failing test or broken build using the systematic triage checklist and find the root cause.

Frequently Asked Questions about debugging-and-error-recovery

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

FAQPage Schema
How do I debug a failing test systematically?

Follow the triage checklist in order: reproduce the failure reliably, localize where it occurs, reduce it to a minimal failing case, then fix the root cause. Do not skip steps or patch symptoms before confirming the underlying issue.

What is root cause debugging versus fixing symptoms?

Root cause debugging fixes the underlying defect rather than its visible effect. For example, if database queries return duplicates, fix the SQL JOIN generating them instead of filtering duplicates in UI code.

How do I create a minimal reproduction for a bug?

Remove unrelated code and configuration until only the bug remains, simplify the input to the smallest triggering example, and strip the test to the bare minimum that reproduces the issue. A minimal case makes the root cause obvious.

When should I stop feature work to fix a bug?

Stop immediately when anything unexpected happens, such as a failing test or broken build. Preserve evidence first, then diagnose and fix before resuming, because errors compound and later work built on a bug becomes invalid.

Can subagents help with debugging large logs?

Yes. Delegate reproduction, localization, and reduction steps to a subagent to run diagnostic shell commands, read long logs, and write repro scripts. This keeps the main context window free of bulky diagnostic output.