fix

Diagnose software bugs and apply verification-driven fixes with parallel checks.

Updated Mar 7, 2026

Search Tags:#ci/cd #debugging #root cause analysis #bug fixing #test failures #type errors #regression verification

Ready to run

npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill fix-minhhoangdono

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: fix
Source: https://github.com/MinhHoangDono/antigravity-kit/tree/main/.agent/skills/fix
Command: npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill fix-minhhoangdono

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Fixes bugs, errors, failing tests, CI/CD issues, and other software breakages by diagnosing root causes and applying a safe, verification-driven repair.

Core Features & Use Cases

Intelligent repair routing: Chooses a quick, standard, deep, or parallel workflow based on issue complexity and whether multiple independent issues are present.
Debug-first consistency: Always activates the debug skill and produces a findings report before implementing any changes.
Verification and regression prevention: Runs parallel verification steps to confirm tests pass and that the fix does not introduce regressions.
Mandatory final accountability: Reports confidence, changes, and files affected, updates docs when warranted, and offers a commit.

Quick Start

Use the fix skill to resolve an error by telling the AI: "Fix the failing CI test for issue #1234 using quick mode."

Frequently Asked Questions about fix

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

FAQPage Schema

How do I fix failing CI/CD tests and runtime errors without introducing regressions?▼

To fix failing CI/CD tests and runtime errors safely, activate a debug-first workflow to generate root-cause findings, implement the code changes, and then run parallel verification checks to prevent regressions before providing a final confidence report.

What is the best way to diagnose root causes for type errors and lint issues in a multi-file codebase?▼

Diagnosing root causes for type errors and lint issues requires selecting a complexity-aware routing workflow that analyzes the codebase, produces a findings report, and implements fixes while verifying that no other files are negatively affected.

Can I resolve multiple independent test failures at the same time?▼

Yes, you can resolve multiple independent test failures simultaneously by using a parallel workflow mode that handles separate issues concurrently and verifies all fixes together to ensure overall codebase stability.

Does this bug fixing approach work for both single-file and multi-file codebases?▼

Yes, this bug fixing approach applies to both single-file and multi-file codebases, routing repair workflows based on issue complexity to ensure accurate root cause analysis and verification regardless of project scale.

Why should I run a debug analysis before implementing a bug fix?▼

Running a debug analysis before implementing a bug fix ensures you identify the actual root cause rather than treating symptoms, producing a findings report that guides complexity-aware changes and mandatory final accountability.

What is included in the final report after repairing a software bug?▼

The final report after repairing a software bug includes a confidence score, a list of affected files, the changes made, updated documentation when warranted, and an offer to commit the verified fix.