Does this skill support specific programming languages?

Yes, it includes specialized debugging references for Python and TypeScript/JavaScript, along with a robust general framework applicable to any programming language.

Can I force the agent to perform a deeper investigation?

Yes, by using the --deep flag when invoking the skill, you can bypass the triage phase and move directly into a full-scale deep-track investigation.

How does it ensure the fix doesn't break other features?

The 'Fix & Verify' phase requires running existing tests, performing a quality check, and writing a new regression test specifically designed to catch that bug in the future.

What are the 'Quick' and 'Deep' tracks in the workflow?

The Quick track is for localized, obvious errors, while the Deep track is triggered for multi-file issues, intermittent bugs, or async problems, often using sub-agents for parallel investigation.

How does Bug Killer differ from standard AI code fixing?

Standard AI often attempts to fix code immediately; Bug Killer enforces a 5-phase professional workflow that requires reproduction, evidence gathering, and hypothesis testing before any modifications are made.

Bug Killer Debugger

Name: Bug Killer Debugger
Author: sequenzia

bysequenzia

•

セキュリティとテスト

Executes a systematic, hypothesis-driven debugging workflow to diagnose and resolve complex software errors through structured evidence gathering.

Bug Killer is a high-performance skill for Claude Code that transforms haphazard debugging into a disciplined, 5-phase engineering process. It utilizes a triage-based routing system to distinguish between simple fixes and deep-track investigations, ensuring that complex, multi-module bugs receive the rigorous analysis they require. By enforcing a hypothesis journal and requiring root cause confirmation before any code is modified, Bug Killer minimizes 'trial-and-error' coding and ensures every fix is backed by verifiable evidence and regression tests.

主な機能

01Automated regression test generation to prevent recurring issues

02Automated hypothesis journaling to track evidence for and against potential causes

0322 GitHub stars

04Triage-based routing for optimized quick-fix or deep-track investigation paths

05Language-specific debugging strategies for Python, TypeScript, and JavaScript

06Parallel multi-agent exploration for complex, distributed codebase issues

ユースケース

01Navigating and debugging complex data flows spanning multiple services or modules

02Providing a structured audit trail for critical production bug investigations

03Diagnosing intermittent or environment-specific failures that resist simple fixes

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sequenzia/agent-alchemy bug-killer

For use in Claude.ai and ChatGPT

Download Skill