How does the skill track progress during a debugging session?

It uses a structured phase-based workflow (Collect, Isolate, Formulate, Test, Verify) tracked via todo tasks to ensure a logical and documented progression toward the fix.

Does it automatically create tests for the bugs it finds?

The framework requires the creation of a failing test case that reproduces the bug before applying the fix, ensuring the solution is verified and prevents future regressions.

Does this skill help with intermittent or flaky tests?

Yes, it provides specific strategies for investigating timing-dependent failures and race conditions by analyzing environment differences, concurrency, and shared state.

What is the 'Iron Law' of this debugging skill?

The Iron Law mandates that no code fixes should be proposed or implemented until a definitive root cause has been identified through systematic investigation and evidence gathering.

Can this skill handle performance issues like memory leaks?

Yes, it includes specialized playbooks for performance debugging that focus on profiling, resource usage metrics, and algorithm complexity analysis to find bottlenecks.

Systematic Debugging

Name: Systematic Debugging
Author: outfitter-dev

byoutfitter-dev

0•

セキュリティとテスト

Implements a rigorous four-phase investigation framework to identify and resolve software defects through evidence-based analysis and root-cause verification.

Systematic Debugging provides a structured methodology for Claude to diagnose and fix complex software issues, ranging from runtime crashes to intermittent race conditions. By enforcing the 'iron law' of root-cause investigation before proposing any code changes, this skill ensures that fixes are based on data rather than guesswork. It guides the agent through evidence collection, variable isolation, and hypothesis testing, resulting in permanent solutions and robust test coverage for runtime, logic, integration, and performance bugs.

主な機能

01Rigorous four-phase debugging framework: Evidence, Isolation, Hypothesis, and Verification

02Evidence-based investigation enforcing root-cause identification before any code changes

030 GitHub stars

04Specialized playbooks for runtime, logic, integration, performance, and intermittent failures

05Mandatory verification via failing test cases to prevent regression

06Structured progress tracking using a phase-based todo workflow

ユースケース

01Troubleshooting integration failures between APIs, databases, and external services

02Diagnosing and fixing intermittent race conditions or timing-dependent failures

03Identifying the root cause of complex logic errors or incorrect data transformations

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add outfitter-dev/agents debugging

For use in Claude.ai and ChatGPT

Download Skill