How does it handle multi-component systems?

It uses diagnostic instrumentation at component boundaries to log data entry/exit and verify environment configuration at every layer, identifying exactly where the chain breaks.

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

The Iron Law states that no fixes can be proposed or attempted until a complete root cause investigation is finished. Symptom-only fixes are considered a failure of the process.

What happens if multiple fixes fail to solve the problem?

If three or more fixes fail, the skill mandates a stop to question the underlying architecture. This prevents 'thrashing' and encourages a discussion on whether the implementation pattern itself is fundamentally flawed.

When should I use the Systematic Debugging skill?

Use it for any technical issue including test failures, performance bottlenecks, or unexpected behavior, especially when under time pressure or after previous fix attempts have failed.

Does this skill require writing tests?

Yes, Phase 4 requires creating a simple, automated failing test case that reproduces the bug before the fix is implemented to ensure the solution is verified and doesn't regress.

Systematic Debugging

Name: Systematic Debugging
Author: tianyuanztm

bytianyuanztm

0•

セキュリティとテスト

Enforces a rigorous, four-phase root cause analysis process to eliminate guesswork and prevent recurring bugs during development.

The Systematic Debugging skill transforms how Claude handles technical issues by mandating a strict "Root Cause First" approach. It prevents the common pitfall of applying superficial patches by guiding the AI through structured evidence gathering, pattern analysis, and minimal hypothesis testing. This skill is particularly effective for complex multi-component systems, elusive race conditions, and architectural failures where "one quick fix" often leads to regression. By following these phases, developers can reduce debugging time from hours of thrashing to minutes of precise, verified resolution.

主な機能

01Backward data flow tracing to identify the original source of corrupted values

020 GitHub stars

03Integration with TDD by requiring automated failing test cases before any code modification

04Architectural 'Stop Loss' logic that triggers a strategy review after three failed fix attempts

05Multi-component diagnostic instrumentation for tracing errors across system boundaries

06Mandatory four-phase workflow: Root Cause, Pattern Analysis, Hypothesis Testing, and Implementation

ユースケース

01Debugging intermittent test failures or production bugs with no obvious origin

02Breaking out of 'fix-and-break' cycles where new patches introduce regressions

03Resolving issues in complex distributed systems with multiple integration points

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tianyuanztm/myplan systematic-debugging

For use in Claude.ai and ChatGPT

主な機能

01Backward data flow tracing to identify the original source of corrupted values

020 GitHub stars

03Integration with TDD by requiring automated failing test cases before any code modification

04Architectural 'Stop Loss' logic that triggers a strategy review after three failed fix attempts

05Multi-component diagnostic instrumentation for tracing errors across system boundaries

06Mandatory four-phase workflow: Root Cause, Pattern Analysis, Hypothesis Testing, and Implementation

ユースケース

01Debugging intermittent test failures or production bugs with no obvious origin

02Breaking out of 'fix-and-break' cycles where new patches introduce regressions

03Resolving issues in complex distributed systems with multiple integration points