Can this skill help with flaking tests?

Yes, it includes specific strategies like bisection to identify 'polluter' tests that leave behind state or files that cause subsequent tests to fail intermittently.

When is the best time to trigger this skill?

Use it whenever you encounter a bug with a long call stack, or when the value causing a failure was generated in a different part of the system than where the crash occurred.

How does this differ from standard AI debugging?

While standard debugging often addresses the immediate error message, Root Cause Tracing enforces a methodology that traces the logic backward until it finds the initial trigger, ensuring the fix is permanent and robust.

What kind of instrumentation does it add?

It typically adds temporary stack trace logging and state inspection (using console.error and Error().stack) to visualize the execution path without relying on complex external debuggers.

Root Cause Tracing

Name: Root Cause Tracing
Author: samjhecht

bysamjhecht

•

セキュリティとテスト

Systematically identifies the original source of software bugs by tracing execution flow backwards through the call stack.

Root Cause Tracing is a specialized debugging skill for Claude Code designed to prevent 'band-aid' fixes. Instead of patching symptoms where they appear, this skill guides Claude to navigate deep call stacks, implement strategic instrumentation, and isolate the exact moment invalid data or state was introduced. It includes methodologies for using stack traces, environment logging, and bisection scripts to solve complex issues like test pollution and configuration errors, ultimately resulting in more robust, defense-in-depth solutions.

主な機能

01Defense-in-depth validation pattern implementation

021 GitHub stars

03Automated instrumentation and stack trace logging for deep visibility

04Context-aware analysis of environment variables and directory resolution

05Bisection logic to identify specific tests causing state pollution

06Systematic backward call chain tracing to identify original triggers

ユースケース

01Locating specific tests that are polluting the file system or global state

02Debugging errors that occur deep in execution far from the entry point

03Identifying the source of 'undefined' or invalid data passed through multiple layers

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add samjhecht/wrangler root-cause-tracing

For use in Claude.ai and ChatGPT

Download Skill