What is included in the Multi-Layer Verification process?

Verification follows a pyramid: starting with static analysis and imports, moving to build success and unit tests, and culminating in critical runtime execution to confirm the fix works in a real environment.

When should I use the auto mode flag?

Use the -a or --auto flag when you want Claude to proceed through the debugging steps autonomously, implementing the most recommended solution without stopping for manual confirmation.

How does the 'Log Technique' help with hard-to-reproduce bugs?

When an error can't be reproduced immediately, the skill adds strategic debug logs to the codebase. The user runs the application, and the resulting console output is fed back into the skill to pinpoint the exact failure point.

What makes the Workflow Debug skill different from standard debugging?

Standard debugging often relies on guesswork. This skill enforces a hypothesis-driven methodology and a 'Verification Pyramid' that ensures the actual code path is executed, preventing false confidence from passing tests alone.

Workflow Debug

Name: Workflow Debug
Author: flosrn

byflosrn

0•

安全与测试

Systematically debugs complex software errors through hypothesis-driven analysis, solution comparisons, and multi-layer verification.

The Workflow Debug skill provides Claude with a rigorous, battle-tested framework for resolving software defects. Instead of jumping to quick fixes, it follows a structured 5-step methodology: reproducing the error, ranking hypotheses for the root cause, researching multiple solution paths with comparative pros and cons, implementing the chosen fix, and validating it through a 'Verification Pyramid' that includes runtime execution. This skill is particularly effective for elusive bugs that require strategic logging or complex issues where multiple architectural solutions need to be weighed before implementation.

主要功能

01Autonomous mode for rapid, hands-free debugging using recommended best practices

020 GitHub stars

03Strategic log instrumentation for debugging non-reproducible or production errors

04Multi-layer verification covering static analysis, build status, and runtime execution

05Hypothesis-driven root cause analysis with ranked probability testing

06Comparative solution research providing 2-3 fix options with detailed trade-offs

使用场景

01Debugging elusive runtime crashes by automatically adding and analyzing strategic debug logs

02Automating the reproduction and verification cycle for recurring software regressions

03Evaluating and implementing the best architectural fix for complex logic errors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add flosrn/.claude workflow-debug

For use in Claude.ai and ChatGPT