How do I invoke the debugger?

Use the command $gsd-debug followed by a description of your issue to start the symptom gathering process.

Where is the debugging state stored?

All session data, including hypotheses and evidence, is stored in .planning/debug/*.md files, allowing for persistent state across Claude Code resets.

Can I resume a previous debugging session?

Yes. If you invoke $gsd-debug without arguments, the skill will list active sessions from the .planning directory and allow you to pick up where you left off.

Why does this skill use subagents for debugging?

Investigation burns context fast when reading files and forming hypotheses. Subagents provide a fresh 200k context per investigation, keeping the main conversation lean.

GSD Systematic Debugger

Name: GSD Systematic Debugger
Author: rapchai-k

byrapchai-k

0•

安全与测试

Automates the debugging process using the scientific method through isolated subagents and persistent state tracking.

The GSD Debugger skill implements a rigorous, systematic approach to troubleshooting software issues within Claude Code. By spawning specialized subagents in fresh context windows, it prevents context bloat during complex investigations while maintaining persistent state in dedicated markdown files. This skill guides users through structured symptom gathering, hypothesis testing, and root cause analysis, ensuring that fixes are verified and documented rather than guessed. It is particularly effective for deep-dive investigations where reading many files would otherwise exhaust the primary conversation context.

主要功能

01Structured symptom gathering (expected vs. actual behavior, timelines)

02Persistent state management via .planning/debug markdown files

03Scientific method-driven hypothesis testing and root cause identification

04Checkpoint-based workflows for human-in-the-loop verification

050 GitHub stars

06Isolated subagent spawning for clean 200k context investigations

使用场景

01Standardizing the debugging workflow for teams using the GSD methodology

02Troubleshooting complex race conditions that require reading extensive log files

03Managing multi-step debugging sessions that span across context resets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rapchai-k/the-third-place gsd-debug

For use in Claude.ai and ChatGPT

主要功能

01Structured symptom gathering (expected vs. actual behavior, timelines)

02Persistent state management via .planning/debug markdown files

03Scientific method-driven hypothesis testing and root cause identification

04Checkpoint-based workflows for human-in-the-loop verification

050 GitHub stars

06Isolated subagent spawning for clean 200k context investigations

使用场景

01Standardizing the debugging workflow for teams using the GSD methodology

02Troubleshooting complex race conditions that require reading extensive log files

03Managing multi-step debugging sessions that span across context resets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rapchai-k/the-third-place gsd-debug

For use in Claude.ai and ChatGPT