debug

Diagnose and fix bugs through a phased reproduce-test-fix-verify pipeline with hypothesis investigation mode.

2|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/DoctorMozg/claude-pipelines --skill debug-doctormozg
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debug
Source: https://github.com/DoctorMozg/claude-pipelines/tree/main/plugins/mz-dev-pipe/skills/debug
Command: npx skills add https://github.com/DoctorMozg/claude-pipelines --skill debug-doctormozg

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Bug reports arrive in many forms — stack traces, failing tests, vague descriptions, or unverified suspicions — and ad-hoc fixing often skips reproduction, regression tests, or root-cause analysis. This Skill orchestrates a disciplined debugging pipeline that reproduces the issue, diagnoses the root cause, writes a failing regression test, applies a minimal fix, and verifies it with review. ## Core Features & Use Cases - Two certainty modes: certainty:high runs a full TDD bug-fix pipeline (reproduce → diagnose → approval gate → regression test → fix → verify → review); certainty:low runs a read-only hypothesis investigation that proves or disproves a suspicion with exploratory tests and a verdict report. - Flexible input handling: accepts free-text bug reports, stack traces, error messages, failing test names, and GitHub issue URLs, with untrusted-content isolation for external input. - Domain research integration: conditionally dispatches web research agents when bugs involve external APIs, protocols, or library behavior. - Use Case: A user reports "the WebSocket reconnection fails on timeout" — the pipeline reproduces the failure, diagnoses the root cause with file:line precision, gets user approval, writes a failing regression test, applies a minimal fix, and delivers a verified report. ## Quick Start Ask the assistant to debug the failing payment test and fix the root cause with a regression test.

Frequently Asked Questions about debug

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I debug a failing test with an automated pipeline?

Invoke the debug skill with the failing test name or a bug description. It reproduces the failure, diagnoses the root cause, writes a regression test that fails against current code, applies a minimal fix, and verifies the full suite passes.

How to verify a hypothesis about code behavior without changing code?

Use the debug skill with certainty:low and your hypothesis. It runs code analysis, optional domain research, and exploratory tests, then produces a read-only verdict report with confidence level and evidence — no code is modified.

Can I debug a bug from a GitHub issue URL?

Yes. Paste the GitHub issue URL as input; the skill fetches the issue via gh CLI with MCP and REST API fallbacks, treats the content as untrusted data, and classifies it to pick the right reproduction strategy.

What happens when a bug cannot be reproduced?

The pipeline stops and asks you for more context via an interactive prompt, reporting what was tried and found. It never proceeds with guesswork or speculative fixes without a confirmed reproduction or static confirmation.

When should I not use an automated debug pipeline?

Avoid it for building new features, polishing known-good code, or pure code explanation with no testable claim. Those cases are better served by build, polish, or explain workflows respectively.