diagnose

Diagnose software bugs and performance regressions through a six-phase methodology.

Updated May 5, 2026

Search Tags:#testing #software engineering #performance #debugging #troubleshooting #root cause analysis

Ready to run

npx skills add https://github.com/wachawo/claude-skills --skill diagnose-wachawo

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/wachawo/claude-skills/tree/main/skills/diagnose
Command: npx skills add https://github.com/wachawo/claude-skills --skill diagnose-wachawo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This skill provides a rigorous, repeatable framework for isolating and fixing complex software bugs and performance regressions, preventing aimless troubleshooting.

Core Features & Use Cases

Feedback Loop Construction: Prioritizes building deterministic, agent-runnable signals (tests, scripts, harnesses) to reliably reproduce issues.
Structured Debugging: Guides the user through a six-phase process from reproduction and hypothesis formulation to instrumentation and regression testing.
Use Case: When a critical production bug is reported, use this skill to systematically build a reproduction harness, test ranked hypotheses, and verify the fix with a regression test.

Quick Start

Use the diagnose skill to help me debug the intermittent crash occurring in the user authentication module.

Frequently Asked Questions about diagnose

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

FAQPage Schema

What is a structured approach to debugging complex software bugs?▼

A structured debugging approach uses a six-phase methodology to systematically reproduce issues, formulate hypotheses, instrument code, and verify fixes. This prevents aimless troubleshooting by prioritizing deterministic, runnable signals to reliably isolate root causes.

How do I build a feedback loop to reproduce an intermittent software crash?▼

To build a feedback loop for an intermittent crash, construct deterministic, agent-runnable scripts and test harnesses. This creates reliable signals that consistently reproduce the issue, allowing you to test ranked hypotheses and isolate the exact trigger.

What's the best way to troubleshoot performance regressions in software engineering?▼

The best way to troubleshoot performance regressions is applying a loop-based methodology that builds reproduction harnesses, tests ranked hypotheses through instrumentation, and verifies the resolution using regression tests to ensure systematic isolation.

Do I need a terminal environment to execute reproduction scripts for root cause analysis?▼

Yes, you need a terminal environment to execute reproduction scripts and manage debugging instrumentation during root cause analysis. This environment runs the deterministic harnesses required to systematically test hypotheses and verify fixes.

When should I use a systematic debugging methodology instead of ad-hoc troubleshooting?▼

Use a systematic debugging methodology when facing complex software bugs or critical production issues that resist ad-hoc troubleshooting. It provides a rigorous, repeatable framework for isolating root causes and verifying fixes through regression testing.

Can I verify software bug fixes using regression testing after isolating the root cause?▼

Yes, you can verify software bug fixes using regression testing as the final phase of the debugging loop. After reproducing the issue and testing hypotheses, regression tests confirm the fix resolves the bug without introducing new defects.