diagnose

Guide structured 6-phase debugging for elusive bugs and performance regressions.

Updated Jul 22, 2025
One-click install
npx skills add https://github.com/MichaelSolati/mkai --skill diagnose-michaelsolati
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/MichaelSolati/mkai/tree/main/profiles/engineering/skills/diagnose
Command: npx skills add https://github.com/MichaelSolati/mkai --skill diagnose-michaelsolati

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Unstructured, ad-hoc debugging of hard bugs and performance regressions wastes developer time, leads to unreliable fixes, and often fails to address the root cause of elusive, intermittent issues.

Core Features & Use Cases

  • Structured 6-Phase Diagnosis Loop: Guides users through building a deterministic feedback loop, reproducing the bug, generating falsifiable hypotheses, targeted instrumentation, implementing fixes with regression tests, and post-mortem cleanup.
  • Non-Deterministic Bug Handling: Provides specific tactics to increase reproduction rates for flaky bugs that only appear intermittently under narrow conditions.
  • Use Case: When a developer encounters a performance regression that only appears under high load, this skill guides them to build a timing harness, generate hypotheses about database query bottlenecks, instrument the code to measure query times, and implement a fix with a regression test to prevent the issue from recurring.

Quick Start

Use the diagnose skill to debug the reported performance regression in the checkout flow by following the structured 6-phase diagnosis loop.

Frequently Asked Questions about diagnose

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

FAQPage Schema
How do I debug flaky tests and intermittent non-deterministic bugs?

To debug flaky tests and non-deterministic bugs, you need specific tactics to increase reproduction rates under narrow conditions. A structured workflow helps build deterministic reproduction loops to reliably trigger the elusive behavior.

What is the best way to find the root cause of a performance regression?

Finding the root cause of a performance regression requires targeted instrumentation to measure bottlenecks, such as database query times. Generating falsifiable hypotheses and building a timing harness ensures you identify the actual source.

How do I stop fixing the same bug repeatedly and prevent future regressions?

To stop repeatedly fixing the same bug and prevent future regressions, implement fixes alongside dedicated regression tests. Conducting post-mortem analysis after the fix eliminates unreliable patches and addresses the root cause.

Does unstructured debugging fail to resolve hard production issues?

Unstructured debugging fails to resolve hard production issues because it wastes developer time and produces unreliable fixes. Applying a 6-phase diagnosis loop replaces ad-hoc guessing with falsifiable hypotheses and targeted instrumentation.

Can I use a structured diagnosis workflow for high load performance issues?

You can use a structured diagnosis workflow for high load performance issues by building a timing harness. It guides you to generate hypotheses about database query bottlenecks, instrument the code, and implement a verified fix.