diagnose

Convert bug symptoms into a deterministic pass/fail feedback loop.

1|Updated May 5, 2026
One-click install
npx skills add https://github.com/yashs33244/my-mac-claude --skill diagnose-yashs33244
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/yashs33244/my-mac-claude/tree/main/skills/mattpocock/engineering/diagnose
Command: npx skills add https://github.com/yashs33244/my-mac-claude --skill diagnose-yashs33244

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps you systematically diagnose hard bugs and performance regressions by building a fast feedback loop that turns symptoms into actionable evidence.

Core Features & Use Cases

  • Build a measurable feedback loop using failing tests, HTTP/CLI repros, headless browser scripts, trace replay, or minimal harnesses so debugging is grounded in a deterministic signal.
  • Reproduce with correctness checks to ensure you’re observing the exact user-reported failure, and that it’s consistently repeatable (or repeatable at a high-enough rate for flaky issues).
  • Generate falsifiable hypotheses, then instrument precisely by changing one variable at a time, tagging logs for cleanup, and using profiling/bisection for perf regressions.
  • Fix with regression confidence by writing the regression test (when a correct seam exists) before implementing the fix, then re-running the original loop to verify the full scenario.

Quick Start

Ask an AI to diagnose your issue by first constructing a tight repro loop, then running a reproduce → hypothesis → targeted instrumentation → fix → regression-test workflow until the original scenario no longer fails.

Frequently Asked Questions about diagnose

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

FAQPage Schema
How do I debug hard bugs and performance regressions systematically?

Debug hard bugs by constructing a deterministic reproduction loop, generating falsifiable hypotheses, and instrumenting with minimal variable changes to convert symptoms into actionable pass/fail evidence.

What's the best way to reproduce flaky bugs for consistent debugging?

Reproduce flaky bugs by building a measurable feedback loop using trace replay or minimal harnesses, ensuring the failure repeats at a high enough rate to generate a reliable pass/fail signal.

How do I fix a performance regression using bisection and profiling?

Fix performance regressions by running profiling and bisecting to isolate the offending change, then validating the fix by re-running the original reproduction loop to confirm the scenario no longer fails.

Can I use this disciplined debugging loop for wrong outputs and thrown errors?

Yes, the disciplined debugging loop applies to broken behavior, thrown errors, and wrong outputs in real codebases by converting those reported symptoms into a reliable pass/fail feedback loop.

Do I need to write a regression test before implementing a bug fix?

Yes, write the regression test before implementing the fix when a correct seam exists, then re-run the original reproduction loop to verify the full scenario no longer fails.