diagnose

Diagnose bugs through a reproducible loop of reproduction, hypothesis, instrumentation, fix, and regression testing.

138|6|Updated Jan 25, 2025
One-click install
npx skills add https://github.com/yamcodes/arkenv --skill diagnose-yamcodes
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/yamcodes/arkenv/tree/main/skills/diagnose
Command: npx skills add https://github.com/yamcodes/arkenv --skill diagnose-yamcodes

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Diagnoses hard bugs and performance regressions by enforcing a disciplined, repeatable diagnostic loop that moves from reproduction to regression testing.

Core Features & Use Cases

  • Provides a phase-based process (reproduce → hypothesise → instrument → fix → regression-test) to structure debugging.
  • Encourages use of domain glossaries and ADRs to maintain consistent mental models during investigation.
  • Supplies a human-in-the-loop template script to guide reproducible experiments.

Quick Start

Run the hitl-loop.template.sh script inside skills/diagnose/scripts to start the interactive diagnostic loop.

Frequently Asked Questions about diagnose

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

FAQPage Schema
How do I debug a hard bug that is difficult to reproduce?

To debug a hard bug, you follow a disciplined diagnosis loop that reproduces, minimises, hypothesises, instruments, fixes, and regression-tests the issue. This structured approach ensures reproducible debugging across codebases.

What is the best way to structure a bug triage workflow for complex regressions?

The best way to structure bug triage is using a phase-based workflow: reproduce, hypothesise, instrument, fix, and regression-test. This enforces a repeatable diagnostic loop for resolving complex regressions.

How do I use instrumentation to diagnose a user-reported error?

You use instrumentation during the hypothesis phase of the diagnostic loop to gather data on a user-reported error. A human-in-the-loop script template guides reproducible experiments to isolate the fault.

Can I apply a structured reproduction loop to any codebase?

Yes, you can apply this structured reproduction loop to any codebase. It guides debugging through tests, instrumentation, and structured hypotheses to accurately address bugs and regressions.

Why do I need domain glossaries and ADRs during debugging?

You need domain glossaries and ADRs during debugging to maintain consistent mental models during investigation. They provide context and architectural decisions to ensure accurate diagnosis across complex codebases.

What are the limitations of ad hoc debugging compared to a diagnostic loop?

Ad hoc debugging lacks a regression-testing phase and structured hypotheses, making it prone to incomplete fixes. A disciplined diagnostic loop enforces reproduction and instrumentation to prevent recurring regressions.