diagnose

Guide a six-phase debugging loop from reproduction to regression testing.

12|1|Updated Feb 13, 2026
One-click install
npx skills add https://github.com/helderberto/agent-skills --skill diagnose-helderberto
Or copy as Structured Prompt for Agentβ–Ό
Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/helderberto/agent-skills/tree/main/skills/diagnose
Command: npx skills add https://github.com/helderberto/agent-skills --skill diagnose-helderberto

SYSTEM DOCUMENTATION & REQUIREMENTS

πŸ’‘ This Skill includes scripts (resource) components.

What problem does it solve?

Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce β†’ minimise β†’ hypothesise β†’ instrument β†’ fix β†’ regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.

Core Features & Use Cases

  • Establish a deterministic, multi-phase debugging workflow that guides reproduction, hypothesis testing, instrumentation, and verification.
  • Emphasize creating a structured feedback loop with explicit checkpoints and measurable signals.
  • Apply across codebases, tests, services, and user flows to locate root causes and validate fixes.

Quick Start

Initiate Phase 1 by building a deterministic feedback loop to reproduce the bug and gather signals.

Frequently Asked Questions about diagnose

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

FAQPage Schema
What is a structured debugging workflow for diagnosing hard bugs and performance regressions?β–Ό

A structured debugging workflow guides you through a six-phase loop: reproduce, hypothesise, instrument, fix, regression test, and cleanup. It requires deterministic signals and explicit checkpoints to locate root causes and validate fixes across codebases, tests, and async systems.

How do I debug a performance regression in an async system across components?β–Ό

Debug a performance regression by first reproducing the issue to gather deterministic signals. Then rank hypotheses, apply targeted probes to instrument the async system across components, and verify the fix through regression testing with clear success criteria.

How do I establish a deterministic reproduction loop for a bug reported in tests?β–Ό

Establish a deterministic reproduction loop by isolating the failing test scenario and gathering measurable signals. This feedback loop forms Phase 1 of the diagnosis process, ensuring you can reliably trigger the bug before hypothesising and instrumenting targeted probes.

Does this disciplined bug diagnosis process work for both code and test failures?β–Ό

Yes, this bug diagnosis process applies to bugs reported in code, tests, and async systems. It uses a multi-phase feedback loop with explicit checkpoints to ensure disciplined hypothesis testing and fix verification regardless of the component layer.

What's the best way to rank hypotheses when hunting for a root cause?β–Ό

The best way to rank hypotheses is to apply targeted probes during the instrumentation phase, collecting deterministic signals that validate or invalidate each assumption. This structured approach minimizes guesswork and isolates the exact root cause efficiently.

When should I not use a multi-phase debugging workflow?β–Ό

Avoid using a multi-phase debugging workflow for obvious or trivial bugs where the root cause is immediately visible. This disciplined loop is designed for hard bugs and performance regressions that require deterministic reproduction, hypothesis ranking, and regression testing to verify.