diagnose

Builds deterministic reproduction loops and structured hypothesis tests to identify root causes of bugs.

5|Updated Mar 30, 2026
One-click install
npx skills add https://github.com/kenlck/skills --skill diagnose-kenlck
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/kenlck/skills/tree/main/ken-swe/skills/diagnose
Command: npx skills add https://github.com/kenlck/skills --skill diagnose-kenlck

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Disciplined diagnosis for hard bugs and performance regressions, turning chaotic symptoms into repeatable, testable steps.

Core Features & Use Cases

  • Build a fast, deterministic feedback loop that reproduces a bug and validates fixes.
  • Follow a phase-driven workflow (reproduce → hypothesise → instrument → fix → regression-test) with guardrails to ensure reliability.
  • Apply across codebases by leveraging domain glossaries and ADRs to maintain a clear mental model of the area being debugged.

Quick Start

Start by constructing a fast, deterministic feedback loop to reproduce the bug and guide hypothesis-driven debugging.

Frequently Asked Questions about diagnose

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

FAQPage Schema
What is the best way to debug a hard bug that is difficult to reproduce?

Disciplined debugging resolves hard bugs by building a deterministic reproduction loop and applying structured hypothesis testing with verifiable instrumentation. This phase-driven workflow transforms chaotic symptoms into repeatable, testable steps to isolate the root cause.

How do I build a deterministic feedback loop for bug reproduction?

To build a deterministic feedback loop for bug reproduction, construct reproducible fixtures and use logs to guide the investigation. This creates a fast, repeatable cycle that validates fixes and ensures the bug is consistently reproduced before applying hypotheses.

Does hypothesis-driven debugging work for end-to-end and integration tests?

Hypothesis-driven debugging applies across unit, integration, and end-to-end scopes. By leveraging domain glossaries and architecture decision records, it maintains a clear mental model of the system while investigating regressions across different testing layers.

What are the phases of a structured debugging workflow for software regressions?

The structured debugging workflow follows five phases: reproduce, hypothesize, instrument, fix, and regression-test. Safe guardrails are applied throughout the process to ensure reliable diagnosis and verifiable fixes for complex performance regressions and bugs.

When should I use instrumentation strategies for diagnosing performance regressions?

Instrumentation strategies are needed when diagnosing performance regressions that lack clear symptoms. By applying targeted instrumentation within a defined feedback loop, you capture the necessary logs and data to validate hypotheses and guide the investigation toward the root cause.

Why do I need safe guardrails when fixing hard bugs?

Safe guardrails are required when fixing hard bugs to ensure verifiable fixes and prevent new regressions. They enforce a disciplined phase-driven workflow where the original bug reproduction loop validates that the applied fix actually resolves the targeted issue.