diagnose

Guide reproduce, hypothesize, instrument, and regression-test phases for debugging.

3|1|Updated Mar 23, 2026

Search Tags:#debugging #bugs #reproduce #diagnose #hitl #instrument #hypothesize

Ready to run

npx skills add https://github.com/malinskibeniamin/skills --skill diagnose-malinskibeniamin

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: diagnose
Source: https://github.com/malinskibeniamin/skills/tree/main/diagnose
Command: npx skills add https://github.com/malinskibeniamin/skills --skill diagnose-malinskibeniamin

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Disciplined diagnosis loops for hard bugs and performance regressions, guiding users from reproduction to hypothesis, instrumentation, fix, and regression testing in a structured, repeatable workflow.

Core Features & Use Cases

Phase-driven debugging workflow spanning reproduction, hypothesis generation, instrumentation, and verification.
Clear guidance and templates to maintain discipline, reduce cognitive load, and increase repro rate.
Human-in-the-loop support via a ready-made script (scripts/hitl-loop.template.sh) to drive deterministic debugging sessions and capture outcomes.
Suitable for bug reports, flaky tests, and performance regressions across codebases and teams, with strict progress checkpoints.

Quick Start

Run the interactive HITL diagnosis loop by executing scripts/hitl-loop.template.sh and follow the prompts.

Frequently Asked Questions about diagnose

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

FAQPage Schema

How do I systematically debug hard bugs and performance regressions?▼

You can systematically debug hard bugs by running a structured six-phase diagnosis loop spanning reproduction, hypothesis generation, instrumentation, fixing with regression tests, and cleanup to reduce cognitive load and increase repro rates.

What is a disciplined diagnosis loop for reproducing flaky tests?▼

A disciplined diagnosis loop for flaky tests is a repeatable workflow that enforces strict progress checkpoints across reproduction, hypothesis, and instrumentation phases. It captures outcomes deterministically, ensuring you verify the bug before attempting a fix.

How do I start a human-in-the-loop debugging session for bug reports?▼

Start a human-in-the-loop debugging session by executing the interactive template script, which drives deterministic debugging phases and prompts you through each step. It captures outcomes and maintains strict progress checkpoints throughout the workflow.

Does this structured debugging workflow work for performance regressions across different teams?▼

Yes, the structured debugging workflow applies to performance regressions, bug reports, and flaky tests across diverse codebases and teams. It provides phase-driven guidance and strict checkpoints to ensure consistent, repeatable diagnosis regardless of the project context.

What is the best way to instrument code during bug diagnosis?▼

The best way to instrument code during bug diagnosis is to follow a structured workflow that moves from reproduction to hypothesis before applying instrumentation. This ensures you collect targeted diagnostic data rather than guessing at the root cause prematurely.