code-debug

Diagnoses unknown software failures through evidence-driven root-cause analysis and reporting.

1|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/pycode4micro/bushserver --skill code-debug-pycode4micro
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-debug
Source: https://github.com/pycode4micro/bushserver/tree/main/src/bushserver/seeds/skills/package/code-debug
Command: npx skills add https://github.com/pycode4micro/bushserver --skill code-debug-pycode4micro

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

The code-debug skill helps you pinpoint the root cause of broken behavior when you know something is failing but do not yet know where or why.

Core Features & Use Cases

  • Reproduction-first diagnosis: Collect symptoms, triggers, environment details, and reproduction status before forming conclusions.
  • Layered triage: Narrow the search space from symptom surface to system scope to code scope.
  • Ranked root-cause hypotheses: Maintain up to three live, evidence-backed hypotheses and verify them one-by-one.
  • Evidence-backed diagnosis reporting: Produce an impact assessment, ruled-out hypotheses, confidence level, and a minimal repair direction without implementing code fixes.
  • Debug-to-evolution handoff: Confirm the diagnosis is strong enough before switching to code-evolution when the user wants the fix.

Quick Start

Ask the AI to reproduce the failure, triage the suspicious scope, rank likely root causes with evidence, and produce a diagnosis report before proposing code changes.

Frequently Asked Questions about code-debug

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

FAQPage Schema
How do I find the root cause of an intermittent software regression when I don't know where the failure is?

Root-cause analysis for intermittent software regressions requires evidence-driven reproduction and layered triage. By narrowing the search space from symptom surface to code scope, you can rank up to three evidence-backed hypotheses and verify them individually to isolate the failing location.

What is the best way to debug unknown software failures from log anomalies and flaky behavior?

Debugging unknown software failures from log anomalies and flaky behavior demands a reproduction-first diagnosis. You must collect symptoms, triggers, and environment details to triage the suspicious scope, form ranked root-cause hypotheses, and produce an evidence-backed diagnosis report before proposing any code changes.

How do I triage an unresponsive page or exception when the failing location is not yet known?

Triage an unresponsive page or exception by applying a layered triage protocol that narrows the search space from system scope to code scope. Maintain up to three live, evidence-backed root-cause hypotheses and verify them one-by-one to pinpoint the exact failing location.

Does evidence-based diagnosis require me to implement code fixes immediately?

Evidence-based diagnosis does not require immediate code fixes. It produces an impact assessment, ruled-out hypotheses, confidence level, and minimal repair direction. A gated handoff to code-evolution occurs only after the diagnosis report confirms the root cause is strong enough.

When do I need a reproduction protocol for software incident response?

You need a reproduction protocol for software incident response when diagnosing flaky behavior or unresponsive pages. Collecting reproduction status, triggers, and environment details before forming conclusions ensures disciplined hypothesis ranking and prevents premature fixes during root-cause validation.