debug

Investigate unknown defects by forming hypotheses and gathering cited evidence.

139|7|Updated May 8, 2026
One-click install
npx skills add https://github.com/arbiterForge/codeArbiter --skill debug-arbiterforge
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debug
Source: https://github.com/arbiterForge/codeArbiter/tree/main/plugins/ca/skills/debug
Command: npx skills add https://github.com/arbiterForge/codeArbiter --skill debug-arbiterforge

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill investigates unknown defects, anomalies, and unexpected behavior when the cause is not yet understood. It helps teams avoid guesswork by forcing evidence-backed diagnosis before any fix is attempted.

Core Features & Use Cases

  • Structured Investigation: Captures the symptom, reproduction steps, and evidence before analysis begins.
  • Root-Cause Reasoning: Requires multiple distinct hypotheses, including at least one boring environmental explanation, before narrowing down the cause.
  • Evidence-Gated Decisions: Reads logs, traces, commits, and configuration in a read-only way to decide whether the issue should route to fix, adr, or no-action close.
  • Safe Handoff: Produces a concise summary and downstream handoff so fix, architecture review, or task tracking can continue without re-reading the full session.

Quick Start

Use the debug skill to investigate the reported failure, collect a minimal repro, and route the outcome to the correct follow-up action.

Frequently Asked Questions about debug

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

FAQPage Schema
How do I find the root cause of a bug without modifying the codebase?

Find the root cause of an intermittent anomaly by capturing symptoms and forming competing hypotheses, including environmental explanations, before narrowing down the cause. Evidence is gathered from logs, traces, and configuration in a read-only way to prevent guesswork.

What is the best way to debug intermittent anomalies when the cause is unknown?

The best way to debug intermittent anomalies is to apply a disciplined root-cause workflow that requires multiple distinct hypotheses and cited evidence from logs or traces. This evidence-gated process ensures you understand the failure before acting.

How do I structure a debugging investigation for reproducible failures?

Structure a debugging investigation by first capturing the symptom and reproduction steps, then forming multiple hypotheses. Gather cited evidence from read-only logs, commits, and configuration to decide whether to route the outcome to fix, adr, or no-action close.

Can I use logs and traces to diagnose unexpected behavior without applying a fix?

Yes, you can use logs and traces to diagnose unexpected behavior without applying a fix. The investigation process requires read-only analysis of logs, traces, commits, and configuration to gather evidence and decide the correct downstream routing action.

When should I close a debugging session with no-action instead of routing to a fix?

You should close a debugging session with no-action instead of routing to a fix when the evidence-gated analysis of logs, traces, and configuration proves the reported defect or unexpected behavior does not require a code change or architecture review.