context-debugging

Official

Diagnose agent failures by fixing context.

AuthorAndurilCode
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill helps you identify why an AI agent produces wrong, ignored, hallucinated, or inconsistent outputs when the root cause is not an obvious code bug but the assembled context or instructions.

Core Features & Use Cases

  • Context-first triage: Verifies what the model actually receives (system prompt, tools, retrieved docs, and history) before diagnosing.
  • Failure classification for agent behavior: Routes issues into common context failure modes like regression, missing context, tool definition problems, buried or conflicting instructions, and context overflow, then matches each to focused fixes.
  • Decision-ready workflow: Provides step-by-step diagnosis questions plus targeted quick and full remediation guidance, and includes an explicit boundary for genuine reasoning failures.

Quick Start

Use the skill when you suspect your agent keeps failing or ignoring instructions by running the context-debugging triage questions and applying the corresponding quick fix for the matched failure category.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: context-debugging
Download link: https://github.com/AndurilCode/craftwork/archive/main.zip#context-debugging

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.