claude-md-authoring

Write and optimize CLAUDE.md files with frontmatter and agent_docs references.

Updated Jul 28, 2025
One-click install
npx skills add https://github.com/hudrazine/claude-code-toolkit --skill claude-md-authoring-hudrazine
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-md-authoring
Source: https://github.com/hudrazine/claude-code-toolkit/tree/main/plugins/claude-md-authoring/skills
Command: npx skills add https://github.com/hudrazine/claude-code-toolkit --skill claude-md-authoring-hudrazine

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Write and optimize CLAUDE.md files to provide universal, maintainable guidance while clearly separating task-specific instructions, ensuring consistent system prompts across projects.

Core Features & Use Cases

  • Establish a concise CLAUDE.md frontmatter with name and description to anchor universal context.
  • Promote progressive disclosure by guiding task-specific details to dedicated agent_docs and references folders.
  • Provide a pointer-based reference pattern that reduces drift and keeps the root CLAUDE.md focused and universally applicable.

Quick Start

Create or update a CLAUDE.md that captures WHY, WHAT, and HOW in the frontmatter and move task-specific content to agent_docs/ and references/ for targeted guidance.

Frequently Asked Questions about claude-md-authoring

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

FAQPage Schema
How do I structure a CLAUDE.md file to keep universal guidance separate from task-specific instructions?

To structure CLAUDE.md files effectively, place universal context in concise frontmatter with name and description, then move task-specific instructions to dedicated agent_docs and references folders for progressive disclosure.

What is progressive disclosure in agent documentation and when do I need it?

Progressive disclosure in agent documentation is the practice of keeping root CLAUDE.md files compact by pointing to detailed task-specific instructions in agent_docs and references, needed when managing complex projects to reduce context drift.

What's the best way to prevent instruction drift in large CLAUDE.md files?

The best way to prevent instruction drift in CLAUDE.md files is using a pointer-based reference pattern, keeping root content universally applicable while delegating task-specific details to structured agent_docs and references folders.

Does CLAUDE.md frontmatter require specific fields for agent context?

Yes, CLAUDE.md frontmatter requires name and description fields to anchor universal context, ensuring the agent receives consistent system prompts across projects while maintaining a compact configuration.

When should I not use progressive disclosure for CLAUDE.md?

You should avoid progressive disclosure for CLAUDE.md when a project lacks task-specific content, as moving minimal instructions to agent_docs and references adds unnecessary structural overhead without reducing drift or improving maintainability.