Why does this skill discourage including style rules in CLAUDE.md?

Best practices suggest using linters, formatters, or git hooks for style rules. Keeping them out of CLAUDE.md saves valuable context space for high-level architectural and operational instructions.

How does progressive disclosure work in this context?

Instead of pasting massive amounts of information, the skill helps you link to external files (like docs/testing.md). This keeps the main CLAUDE.md file concise while providing paths for the AI to find more detail when needed.

What is the primary benefit of the update-claude-md skill?

It ensures your project context is optimized for AI consumption by following strict standards for brevity and clarity, which prevents the AI from ignoring or misinterpreting instructions.

Can I use this skill to update an existing CLAUDE.md file?

Yes, it can read your current @CLAUDE.md and make targeted updates or refactor it to better align with the recommended 60-150 line standard.

CLAUDE.md Context Optimizer

Name: CLAUDE.md Context Optimizer
Author: doodledood

bydoodledood

•

生産性とワークフロー

Optimizes CLAUDE.md configuration files using best practices to improve Claude Code's project comprehension and reliability.

The update-claude-md skill automates the creation and maintenance of your project's CLAUDE.md file, ensuring it serves as a high-leverage configuration point for agentic workflows. By enforcing principles such as progressive disclosure, brevity (limiting instructions to 300 lines), and the use of verified commands over static style rules, this skill ensures that Claude receives the most relevant context without overwhelming its instruction window. It standardizes documentation into 'What', 'Why', and 'How' sections, making it easier for AI agents to understand your tech stack, architectural relationships, and build processes instantly.

主な機能

01Standardizes CLAUDE.md structure with What, Why, and How sections

02Prevents context bloat by removing style rules and auto-generated boilerplate

038 GitHub stars

04Optimizes instruction length for maximum LLM reliability and performance

05Enforces progressive disclosure by referencing external documentation

06Uses imperative language and verified commands for actionable guidance

ユースケース

01Refactoring an over-long CLAUDE.md to improve agentic task performance

02Bootstrapping a new repository with an AI-optimized configuration file

03Syncing build and test commands across a complex monorepo structure

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doodledood/claude-code-plugins update-claude-md

For use in Claude.ai and ChatGPT

Download Skill