What is the purpose of grading CLAUDE.md and AGENTS.md?

Grading identifies structural issues, duplicate content, and bloat that can confuse AI agents or waste tokens, ensuring your instructions are clear and cost-effective.

How does the skill help with token optimization?

It identifies sections that are too long and automatically moves them to reference files or documentation folders, using pointers to keep the main config file lean.

Can this skill automatically fix my configuration files?

Yes. After performing an audit, the skill presents a prioritized improvement plan and can implement changes—such as splitting files or adding imports—upon your approval.

Does it support monorepo structures?

Absolutely. The skill can detect monorepos and implement package-level CLAUDE.md files to ensure specific rules are only loaded when relevant.

Claude Config Grader & Optimizer

Name: Claude Config Grader & Optimizer
Author: SpillwaveSolutions

bySpillwaveSolutions

Herramientas para Desarrolladores

Grades, audits, and automatically refactors CLAUDE.md and AGENTS.md configuration files to optimize agent performance and token usage.

Acerca de

This skill provides a comprehensive auditing and improvement framework for Claude Code and Codex configuration files, ensuring your AI agents operate at peak efficiency. It automatically evaluates CLAUDE.md and AGENTS.md files against a rigorous 100-point rubric covering structure, content quality, PDA implementation, and maintainability. Beyond just scoring, the skill generates prioritized improvement plans and can automatically refactor oversized files by splitting content into subdirectories, adding table of contents, and implementing @imports to reduce token consumption and improve agent responsiveness.

Características Principales

Intelligent file refactoring to split oversized configurations into manageable structures
Token usage optimization by extracting large sections to reference documentation
Comprehensive before-and-after reporting to verify configuration improvements
Automated scoring based on standardized rubrics for CLAUDE.md and AGENTS.md
0 GitHub stars
Automatic generation of Table of Contents and @imports for enhanced organization

Casos de Uso

Auditing existing project instructions to ensure compliance with agentic best practices
Standardizing agent configurations across multiple projects or within a complex monorepo
Optimizing a bloated CLAUDE.md file that has become too large and token-intensive

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add spillwavesolutions/grading-claude-agents-md-agentic-skill grading-claude-agents-md

For use in Claude.ai and ChatGPT

Download Skill

GitHub