Audits and optimizes CLAUDE.md files to improve AI instruction adherence and reduce context bloat.
The Claude MD Optimizer is a specialized skill designed to maintain the health and efficiency of your project's CLAUDE.md file. It performs a rigorous six-dimension audit—evaluating instruction budgets, safety, and structural quality—to ensure your AI agent remains performant and follows best practices. By identifying sensitive data, extracting path-specific rules into dedicated modular files, and condensing verbose instructions, it prevents 'instruction fatigue' and helps Claude maintain consistent behavior even in large or complex codebases.
主な機能
01Refactoring of bloated instructions into modular .claude/rules/ files
02Six-dimension scoring audit based on Claude Code best practices
031 GitHub stars
04Automatic detection of leaked credentials and sensitive secrets
05Standardized reporting with letter grades and prioritized recommendations
06Instruction budget measurement and line-count analysis
ユースケース
01Auditing project documentation for accidentally committed API keys or tokens
02Refactoring a bloated CLAUDE.md file that is causing inconsistent AI behavior
03Implementing progressive disclosure for complex, multi-directory projects
