Audits and optimizes LLM context window usage to improve agent performance and reduce token consumption.
The Context Auditor skill provides a comprehensive diagnostic framework for managing how Claude interacts with your codebase by analyzing memory files, MCP configurations, and command structures. It identifies 'context bloat' that causes sluggish performance or high token costs, applying the R&D (Reduce and Delegate) framework to generate a graded health score and actionable recommendations. This ensures your AI interactions remain sharp, focused, and cost-effective by transitioning from static memory to dynamic priming patterns.
Key Features
011 GitHub stars
02Automated analysis of CLAUDE.md and secondary memory files
03Actionable optimization recommendations based on the R&D framework
04Command infrastructure and priming pattern evaluation
05Context engineering scoring and grading system (A-F)
06MCP server overhead and token consumption estimation
Use Cases
01Optimizing a new codebase for high-efficiency AI collaboration
02Reducing token costs by identifying and removing redundant context files
03Troubleshooting sluggish AI performance or context window warnings
