Claude Skill Usage System

About

Acts as the core infrastructure for a filesystem-based skill architecture, enabling automatic skill discovery and invocation through a progressive disclosure model. It optimizes token usage by loading only high-level metadata at startup and fetching detailed instructions, executable scripts, or reference resources only when contextually relevant. By integrating directly with project configuration files, it streamlines how Claude identifies and applies specialized domain knowledge to complex development tasks.

Key Features

• Progressive disclosure system to optimize token consumption
• Three-level loading architecture for metadata, instructions, and resources
• Support for explicit and automatic skill invocation based on context
• 0 GitHub stars
• Filesystem-based access to bundled executable scripts and documentation
• Automatic skill discovery and registration via CLAUDE.md integration

Use Cases

• Optimizing context window usage in projects with numerous specialized instructions
• Automating the registration of new domain-specific capabilities into existing workflows
• Standardizing skill discovery and management across different development environments