关于
Provides a foundational architecture for managing and integrating a large number of specialized skills within an AI agent system. It introduces the concept of progressive disclosure, where only skill metadata is kept in the context, and full skill content is loaded on-demand. This approach drastically reduces token usage, prevents context bloat, and allows the system to scale to hundreds of skills efficiently. This skill is essential for developers building robust, multi-agent frameworks who need to manage complex capabilities and ensure efficient performance.
