progressive-disclosure

What problem does it solve?

This Skill teaches an architectural pattern to optimize AI context window usage, preventing information overload and reducing token costs by loading details only when needed. It ensures AI agents receive relevant information efficiently.

Core Features & Use Cases

• 3-Tier Architecture: Organize skills into metadata, entry point (SKILL.md), and on-demand references.
• Context Window Efficiency: Keep SKILL.md concise (<500 lines) and move detailed documentation to linked reference files.
• Workflow-Based Organization: Structure skills by user journey rather than technical features for better discoverability.
• Use Case: Refactor an overly verbose skill that frequently hits context limits, making it more efficient and faster for the AI to process.

Quick Start

Use the progressive-disclosure skill to refactor the 'my-large-skill' by extracting detailed API documentation into a new 'references/api.md' file.