protext

What problem does it solve?

This Skill addresses the challenge of providing AI agents with token-efficient, dynamic, and layered context about a project, preventing information overload and enabling faster, more accurate task completion.

Core Features & Use Cases

• Layered Context: Provides a three-layer hierarchy (PROTEXT.md, index.yaml, Deep Context) for progressive information disclosure.
• Dynamic State Management: PROTEXT.md reflects the project's current state, separate from stable behavior files (CLAUDE.md, etc.).
• Token Efficiency: Balances concise orientation with on-demand deep context extraction to manage token budgets.
• Use Case: When starting a new session on a large codebase, invoke /protext to get a quick ~500 token overview, then use @deep:network to pull specific details about network configuration without loading the entire project.

Quick Start

Invoke /protext at session start to load token-efficient project orientation.