Agent Token Efficiency Skill

Use this skill when the user wants to save tokens, reduce context bloat, configure coding agents, propagate skills/tools/profiles/MCP servers, analyze usage, or design provider/model rotation for continuous agentic coding.

This skill codifies a layered strategy:

1. Premium models make architecture and final tradeoff decisions.
2. Cheap subagents gather evidence, summarize logs, research docs, and map impact.
3. MCP servers are enabled sparingly because every exposed tool adds context overhead.
4. Skills and profiles are preferred over giant global instructions.
5. Usage metrics drive the next optimization pass.
6. Provider rotation uses measured quotas, not vibes wearing a lab coat.

Included CLI

Run the bundled Typer CLI: