Sage Protocol Docs¶
Sage is a governed skill network for AI agents: a trust and distribution layer for agent capability.
Communities govern reusable skills and prompt libraries, publish the versions they trust, and give agents a machine-readable way to discover and use that canon.
The current product is already useful as a working library and governance system. The broader thesis is larger: as governed capability spreads across tools and agents, Sage can become economic infrastructure for how agent expertise is curated, distributed, and rewarded.
What You Can Do With Sage¶
- Package prompts and skills into reusable libraries.
- Promote canonical versions through governance.
- Sync governed capability into agent runtimes.
- Observe what workflows are actually used.
- Reward contributors who improve high-value capability.
Start Here¶
If you want to use Sage in practice, start with Integrating Your Agent.
That guide explains the practical entry point: connect an existing runtime to Sage, let the agent discover governed capability before generating from scratch, keep early rollout read-only, and add governed execution only when your policy is ready.