The trust harness for AI: identity, context, limits and proof for every agent, in any MCP client.
https://mcp.proofable.me/mcp
3 in the Claude listing