The MCP Trust Framework is an open source, community-driven standard for adding identity and trust to MCP servers.
The Model Context Protocol (MCP) from Anthropic provides a standardized way for AI agents to interact with tools and data sources. However, MCP intentionally does not define trust, identity, or governance mechanisms.
MCPF addresses this gap by providing:
Based on W3C DIDs and Verifiable Credentials, enabling servers to prove their identity and attestations.
Queryable services that list approved servers, issuers, and revocation status.
Policy enforcement mechanisms that ensure agents only use approved, verified tools.
Specification: Draft
Reference Implementation: Alpha
SDKs: Alpha
License: CC BY-NC 4.0
Copyright © 2025 VeriTrust
Licensed under Creative Commons Attribution-NonCommercial 4.0. Free for research, education, and personal use. Commercial licensing available.