The rise of AI agents has created a critical identity gap online: How can a service know which agent is interacting with it, who authorized the action, and what permissions have been granted? To solve this, Vouched developed the Model Context Protocol - Identity (MCP-I), an open-source framework that provides a complete identity and delegation layer for AI agents, using cryptographically verifiable identities, Decentralized Identifiers (DIDs), and Verifiable Credentials (VCs).

Vouched is now formally donating MCP-I to the Decentralized Identity Foundation (DIF) to be advanced as a community-driven open standard under the Trusted AI Agents Working Group, ensuring a secure and interoperable way for agents to operate on the user's behalf.

MCP-I makes identity and delegation verifiable; services still apply risk policy, monitoring, and abuse defenses using audit and reputation signals. MCP-I defines three conformance levels to accommodate different security requirements and adoption stages.

Originally published on the Decentralized Identity Foundation (DIF).

For full article, visit the source.