UTA is an open-source trust interoperability layer for AI agents. It translates between 8 credential formats — including ATC, EAT-AI, ZTA, A2A, MCP, W3C VC, OAuth, and SPIFFE — through a canonical Universal Trust Schema. It provides a 12-stage fail-closed verification pipeline with Ed25519 signatures and RFC 8785 JCS canonicalization, making trust verification portable across agent ecosystems.
AI agents are becoming interoperable, but trust isn't.
We built UTA to solve that gap: one open-source verification layer that lets agents recognize and verify credentials across different ecosystems.
We're looking for developers, security teams, and AI infrastructure builders to test UTA and help shape the standard.
GitHub: github.com/alicelabs-llc/universal-trust-adapter