OpenAgenet/OAN: A trust layer for agent identity and discovery

OpenAgenet/OAN (arXiv:2606.03163) introduces a protocol-neutral trust layer designed to solve the foundational problem of agent identity and discovery across heterogeneous frameworks. Authored by Jinliang Xu, the architecture specifies role architecture, identity objects, registration workflow, Root-governed lifecycle, Root-verified package model, authorization-aware Discovery, signed trusted invocation, verification requirements, state transitions, and security properties. Critically, OAN does not define full business conversations between agents; instead, it defines how identities become admissible, discoverable, verifiable, and safe to approach before any protocol-specific interaction begins. It supports existing frameworks including MCP, A2A, ANP-like systems, and domain-specific agent protocols, making it a potential universal trust substrate for open agent ecosystems.

The implementation boundaries and deployment considerations are carefully laid out, ensuring the system can be integrated into existing multi-agent infrastructures without requiring a complete rewrite of interaction protocols. By focusing on the pre-interaction trust layer, OAN addresses a gap that has hindered secure, scalable agent-to-agent communication in the real world. The design prioritizes security properties such as signed trusted invocation and verification requirements, while maintaining flexibility through its protocol-agnostic approach. Researchers and practitioners building multi-agent systems can use OAN to ensure that only verified, trusted agents can discover and invoke each other, reducing risk of malicious actors or spoofed identities in open networks.