The identity layer for AI agents is finally being built. What MCP, A2A and new research delivered since March, and what's still missing.

Check more stories related to machine-learning at: https://hackernoon.com/c/machine-learning.

You can also check exclusive content about #ai-agents, #ai-security, #a2a-protocol, #mcp-server, #ai-agent-identity, #oauth-2.1, #enterprise-ai, #hackernoon-top-story, and more.

This story was written by: @garagon. Learn more about this writer by checking @garagon's about page,

and for more stories, please visit hackernoon.com.

In March I wrote that AI agents lack verifiable identity and called it a security crisis. Four months later the diagnosis holds, but the ecosystem started to respond. MCP made authorization a first-class roadmap item and shipped Enterprise-Managed Authorization. A2A reached v1.0 with signed Agent Cards and modern OAuth. Research delivered task-scoped authorization and verifiable delegation chains. Still missing: per-instance identity, multi-hop attenuation and end-to-end provenance.