A2A (Agent-to-Agent Protocol)
The Agent-to-Agent protocol (A2A) is an open standard developed by Google and the Linux Foundation that defines how AI agents delegate work to other AI agents. It provides a structured way for agents to discover capabilities, negotiate tasks, and exchange results, including a built-in state for pausing to request human authorization.
Why It Matters
MCP governs how agents connect to tools. A2A governs how agents connect to each other. As agent architectures become more complex, a single user action may trigger a chain of agents: one that understands the request, one that retrieves data, one that processes it, and one that takes action.
A2A includes a pause-and-authorize state, which acknowledges that some delegated tasks need human approval before proceeding. However, the protocol defines the pause mechanism without defining a decision engine.
This gap is where agent governance layers sit. A2A handles the communication between agents; access management handles the authorization decisions within that communication.