A2A Protocol

The Agent-to-Agent (A2A) protocol, introduced by Google, defines a standard for how independent AI agents discover each other, assign tasks, and exchange messages, filling the gap that the Model Context Protocol (MCP) leaves by covering agent-to-tool connections but not agent-to-agent coordination. Where MCP standardizes how a single agent connects to tools and data sources, A2A standardizes how one agent hands off work to another, enabling a planning agent to delegate to a research agent or a coding agent to coordinate with a testing agent. A2A is still early in adoption, but it signals the direction the ecosystem is moving: from individual agents augmented by tools toward networks of specialized agents that compose like software modules.