Seonglae ChoAgent-to-Agent
Interoperability
Uses HTTP, SSE, and JSON-RPC to ensure compatibility with existing systems. It defines a “task” object with a lifecycle and an “artifact” as the task output. It supports message negotiation using MIME-type “parts” including JSON, images, or video.
Agents describe their functions in structured JSON files called Agent Cards. Cards list each agent’s capabilities, supported modalities, and accepted content types. Client agents use cards to select the correct remote agent for each task. Content-type negotiation supports iframes, video, HTML forms, and structured outputs. No requirement for shared context or memory between agents.
The protocol supports both synchronous and asynchronous execution patterns. It supports short tasks and long-running operations with real-time status updates. It enables state synchronization and failure handling between client and remote agents.
A2A Usages
A2A Notion
- Task
- Message
- Artifact
- Part
Google for Developers Blog - News about Web, Mobile, AI and Cloud
Explore A2A, Google's new open protocol empowering developers to build interoperable AI solutions.
https://developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability/
A2A Protocol
The official documentation for the Agent2Agent (A2A) protocol. The A2A protocol is an open standard that allows different AI agents to securely communicate, collaborate, and solve complex problems together.
https://a2a-protocol.org/latest/
What Is Agent2Agent (A2A) Protocol? | IBM
The Agent2Agent (A2A) protocol is an open communication protocol for artificial intelligence (AI) agents designed for multi-agent systems, allowing interoperability between AI agents from varied providers or those built using different AI agent frameworks.
https://www.ibm.com/think/topics/agent2agent-protocol
