An open protocol for agents to hand tasks to one another, at version 1.0 since March 2026 1 and a Growth Stage project of the Linux Foundation-directed Agentic AI Foundation since August 2026 2. Servers must authenticate every request, but authorisation, and the scope and revocation of authority granted mid-task, are left to the implementer 1.
- Developed in
- ch. 23, Multi-agent systems and delegation chains
- Chapters
- ch. 23, AI Agents
- Contrast with
- MCP
- Source
- 2 numbered references, listed below
Where it is used
One chapter of the Body of Knowledge uses the term. Each link opens the first section that does.
- 23 · AI Agents Kill switch and per-agent circuit breakers 4 mentions
Related terms
Sources
- [1] Agent2Agent (A2A) Protocol Specification, v1.0 (v1.0.0 released 2026-03-12 and v1.0.1 on 2026-05-28 in github.com/a2aproject/A2A; Agent Card at /.well-known/agent-card.json, signed with JWS over JCS-canonicalised JSON; servers authenticate every request; authorisation implementation-specific; scope and revocation of in-task authorisation not defined). A2A Project (Linux Foundation). 2026-05-28. https://a2a-protocol.org/latest/specification/ (verified: primary)
- [2] "A New Chapter for A2A: Joining the Agentic AI Foundation" (A2A accepted as a Growth Stage project of the Linux Foundation-directed Agentic AI Foundation, alongside MCP). A2A Project. 2026-08-27. https://a2a-protocol.org/latest/blog/2026/08/27/a-new-chapter-for-a2a-joining-the-agentic-ai-foundation/ (verified: primary)
Definitions of legal terms paraphrase the cited text, which governs. Dated statements are as of .