a2a-protocol

name: a2a-protocol description: >- Reference documentation for the Agent2Agent (A2A) protocol. Use when building A2A servers or clients, configuring Ark A2AServer resources, debugging A2A communication, or answering questions about the A2A specification, Agent Cards, task lifecycle, streaming, extensions, or protocol bindings. allowed-tools: Read, Grep

A2A Protocol Reference

Reference skill for the Agent2Agent (A2A) Protocol, an open standard by Google / the A2A Project (Linux Foundation) for communication between independent AI agent systems.

Attribution: All specification and topic content in references/ is sourced from the official A2A repository under the Apache-2.0 license. Each file includes source attribution in its YAML frontmatter.

IMPORTANT — Use v0.3 by default. The primary references in references/v0.3/ are v0.3.0, the version currently supported by the official Go and Python SDKs and most implementations. The v1.0 RC spec is in references/v1.0-rc/ and should only be consulted when explicitly working on v1.0 features or migration planning.

When to use this skill

• Building or debugging an A2A server or client
• Configuring Ark A2AServer custom resources
• Understanding Agent Cards, task lifecycle, or message formats
• Implementing streaming (SSE) or push notifications
• Working with A2A extensions
• Comparing A2A with MCP

Protocol overview

A2A enables agents built on different frameworks to discover capabilities, negotiate interaction modes, manage collaborative tasks, and exchange information — without exposing internal state, memory, or tools.

Core actors:

• User — human or automated service initiating a request
• A2A Client — application or agent acting on behalf of the user
• A2A Server — agent exposing an HTTP endpoint implementing A2A

Core elements:

Interaction patterns:

• Request/Response — synchronous with polling for long-running tasks
• Streaming (SSE) — real-time incremental updates over open connection
• Push Notifications — async webhooks for disconnected/long-running tasks

Task lifecycle: submitted → working → input-required → completed / failed / canceled

Agent discovery: Clients find agents via /.well-known/agent.json

Specification reference (v0.3)

The primary A2A specification (v0.3.0) is split into sections in references/v0.3/:

Topic guides (v0.3)

Conceptual guides from the A2A documentation:

Key JSON-RPC methods (v0.3)

Agent Card example

{
  "name": "My Agent",
  "description": "An agent that does useful things",
  "url": "https://myagent.example.com/a2a",
  "version": "1.0.0",
  "capabilities": {
    "streaming": true,
    "pushNotifications": true
  },
  "skills": [
    {
      "id": "summarize",
      "name": "Summarize Text",
      "description": "Summarizes long text into key points"
    }
  ],
  "securitySchemes": {
    "bearer": {
      "type": "http",
      "scheme": "bearer"
    }
  },
  "security": [{ "bearer": [] }]
}

Discovered at: https://myagent.example.com/.well-known/agent.json

Quick lookup guide

• "How do I discover agents?" → topic-agent-discovery.md, spec-05-agent-card.md
• "What are the task states?" → topic-life-of-a-task.md, spec-06-data-objects.md
• "How does streaming work?" → topic-streaming-and-async.md, spec-07-rpc-methods.md
• "What's the difference between A2A and MCP?" → topic-a2a-and-mcp.md
• "How do extensions work?" → topic-extensions.md
• "What security/auth is needed?" → spec-04-authentication.md, topic-enterprise-ready.md
• "What are the error codes?" → spec-08-error-handling.md
• "Show me workflow examples" → spec-09-workflows.md

v1.0 RC reference (use only when needed)

The v1.0 Release Candidate spec is available in references/v1.0-rc/ for forward-looking work. Key differences from v0.3 are documented in v1.0-rc/topic-whats-new-v1.md.

Only consult v1.0 content when:

• Explicitly planning migration from v0.3 to v1.0
• Working on features that require v1.0-specific capabilities
• The user specifically asks about v1.0