a2a-protocol-pro
A2A protocol and multi-agent systems (professional)
Skill text is English; match the user's response language from Cursor User Rules / project rules when applicable.
Use the official A2A specification and AAIF documentation as authority. This skill encodes agent discovery (Agent Cards), task lifecycle, orchestration patterns (Gateway, Pipeline, Mesh), and security for multi-agent systems. Confirm number of agents, communication pattern, trust model, and deployment environment before proposing an architecture.
Boundary
a2a-protocol-pro owns agent discovery (Agent Cards), task delegation, task lifecycle, orchestration patterns, and A2A security. It does not own single-agent tool access (MCP), general API design, or the agent implementation itself — combine with mcp-server-pro, api-design-pro, and ai-integration-pro as needed.
| Skill | When to combine with a2a-protocol-pro |
|---|---|
mcp-server-pro |
When each agent needs MCP tool access |
api-design-pro |
When A2A interacts with broader HTTP APIs |
ai-integration-pro |
When designing the LLM and reasoning inside each agent |
security-pro |
When threat modeling the multi-agent attack surface |
docker-pro / deployment-pro |
When deploying agent clusters |