adk-multi-agent

Design an ADK 2.0 multi-agent system (coordinator, A2A, graph workflow, or dynamic workflow)

Use this skill to call a remote A2A (Agent-to-Agent) endpoint from inside an ADK 2.0 agent. Triggers on: "call A2A agent", "consume A2A endpoint", "ADK calls remote agent", "use external agent as a tool", "A2A client", "agent-to-agent client", "remote agent reference". Generates an RemoteA2AAgent (or A2A tool wrapper) bound to the remote agent card URL, ready to drop into a coordinator's sub_agents=[] or tools=[] list.

Use this skill to expose an ADK 2.0 agent as an A2A (Agent-to-Agent) protocol service so other agents — including non-Python agents in Go, Java, or TypeScript — can call it. Triggers on: "expose ADK agent as A2A", "publish agent over A2A", "make ADK agent callable from another language", "A2A server", "agent-to-agent protocol", "A2A extension". Generates a server entrypoint that wraps your root_agent with the A2A protocol layer and serves it over HTTP.

Use this skill to design an ADK 2.0 coordinator agent with multiple subagents — the canonical multi-agent pattern. Triggers on: "ADK coordinator", "ADK subagents", "build an ADK agent team", "parent agent with sub_agents", "delegating ADK agent", "ADK collaborative agents", "agent that routes to specialists". Generates a coordinator LlmAgent with sub_agents=[...] containing N specialist agents, plus instruction text that teaches the coordinator when to delegate to each.

Use this skill to build ADK 2.0 dynamic workflows — code-defined logic for iterative loops, conditional branching, and runtime-determined agent composition. New in ADK 2.0. Triggers on: "ADK dynamic workflow", "ADK code-driven agent flow", "iterative agent loop with custom logic", "branching ADK workflow", "dynamic agent composition", "runtime agent graph", "ADK 2.0 imperative workflow". Generates a Python orchestration function that decides next steps based on session state.

Use this skill to design ADK 2.0 graph-based workflows — the new-in-2.0 feature for deterministic task routing with explicit node/edge control flow. Triggers on: "ADK graph workflow", "graph-based agent routing", "ADK 2.0 deterministic routing", "agent graph", "node-edge agent flow", "directed graph of agents", "ADK GraphAgent". Generates a graph definition with typed edges, conditional transitions, and entry/exit nodes — strictly deterministic, unlike the LLM-driven coordinator pattern.