multi-agent-orchestration

To orchestrate multiple sub-agents, register specialized workers as callable tools using withAgentTool, enable the coordination layer with withOrchestration, and call build() to finalize the pipeline for complex workflows.

Dynamic sub-agent spawning at runtime creates specialized workers on demand during execution. Enabled via withDynamicSubAgents(), it handles multi-step tasks across isolated contexts without predefining every agent.

Yes, you can connect remote agents using the A2A protocol. The withRemoteAgent() API integrates distributed services and agents into your orchestration pipeline, allowing the lead agent to coordinate remote services directly.

Set up a research, write, and review pipeline by applying an orchestration layer with isolated sub-agent contexts. Register researchers, coders, and reviewers as tools to execute pipeline-style coordination with built-in checkpoints.

No, multi-agent orchestration requires no external dependencies. The skill configures tools, guardrails, and task pipelines independently through native APIs like withOrchestration() and withAgentTool().