agents-team

You can coordinate multiple AI agents by spawning an orchestrator agent to plan, delegate, and assign roles like worker, reviewer, and researcher. The orchestrator manages task submissions, monitors progress via wait, and merges individual agent outputs into a cohesive final project result.

Role-based task delegation is managed by defining specific roles such as writer, editor, and researcher within your multi-agent system. An orchestrator agent uses spawn_agent to assign tasks to these roles and tracks their submissions, ensuring work is split efficiently across the collaborative workflow.

Yes, multi-agent orchestration supports long-running, chaptered, and project-based workflows. It provides explicit role definitions and traceability for creative or analytical projects, allowing agents to handle planning, review, and synthesis across extended tasks.

You track task progress and merge results by leveraging the wait and list_agents tools to monitor spawned agents. The orchestrator waits for task submissions from workers, reviewers, or researchers, and then synthesizes the individual outputs into a single cohesive deliverable.

Yes, you need to define explicit roles such as orchestrator, worker, reviewer, researcher, writer, and editor when setting up a collaborative multi-agent workflow. Defining these roles allows the orchestrator to properly delegate tasks, monitor submissions, and merge the results effectively.

You should use a multi-agent team for complex, long-running projects that require planning, delegation, review, and synthesis. Splitting work across specialized roles like researcher and editor provides traceability and scalable collaboration that a single agent cannot achieve for long-form workflows.