agents

To coordinate multi-agent Claude workflows, you use an orchestrator-workers pattern where a lead agent decomposes tasks and assigns them to subagents for parallel execution. It enables scalable task decomposition across multiple Claude instances.

The orchestrator-workers pattern is a multi-agent coordination structure where a lead agent breaks down complex tasks and delegates them to worker agents. It aggregates the subagents' results to complete the overarching workflow.

Yes, you can run parallel task execution across multiple Claude instances by implementing inter-agent communication and worker collaboration patterns. A lead agent delegates subtasks to different agents which execute simultaneously.

Yes, multi-agent orchestration supports building debate patterns and router coordination for complex workflows. It provides pattern-based decision making and inter-agent communication to manage these collaborative agent interactions.

You start by creating a simple orchestrator that decomposes a task into two subagents, assigns parallel work, and aggregates their results. This establishes the foundational lead-agent to subagents communication pattern.

You should use multi-agent orchestration when your workflows require scalable task decomposition, team coordination, and parallel execution that exceeds a single agent's capacity. It is applicable for building complex orchestrator-worker systems and agent teams.