swarm-advanced

Distributed swarm orchestration initializes topology via mcp__claude-flow__swarm_init, spawns specialized agents with mcp__claude-flow__agent_spawn, and coordinates parallel or pipeline tasks using mcp__claude-flow__task_orchestrate. This enables parallel task execution across agents in research, development, and testing workflows.

Swarm topologies include mesh, hierarchical, star, and ring patterns with adaptive coordination strategies. Each topology controls agent communication and task routing, allowing you to match the pattern to your workflow's parallelism and dependency structure.

Yes. Cross-swarm memory and performance tracking interfaces enable agents to access shared state and monitor execution metrics. This supports coordinated decision-making and visibility across the distributed workflow.

MCP tools (swarm_init, agent_spawn, task_orchestrate) and associated CLI commands provide functional APIs to initialize topology, spawn agents with specialized roles, and route tasks through parallel or sequential execution modes.

Swarm orchestration patterns address research, development, and testing scenarios requiring parallel task coordination. They handle complexity at scale by distributing work across specialized agents and managing memory and topology state.

Hierarchical topology suits workflows with clear leader-follower dependencies and staged execution. Mesh topology enables full peer-to-peer communication for highly parallel, adaptive coordination where agents need direct access to all other agents.