swarm-advanced

Distributed agent swarms are orchestrated by initializing a topology like mesh, hierarchical, star, or ring, then spawning specialized agents to execute tasks in parallel. This Skill manages the coordination, memory, and performance monitoring for those workflows.

Swarm topologies define how agents communicate and coordinate. Mesh enables decentralized peer-to-peer task sharing, hierarchical uses a top-down command structure, star routes all communication through a central agent, and ring passes tasks sequentially around a loop.

Yes, Claude Flow MCP tools are required. This Skill depends on them to dynamically initialize swarms, spawn agents, orchestrate tasks, and monitor performance across the chosen distributed topology during research or testing workflows.

You initialize a swarm by defining the topology and agent count, such as a mesh swarm with 6 agents. You then spawn a researcher agent by naming it and assigning its role within the orchestrated workflow for parallel information gathering.

Yes, specialized swarms can be deployed for diverse needs. You can configure tailored agent swarms for research, development, testing, and analysis, facilitating end-to-end orchestration from initial planning through execution and reporting.

Agent-based orchestration relies entirely on Claude Flow MCP tools for dynamic initialization and memory management. Without these dependencies, the advanced swarm patterns and parallel execution across topologies cannot function.