Orchestrate distributed agent swarms via MCP tools and CLI commands.
npx skills add https://github.com/ruvnet/claude-flow --skill swarm-advanced-ruvnet
Please help me install this Agent Skill. Skill: swarm-advanced Source: https://github.com/ruvnet/claude-flow/tree/main/.claude/skills/swarm-advanced Command: npx skills add https://github.com/ruvnet/claude-flow --skill swarm-advanced-ruvnet
This Skill provides advanced patterns and techniques for orchestrating AI agent swarms across diverse applications like research, development, and testing. It enables users to design, deploy, and manage complex distributed workflows with optimal topologies, agent strategies, and robust error handling, maximizing efficiency and reliability.
hierarchical swarm topology with specialized agents (architects, coders, testers). This skill can orchestrate parallel implementation, comprehensive testing, and automated code reviews, ensuring efficient progress and high-quality deliverables.// 1. Initialize swarm topology mcp__claude-flow__swarm_init({ topology: "mesh", maxAgents: 6 })
// 2. Spawn a specialized researcher agent mcp__claude-flow__agent_spawn({ type: "researcher", name: "Agent 1" })
// 3. Orchestrate a task mcp__claude-flow__task_orchestrate({ task: "research AI trends", strategy: "parallel" })
High-intent search queries and answers about installing and using this skill.
Orchestrate distributed agents by initializing a swarm topology (mesh, hierarchical, star, or ring), spawning specialized agents with defined roles, and assigning tasks using parallel or adaptive execution strategies. This coordinates complex workflows across research, development, and testing with optimal agent coordination.
Swarm topologies include mesh (fully connected), hierarchical (manager-worker), star (hub-spoke), and ring (circular communication). Choose based on your workflow complexity: hierarchical suits large teams with specialized roles; mesh maximizes parallel coordination; star and ring reduce communication overhead.
Yes. Adaptive, balanced, specialized, and parallel strategies let agents adjust behavior per task. Specialized agents (architects, coders, testers) excel in hierarchical topologies for large projects; parallel strategies maximize throughput for independent subtasks.
Agents learn from successful workflows to optimize future coordination and task execution. Neural learning continuously refines agent behavior, reducing coordination overhead and improving efficiency across repeated or similar distributed operations.
Mesh topologies enable all agents to communicate directly, maximizing parallelism but increasing complexity. Hierarchical topologies use managers to coordinate workers, reducing direct communication and scaling better for large teams with specialized agent roles.
Yes. Swarm initialization, agent spawning, and task orchestration require claude-flow@alpha dependency, which provides MCP tools and CLI commands for topology setup, memory operations, pattern recognition, and knowledge management.
script-repoCoordinate multi-agent swarms for parallel task execution with dynamic topologies.
CornjebusOrchestrate distributed research, development, and testing workflows with multi-agent swarm topologies.
fableindigo-gifCoordinate pre/post task hooks, session management, and memory across Claude Code operations.
ricableOrchestrate distributed AI agent swarms across mesh, hierarchical, star, and ring topologies.