| name | swarm-coordination |
| description | Orchestrate multi-agent swarms for complex parallel task execution |
| version | 1.0.0 |
| invocable | true |
| author | agentic-flow |
| capabilities | ["swarm_init","agent_spawn","task_orchestration","topology_optimization"] |
Swarm Coordination Skill
Coordinate multi-agent swarms for parallel task execution with intelligent topology selection.
Quick Start
npx agentic-flow@alpha swarm init --topology mesh --agents 5
npx agentic-flow@alpha swarm spawn --type researcher --name "Research Agent"
npx agentic-flow@alpha swarm spawn --type coder --name "Code Agent"
npx agentic-flow@alpha swarm orchestrate "Implement feature X with tests"
npx agentic-flow@alpha swarm status
Topologies
Mesh (Default)
- All agents communicate directly
- Best for: Collaborative tasks, code review
- Latency: Low
- Scalability: Medium
Hierarchical
- Tree structure with coordinator
- Best for: Large projects, delegation
- Latency: Medium
- Scalability: High
Ring
- Sequential communication
- Best for: Pipeline processing, CI/CD
- Latency: Higher
- Scalability: Medium
Star
- Central hub coordinates all
- Best for: Simple coordination
- Latency: Low
- Scalability: Low
Agent Types
| Type | Description | Use Case |
|---|
| researcher | Deep analysis | Requirements gathering |
| coder | Implementation | Feature development |
| tester | Quality assurance | Test creation |
| reviewer | Code quality | PR review |
| architect | System design | Architecture decisions |
| coordinator | Task routing | Complex workflows |
MCP Tools
mcp__claude-flow__swarm_init({ topology: "mesh", maxAgents: 8 })
mcp__claude-flow__agent_spawn({ type: "coder", name: "Feature Dev" })
mcp__claude-flow__task_orchestrate({
task: "Implement OAuth",
strategy: "parallel"
})
Best Practices
- Right-size your swarm: Start with 3-5 agents
- Choose topology wisely: Match to task structure
- Use Claude Code Task tool: For actual agent spawning
- Monitor status: Check for bottlenecks
- Clean up: Destroy swarm when done