Advanced Swarm Coordination SOP

Overview

This skill implements advanced swarm patterns with dynamic topology switching, self-organizing behaviors, and intelligent coordination for complex multi-agent systems. It enables sophisticated swarm orchestration with adaptive topology selection and performance optimization.

Agents & Responsibilities

hierarchical-coordinator

Role: Tree-based coordination with leader-follower patterns Responsibilities:

• Manage hierarchical swarm structures
• Coordinate parent-child agent relationships
• Handle task delegation cascades
• Monitor hierarchy performance

mesh-coordinator

Role: Peer-to-peer coordination with full connectivity Responsibilities:

• Enable direct agent-to-agent communication
• Manage mesh network topology
• Coordinate distributed consensus
• Handle fault tolerance

adaptive-coordinator

Role: Dynamic topology switching based on workload Responsibilities:

• Analyze task complexity and requirements
• Switch topologies dynamically
• Optimize resource allocation
• Monitor and adapt to performance

Phase 1: Initialize Swarm Infrastructure

Objective

Establish foundation for advanced swarm coordination with proper topology and agent configuration.

Evidence-Based Validation

• Swarm initialized with confirmed topology
• All agents spawned successfully
• Memory coordination active
• Health checks passing

Scripts

# Initialize hierarchical swarm
npx claude-flow@alpha swarm init --topology hierarchical --max-agents 10

# Initialize mesh swarm
npx claude-flow@alpha swarm init --topology mesh --max-agents 8

# Initialize adaptive swarm
npx claude-flow@alpha swarm init --topology adaptive --max-agents 12 --strategy balanced

# Verify initialization
npx claude-flow@alpha swarm status --verbose

# Setup memory coordination
npx claude-flow@alpha memory store --key "swarm/topology" --value "hierarchical"
npx claude-flow@alpha memory store --key "swarm/max-agents" --value "10"

MCP Integration

// Initialize swarm with MCP
mcp__claude-flow__swarm_init({
  topology: "hierarchical",
  maxAgents: 10,
  strategy: "balanced"
})

// Alternative: Mesh topology
mcp__claude-flow__swarm_init({
  topology: "mesh",
  maxAgents: 8,
  strategy: "specialized"
})

// Alternative: Adaptive topology
mcp__claude-flow__swarm_init({
  topology: "adaptive",
  maxAgents: 12,
  strategy: "adaptive"
})

Memory Patterns

# Store swarm configuration
npx claude-flow@alpha memory store \
  --key "swarm/config" \
  --value '{"topology":"hierarchical","maxAgents":10,"strategy":"balanced"}'

# Store agent assignments
npx claude-flow@alpha memory store \
  --key "swarm/agents/coordinator-1" \
  --value '{"type":"hierarchical-coordinator","status":"active","level":0}'

Validation Criteria

1. Swarm ID generated and confirmed
2. Topology matches requested configuration
3. Agent count within specified limits
4. Memory coordination operational
5. Health endpoint responding

Phase 2: Configure Topology

Objective

Select and configure optimal topology pattern based on task requirements and complexity.

Evidence-Based Validation

• Topology selected based on analysis
• Coordinator agents spawned
• Agent connections established
• Topology metrics baseline recorded

Scripts

# Spawn hierarchical coordinator
npx claude-flow@alpha agent spawn \
  --type coordinator \
  --role "hierarchical-coordinator" \
  --capabilities "task-delegation,hierarchy-management"

# Spawn mesh coordinator
npx claude-flow@alpha agent spawn \
  --type coordinator \
  --role "mesh-coordinator" \
  --capabilities "peer-coordination,consensus"

# Spawn adaptive coordinator
npx claude-flow@alpha agent spawn \
  --type coordinator \
  --role "adaptive-coordinator" \
  --capabilities "topology-switching,optimization"

# Configure topology
npx claude-flow@alpha swarm configure \
  --topology hierarchical \
  --levels 3 \
  --branching-factor 3

# Verify topology
npx claude-flow@alpha swarm status --show-topology

MCP Integration

// Spawn coordinator agents
mcp__claude-flow__agent_spawn({
  type: "coordinator",
  name: "hierarchical-coordinator",
  capabilities: ["task-delegation", "hierarchy-management"]
})

mcp__claude-flow__agent_spawn({
  type: "coordinator",
  name: "mesh-coordinator",
  capabilities: ["peer-coordination", "consensus"]
})

mcp__claude-flow__agent_spawn({
  type: "coordinator",
  name: "adaptive-coordinator",
  capabilities: ["topology-switching", "optimization"]
})

Topology Selection Guide

Hierarchical:

• Best for: Clear task hierarchies, delegation workflows
• Pros: Efficient delegation, clear authority
• Cons: Single point of failure at root
• Use when: Tasks have natural parent-child relationships

Mesh:

• Best for: Peer collaboration, distributed consensus
• Pros: High fault tolerance, no bottlenecks
• Cons: Higher communication overhead
• Use when: Agents need direct communication

Star:

• Best for: Centralized coordination, simple workflows
• Pros: Simple control, low complexity
• Cons: Central coordinator bottleneck
• Use when: Single coordinator can handle all traffic

Ring:

• Best for: Sequential processing, pipeline workflows
• Pros: Predictable flow, ordered execution
• Cons: Latency accumulation
• Use when: Tasks must be processed in sequence

Adaptive:

• Best for: Dynamic workloads, variable complexity
• Pros: Automatic optimization, flexible
• Cons: Overhead from topology switching
• Use when: Workload patterns vary significantly

Memory Patterns

# Store topology configuration
npx claude-flow@alpha memory store \
  --key "swarm/topology/config" \
  --value '{"type":"hierarchical","levels":3,"branchingFactor":3}'

# Store baseline metrics
npx claude-flow@alpha memory store \
  --key "swarm/metrics/baseline" \
  --value '{"latency":45,"throughput":120,"agentUtilization":0.75}'

Validation Criteria

1. Coordinator agents active and responsive
2. Topology structure matches configuration
3. Agent connections verified
4. Baseline metrics recorded
5. No configuration errors

Phase 3: Deploy Agents

Objective

Spawn specialized agents based on topology and assign roles with proper coordination.

Evidence-Based Validation

• All required agents spawned
• Agent roles assigned correctly
• Coordination protocols active
• Agent health checks passing

Scripts

# Spawn specialized agents for hierarchical topology
npx claude-flow@alpha agent spawn --type researcher --capabilities "analysis,patterns"
npx claude-flow@alpha agent spawn --type coder --capabilities "implementation,testing"
npx claude-flow@alpha agent spawn --type reviewer --capabilities "quality,security"

# Assign agents to hierarchy levels
npx claude-flow@alpha swarm assign \
  --agent-id "agent-001" \
  --level 1 \
  --parent "coordinator-1"

# Spawn agents for mesh topology
npx claude-flow@alpha agent spawn --type analyst --peer-mode enabled
npx claude-flow@alpha agent spawn --type optimizer --peer-mode enabled

# Configure peer connections
npx claude-flow@alpha swarm connect-peers --all

# List all agents
npx claude-flow@alpha agent list --show-roles --show-connections

MCP Integration

// Spawn specialized agents
mcp__claude-flow__agent_spawn({
  type: "researcher",
  capabilities: ["analysis", "patterns", "research"]
})

mcp__claude-flow__agent_spawn({
  type: "coder",
  capabilities: ["implementation", "testing", "debugging"]
})

mcp__claude-flow__agent_spawn({
  type: "analyst",
  capabilities: ["optimization", "performance", "metrics"]
})

// Check agent status
mcp__claude-flow__agent_list({
  filter: "active"
})

mcp__claude-flow__agent_metrics({
  metric: "all"
})

Agent Assignment Patterns

Hierarchical Assignment:

# Level 0: Root coordinator
# Level 1: Department coordinators
# Le