Agent Adaptive Coordinator

This is a swarm orchestration layer that automatically switches between coordination patterns (hierarchical, mesh, ring, hybrid) based on real-time workload analysis. You'd use it when you need Claude to dynamically adapt how multiple agents coordinate rather than locking into a single topology. The hooks show it analyzes task characteristics like complexity and parallelizability, then picks or switches topologies mid-flight using performance metrics and ML predictions. It's ambitious in scope with neural pattern training and predictive scaling, though the actual switching logic relies heavily on threshold-based heuristics. Best suited for complex multi-agent workflows where coordination overhead meaningfully impacts performance and you want the system to self-optimize rather than manually tuning topology per task type.

skills.sh installs: 1.1K
GitHub stars: 81.4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add ruvnet/ruflo --skill agent-adaptive-coordinator --agent claude-code

Installs into .claude/skills of the current project.

Files

name: adaptive-coordinator type: coordinator color: "#9C27B0"
description: Dynamic topology switching coordinator with self-organizing swarm patterns and real-time optimization capabilities:

topology_adaptation
performance_optimization
real_time_reconfiguration
pattern_recognition
predictive_scaling
intelligent_routing priority: critical hooks: pre: | echo "🔄 Adaptive Coordinator analyzing workload patterns: $TASK"
Initialize with auto-detection
mcp__claude-flow__swarm_init auto --maxAgents=15 --strategy=adaptive
Analyze current workload patterns
mcp__claude-flow__neural_patterns analyze --operation="workload_analysis" --metadata="{"task":"$TASK"}"
Train adaptive models
mcp__claude-flow__neural_train coordination --training_data="historical_swarm_data" --epochs=30
Store baseline metrics
mcp__claude-flow__memory_usage store "adaptive:baseline:${TASK_ID}" "$(mcp__claude-flow__performance_report --format=json)" --namespace=adaptive
Set up real-time monitoring
mcp__claude-flow__swarm_monitor --interval=2000 --swarmId="${SWARM_ID}" post: |