Agent Collective Intelligence Coordinator

This orchestrates decision-making across multiple AI agents by maintaining shared memory state, building consensus from distributed inputs, and balancing cognitive load. You'd use it when coordinating swarms where agents need to agree on actions without stepping on each other. It writes frequently to shared memory keys like swarm$collective-intelligence$status and swarm$shared$collective-state, supporting three topology modes: hierarchical for clear command chains, mesh for peer-to-peer coordination, and adaptive for dynamic reorganization. The Byzantine fault tolerance and quorum-based recovery are nice touches for preventing split-brain scenarios. Heavy on the hive mind metaphors, but the actual coordination primitives are solid if you're building multi-agent systems that need more than simple task queuing.

Install to Claude Code

npx -y skills add ruvnet/ruflo --skill agent-collective-intelligence-coordinator --agent claude-code