Ruflo

Skill: ruflo — AI Agent Orchestration Platform

Executable: ruflo (installed at ~/.npm-global/bin/ruflo) Version: v3.7.0-alpha.34 Source: ruv.io / github.com/ruvnet/claude-flow Emoji: 🔮

What It Does

ruflo is an enterprise AI agent orchestration platform — deploy 60+ specialized agents in coordinated swarms with self-learning, fault-tolerant consensus, vector memory, and MCP integration.

Core capabilities:

• Agent orchestration — spawn, coordinate, and manage multi-agent swarms
• Vector memory — semantic search over agent memory with HNSW indexing (150x–12,500x faster)
• MCP server — Model Context Protocol server management
• Self-learning hooks — intelligent workflow automation
• Guidance Control Plane — compile, retrieve, enforce, and optimize guidance rules
• Autopilot — persistent swarm completion (keeps agents working until ALL tasks done)
• hive-mind — Queen-led consensus-based multi-agent coordination
• Flash Attention — 2.49x–7.47x speedup for embedding operations

When to Use

• You need to orchestrate multiple specialized agents working in coordination
• You want persistent autonomous task completion (autopilot mode)
• You need semantic search over agent memory/knowledge
• You want to route tasks intelligently to specialized agents (Q-Learning-based routing)
• You need MCP server integration for external tools
• You want consensus-based multi-agent decision making (hive-mind)

Key Commands

Architecture (V3)

• 15-agent hierarchical mesh coordination
• AgentDB with HNSW indexing
• Unified SwarmCoordinator engine
• Event-sourced state management
• DDD architecture

Examples

Spawn a coder agent

ruflo agent spawn -t coder

Initialize a V3 swarm

ruflo swarm init --v3-mode

Semantic memory search

ruflo memory search -q "auth patterns"

Start MCP server

ruflo mcp start

System diagnostics

ruflo doctor

Intelligent task routing

ruflo route --task "analyze this codebase"

Autopilot: persistent swarm

ruflo autopilot --goal "complete all open PR reviews"

Integration Notes

Path: ~/.npm-global/bin/ruflo Config: ~/.ruflo/ or project-local .ruflo/

ruflo complements OpenClaw for scenarios requiring:

1. Multi-agent coordination — spawn specialized ruflo agents for complex tasks
2. Enhanced memory — leverage ruflo's vector memory for semantic search
3. Swarm workflows — use autopilot + hive-mind for autonomous task completion
4. Code analysis — route complex analysis tasks via ruflo analyze

Skill Metadata