{"slug": "ruflo-an-agent-meta-harness-for-claude-code-and-codex", "title": "Ruflo: An agent meta-harness for Claude Code and Codex", "summary": "Ruflo, an agent meta-harness built by rUv and powered by Cognitum.One, adds 100+ specialized agents, coordinated swarms, self-learning memory, federated comms across machines, and enterprise security guardrails to Claude Code and Codex. The tool, previously called Claude Flow, can be installed via a CLI (`npx ruflo init`) for full functionality including 98 agents, 60+ commands, 30 skills, MCP server, hooks, and daemon, or as a plugin with limited features.", "body_md": "Agent = Model + Harness.The model writes; the harness gives it tools, memory, loops, sandboxes, and controls so it can actually work.Ruflo is the harness— the execution layer around Claude Code and Codex that adds 100+ specialized agents, coordinated swarms, self-learning memory, federated comms across machines, and enterprise security guardrails. So agents don't just run, they collaborate.\n\nOne `npx ruflo init`\n\ngives Claude Code a nervous system: agents self-organize into swarms, learn from every task, remember across sessions, and — with federation — securely talk to agents on other machines without leaking data. You keep writing code. Ruflo handles the coordination.\n\n``` php\nSelf-Learning / Self-Optimizing Agent Architecture\n\nUser --> Ruflo (CLI/MCP) --> Router --> Swarm --> Agents --> Memory --> LLM Providers\n                          ^                           |\n                          +---- Learning Loop <-------+\n```\n\nNew to Ruflo?You don't need to learn 314 MCP tools or 26 CLI commands. After`init`\n\n, just use Claude Code normally — the hooks system automatically routes tasks, learns from successful patterns, and coordinates agents in the background.\n\n**📖 Background — where the name comes from**\n\nClaude Flow is now Ruflo — named by\n\n[, who loves Rust, flow states, and building things that feel inevitable. The \"Ru\" is the rUv. The \"flo\" is working until 3am. Underneath, powered by]`rUv`\n\n[agentic architecture, running a supercharged Rust-based AI engine, embeddings, memory, and plugin system.]`Cognitum.One`\n\nThere are **two different install paths** with very different surface areas. Pick based on what you need (#1744):\n\nClaude Code Plugin |\nCLI install (`npx ruflo init` ) |\n|\n|---|---|---|\n| What it gives you | Slash commands + a few skills + agent definitions per-plugin | Full Ruflo loop — 98 agents, 60+ commands, 30 skills, MCP server, hooks, daemon |\n| Files in your workspace | Zero |\n`.claude/` , `.claude-flow/` , `CLAUDE.md` , helpers, settings |\n| MCP server registered | No (`memory_store` , `swarm_init` , etc. unavailable to Claude) |\nYes |\n| Hooks installed | No | Yes |\n| Best for | Try a single plugin's commands without committing to the full install | Production use — everything works as documented |\n\n```\n# Add the marketplace\n/plugin marketplace add ruvnet/ruflo\n\n# Install core + any plugins you need\n/plugin install ruflo-core@ruflo\n/plugin install ruflo-swarm@ruflo\n/plugin install ruflo-rag-memory@ruflo\n/plugin install ruflo-neural-trader@ruflo\n```\n\nThis adds slash commands and agent definitions only. The Ruflo MCP server is NOT registered, so `memory_store`\n\n, `swarm_init`\n\n, `agent_spawn`\n\n, etc. won't be callable from Claude. For the full loop, use Path B below.\n\n**🔌 All 35 plugins**\n\n| Plugin | What it does |\n|---|---|\nruflo-core |\n\n**ruflo-swarm****ruflo-autopilot****ruflo-loop-workers****ruflo-workflows****ruflo-federation**| Plugin | What it does |\n|---|---|\nruflo-agentdb |\n\n**ruflo-rag-memory****ruflo-rvf****ruflo-ruvector**[— GPU-accelerated search, Graph RAG, 103 tools](https://npmjs.com/package/ruvector)`ruvector`\n\n**ruflo-knowledge-graph**| Plugin | What it does |\n|---|---|\nruflo-intelligence |\n\n**ruflo-graph-intelligence****ruflo-daa****ruflo-ruvllm****ruflo-goals**| Plugin | What it does |\n|---|---|\nruflo-testgen |\n\n**ruflo-browser****ruflo-jujutsu****ruflo-docs**| Plugin | What it does |\n|---|---|\nruflo-security-audit |\n\n**ruflo-aidefence**| Plugin | What it does |\n|---|---|\nruflo-adr |\n\n**ruflo-ddd****ruflo-sparc****ruflo-metaharness**[guide](/ruvnet/ruflo/blob/main/docs/metaharness-user-guide.md))** ruflo-arena**| Plugin | What it does |\n|---|---|\nruflo-migrations |\n\n**ruflo-observability****ruflo-cost-tracker**| Plugin | What it does |\n|---|---|\nruflo-agent |\n\n**ruflo-plugin-creator**| Plugin | What it does |\n|---|---|\nruflo-iot-cognitum |\n\n**ruflo-neural-trader**[— AI trading with 4 agents, backtesting, 112+ tools](https://npmjs.com/package/neural-trader)`neural-trader`\n\n**ruflo-market-data****macOS / Linux / WSL / Git-Bash:**\n\n```\n# One-line install (POSIX shells only — see Windows note below)\ncurl -fsSL https://cdn.jsdelivr.net/gh/ruvnet/ruflo@main/scripts/install.sh | bash\n```\n\n**All platforms (including native Windows PowerShell / cmd):**\n\n```\n# Interactive setup wizard — runs identically on every platform\nnpx ruflo@latest init wizard\n\n# Quick non-interactive init\n# npx ruflo@latest init\n\n# Or install globally\nnpm install -g ruflo@latest\n```\n\n💡\n\nWindows users:the`curl ... | bash`\n\nform needs a POSIX shell (Git-Bash, WSL, MSYS). The`npx ruflo@latest init wizard`\n\nline works natively in PowerShell and cmd. If you hit an`'bash' is not recognized`\n\nerror, use the`npx`\n\nline instead — both end up running the same init flow.\n\n```\n# Add Ruflo as an MCP server in Claude Code (canonical form, matches USERGUIDE.md)\nclaude mcp add ruflo -- npx ruflo@latest mcp start\n```\n\n| Capability | Description |\n|---|---|\n🤖 100+ Agents |\nSpecialized agents for coding, testing, security, docs, architecture |\n📡 Comms Layer |\nZero-trust federation — agents across machines/orgs discover, authenticate, and exchange work securely |\n🐝 Swarm Coordination |\nHierarchical, mesh, and adaptive topologies with consensus |\n🧠 Self-Learning |\nSONA neural patterns, ReasoningBank, trajectory learning |\n💾 Vector Memory |\nHNSW-indexed AgentDB — measured ~1.9x faster at N=20k, ~3.2x–4.7x at N=5k vs brute force (recall@10 ~0.99); ANN wins above the crossover, ties/loses at small N. See\n`scripts/benchmark-intelligence.mjs` |\n\n**Background Workers****Plugin Marketplace****Multi-Provider****Security****Agent Federation**[MetaHarness](/ruvnet/ruflo/blob/main/docs/metaharness-user-guide.md)`ruflo eject`\n\nturns a ruflo project into a standalone agent toolkit with its own name. [Full guide](/ruvnet/ruflo/blob/main/docs/metaharness-user-guide.md).[Web UI Beta](https://flo.ruv.io/)[RuFlo Research](https://goal.ruv.io/)[/agents](https://goal.ruv.io/agents)### Web UI (Beta) — self-hostable, hosted demo at [flo.ruv.io](https://flo.ruv.io/)\n\n**RuFlo's web UI is a multi-model AI chat with built-in Model Context Protocol (MCP) tool calling.** Talk to Qwen, Claude, Gemini, or OpenAI while RuFlo invokes the same MCP tools the CLI uses — agent orchestration, persistent memory, swarm coordination, code review, GitHub ops — directly from chat. No install, no API key needed to try it.\n\n| What it is | Why it matters | |\n|---|---|---|\n| 🧠 | Any model, local or remote |\n6 curated frontier models out-of-the-box — Qwen 3.6 Max (default), Claude Sonnet 4.6, Claude Haiku 4.5, Gemini 2.5 Pro, Gemini 2.5 Flash, OpenAI — via OpenRouter. Add your own: any OpenAI-compatible endpoint (vLLM, Ollama, LM Studio, Together, Groq, self-hosted). |\n| 🦾 | ruvLLM self-learning AI |\nNative support for\n`ruvnet/RuVector/examples/ruvLLM` ) — RuFlo's self-improving local model layer. Routes to MicroLoRA adapters, learns from your trajectories via SONA, and stays on your machine. Pair with the cloud models or run fully offline. |\n\n**~210 tools, ready to call****Bring your own MCP servers****MCP (n)** pill in the chat input →*Add Server*and paste any MCP endpoint (HTTP, SSE, or stdio). Your tools join RuFlo's native ones in the same parallel-execution flow. Run a local MCP server on`localhost:3000`\n\nand it just works.**Tools run in parallel*** Step 1 — 2 tools completed*badge so you can see exactly what ran.**Memory that sticks***\"remember my favorite color is indigo\"*and ask weeks later — RuFlo recalls it. Backed by AgentDB + HNSW vector search (measured ~1.9x–4.7x faster than brute force above the crossover, recall@10 ~0.99).**Built-in capabilities tour****Self-hostable**`ruflo/src/ruvocal/Dockerfile`\n\n) with embedded Mongo. Deploy to your own Cloud Run / Fly / Kubernetes / docker-compose. The hosted [flo.ruv.io](https://flo.ruv.io/)demo is one option; running your own is fully supported.**Zero install to try****Try the hosted demo:** [https://flo.ruv.io/](https://flo.ruv.io/) — no account, no API key. **Run your own:** the source lives in [ ruflo/src/ruvocal/](/ruvnet/ruflo/blob/main/ruflo/src/ruvocal) with a multi-stage Dockerfile (\n\n`INCLUDE_DB=true`\n\nbuilds in MongoDB) and a `cloudbuild.yaml`\n\nfor Google Cloud Run. See [ADR-033](/ruvnet/ruflo/blob/main/ruflo/docs/adr/ADR-033-RUVOCAL-WASM-MCP-INTEGRATION.md)for the architecture and\n\n[issue #1689](https://github.com/ruvnet/ruflo/issues/1689)for the roadmap.\n\n### Goal Planner UI — autonomous agents at [goal.ruv.io](https://goal.ruv.io/)\n\n**Turn high-level goals into executable agent plans.** `goal.ruv.io`\n\nis RuFlo's hosted Goal-Oriented Action Planning (GOAP) front-end — describe an outcome in plain English and watch RuFlo decompose it into preconditions, actions, and an A* path through state space, then dispatch the work to live agents at [ /agents](https://goal.ruv.io/agents).\n\n| What it is | Why it matters | |\n|---|---|---|\n| 🎯 | Plain-English goals |\nType \"ship the auth refactor with tests and a PR\" — RuFlo extracts the success criteria, the constraints, and the implicit preconditions. No JSON, no DSL. |\n| 🧭 | GOAP A* planner |\nClassic gaming-AI planning ported to software work: state-space search through actions with preconditions/effects to find the shortest viable path. Replans on the fly when state changes. |\n| 🤖 | Live agent dashboard |\n|\n\n**Visual plan tree*** exactly*why an agent picked a path — no opaque chain-of-thought.**Adaptive replanning****Shared memory + SONA****Wired to MCP tools****Zero install to try**[goal.ruv.io](https://goal.ruv.io/), describe a goal, watch it run. Source lives in[— Vite + Supabase, self-hostable.](/ruvnet/ruflo/blob/main/v3/goal_ui)`v3/goal_ui/`\n\n**Try it:** [https://goal.ruv.io/](https://goal.ruv.io/) for goals · [https://goal.ruv.io/agents](https://goal.ruv.io/agents) for live agents. **Run your own:** clone the `goal`\n\nbranch and `cd v3/goal_ui && npm install && npm run dev`\n\n.\n\n``` php\nYour Agent --> [ Remove secrets ] --> [ Sign message ] --> [ Encrypted channel ]\n                 Emails, SSNs,        Proves it came       No one reads it\n                 keys stripped         from you              in transit\n                                                                |\n                                                                v\nTheir Agent <-- [ Block attacks ] <-- [ Check identity ] <------+\n                 Stops prompt          Rejects forgeries\n                 injection\n\n                          Audit trail on both sides.\n                  Trust builds over time. Bad behavior = instant downgrade.\n```\n\nSlack gave teams channels. Federation gives agents the same thing — **shared workspaces across trust boundaries**, where agents on different machines, orgs, or cloud regions can discover each other, prove who they are, and collaborate on tasks.\n\nThe difference: some channels are trusted, some aren't. [ @claude-flow/plugin-agent-federation](https://github.com/ruvnet/ruflo/issues/1669) handles that automatically. Your agents join a federation, get verified via mTLS + ed25519, and start exchanging work — with PII stripped before anything leaves your node and every message auditable. Untrusted agents can still participate at lower privilege: they see discovery info, not your memory. As they prove reliable, trust upgrades. If they misbehave, they get downgraded instantly — no human in the loop required.\n\nYou don't configure handshakes or manage certificates. You `federation init`\n\n, `federation join`\n\n, and your agents start talking. The protocol handles identity, the PII pipeline handles data safety, and the audit trail handles compliance.\n\n📘 Full user guide:[— setup, MCP tools, trust levels, circuit breaker, and the (opt-in) WireGuard mesh layer that ties packet-layer reachability to federation trust. ADR-111 deep-dive at]`docs/federation/`\n\n[.]`docs/federation/phase7-mesh-bringup.md`\n\n**Federation capabilities**\n\n| Capability | How it works | |\n|---|---|---|\n| 🔒 | Zero-trust federation |\nRemote agents start untrusted. Identity proven via mTLS + ed25519 challenge-response. No API keys, no shared secrets. |\n| 🛡️ | PII-gated data flow |\n14-type detection pipeline scans every outbound message. Per-trust-level policies: BLOCK, REDACT, HASH, or PASS. Adaptive calibration reduces false positives. |\n| 📊 | Behavioral trust scoring |\nFormula (`0.4×success + 0.2×uptime + 0.2×threat + 0.2×integrity` ) continuously evaluates peers. Upgrades require history; downgrades are instant. |\n| 📋 | Compliance built-in |\nHIPAA, SOC2, GDPR audit trails as compliance modes. Every federation event produces a structured record searchable via HNSW. |\n| 🤝 | 9 MCP tools + 10 CLI commands |\nFull lifecycle: `federation_init` , `federation_send` , `federation_trust` , `federation_audit` , and more. |\n\n**Example: two teams sharing fraud signals without sharing customer data**\n\n```\n# Team A: initialize federation and generate keypair\nnpx claude-flow@latest federation init\n\n# Team A: join Team B's federation endpoint\nnpx claude-flow@latest federation join wss://team-b.example.com:8443\n\n# Team A: send a task — PII is stripped automatically before it leaves\nnpx claude-flow@latest federation send --to team-b --type task-request \\\n  --message \"Analyze transaction patterns for account anomalies\"\n\n# Team A: check peer trust levels and session health\nnpx claude-flow@latest federation status\n```\n\nSee [issue #1669](https://github.com/ruvnet/ruflo/issues/1669) for the complete architecture, trust model, and implementation roadmap.\n\n```\n# Claude Code plugin\n/plugin install ruflo-federation@ruflo\n\n# Or via CLI\nnpx claude-flow@latest plugins install @claude-flow/plugin-agent-federation\n```\n\n**Claude Code: With vs Without Ruflo**\n\n| Capability | Claude Code Alone | + Ruflo |\n|---|---|---|\n| Agent Collaboration | Isolated, no shared context | Swarms with shared memory and consensus |\n| Coordination | Manual orchestration | Queen-led hierarchy (Raft, Byzantine, Gossip) |\n| Memory | Session-only | HNSW vector memory with sub-ms retrieval |\n| Learning | Static behavior | SONA self-learning with pattern matching |\n| Task Routing | You decide | Intelligent routing (89% accuracy) |\n| Background Workers | None | 12 auto-triggered workers |\n| LLM Providers | Anthropic only | 5 providers with failover |\n| Security | Standard | CVE-hardened with AIDefence |\n\n**Architecture overview**\n\n``` php\nUser --> Claude Code / CLI\n          |\n          v\n    Orchestration Layer\n    (MCP Server, Router, 27 Hooks)\n          |\n          v\n    Swarm Coordination\n    (Queen, Topology, Consensus)\n          |\n          v\n    100+ Specialized Agents\n    (coder, tester, reviewer, architect, security...)\n          |\n          v\n    Memory & Learning\n    (AgentDB, HNSW, SONA, ReasoningBank)\n          |\n          v\n    LLM Providers\n    (Claude, GPT, Gemini, Cohere, Ollama)\n```\n\nFour docs for four audiences:\n\n| Doc | When to read it |\n|---|---|\n|\nSee what currently works — capability counts, test baselines, recent fixes, what's next. The is-it-ready doc. |\n|\nDaily reference — every command, every config flag, every plugin. The how-do-I doc. |\n|\nHow to grade your agent setup, scan tool configs for security, detect changes between runs, and eject a project into a standalone agent toolkit. The audit-my-setup doc. |\n|\nv3.8.0 SOTA matrix vs LangGraph / AutoGen / CrewAI on darwin-arm64 + linux-x64. ruflo wins cold start, single turn, RSS by 1.3×–1953×. The is-it-fast doc. |\n|\nCryptographically prove your installed bytes match the signed witness — `ruflo verify` . The trust-but-verify doc. |\n|\nBefore-merge gates, dual-mode handoff, memory namespace sharing, and witness manifest entry per merge. The safer-team-workflows doc. |\n\nBenchmark internals (for reproduction): [ sota-workload-spec.md](https://github.com/ruvnet/ruflo/blob/perf/sota-comparator-benchmarks/docs/benchmarks/sota-workload-spec.md) ·\n\n[·](https://github.com/ruvnet/ruflo/blob/perf/sota-comparator-benchmarks/docs/benchmarks/SOTA-PROGRESS.md)\n\n`SOTA-PROGRESS.md`\n\n[raw matrix JSON: darwin](https://github.com/ruvnet/ruflo/blob/perf/sota-comparator-benchmarks/docs/benchmarks/sota-matrix.json)·\n\n[linux](https://github.com/ruvnet/ruflo/blob/perf/sota-comparator-benchmarks/docs/benchmarks/sota-matrix-linux.json)\n\nUser Guide section index:\n\n| Section | Topics |\n|---|---|\n|\n\n[Core Features](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#-core-features)[Intelligence & Learning](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#-intelligence--learning)[Swarm & Coordination](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#-swarm--coordination)[Security](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#%EF%B8%8F-security)[Ecosystem](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#-ecosystem--integrations)[Configuration](/ruvnet/ruflo/blob/main/docs/USERGUIDE.md#%EF%B8%8F-configuration--reference)[Plugin Marketplace](https://ruvnet.github.io/ruflo)| Resource | Link |\n|---|---|\n| Documentation |\n|\n\n[GitHub Issues](https://github.com/ruvnet/claude-flow/issues)[ruv.io](https://ruv.io)[Agentics Foundation Discord](https://discord.com/invite/dfxmpwkG2D)[Cognitum.one](https://cognitum.one)MIT - [RuvNet](https://github.com/ruvnet)", "url": "https://wpnews.pro/news/ruflo-an-agent-meta-harness-for-claude-code-and-codex", "canonical_source": "https://github.com/ruvnet/ruflo", "published_at": "2026-07-25 20:12:41+00:00", "updated_at": "2026-07-25 20:22:18.961576+00:00", "lang": "en", "topics": ["ai-agents", "developer-tools", "ai-infrastructure", "ai-tools"], "entities": ["Ruflo", "Claude Code", "Codex", "Cognitum.One", "rUv", "Claude Flow"], "alternates": {"html": "https://wpnews.pro/news/ruflo-an-agent-meta-harness-for-claude-code-and-codex", "markdown": "https://wpnews.pro/news/ruflo-an-agent-meta-harness-for-claude-code-and-codex.md", "text": "https://wpnews.pro/news/ruflo-an-agent-meta-harness-for-claude-code-and-codex.txt", "jsonld": "https://wpnews.pro/news/ruflo-an-agent-meta-harness-for-claude-code-and-codex.jsonld"}}