turbo-flow-claude

作者 marcuspat已验证

Advanced Agentic Development Environment Supporting Devpods, Rackspace Spot Instances, Github Codespaces, Google Cloud Shell, and more! Features 600+ AI agents, Claude Flow, SPARC methodology, and automatic context loading! Deploy intelligent multi-agent swarms, coordinate autonomous workflows.

124
Stars
35
Forks
Shell
语言
2026/8/23
添加时间

⚠️ 第三方软件声明

本 Skill 为第三方开源软件,独立托管于 GitHub。SkillTip 仅为信息目录,不控制或维护底层仓库。所显示的安全检查为自动化且范围有限,安装前请自行审查源码。

阅读服务条款

安装

添加到你的 Claude Code skills 目录:

# Add to your Claude Code skills
git clone https://github.com/marcuspat/turbo-flow-claude

快速入门

使用 turbo-flow-claude 等 Skills 的指南。

安全报告

已验证

上次扫描:—

{
  "status": "PASSED",
  "issues": []
}

README.md

Turbo Flow v4.1.0 — Ruv Stack Edition

Version Ruflo Ruv Stack MCP Tools Plugins License Adventure Wave Labs

Complete Agentic Development Environment — Ruflo v3.5 + Beads + Worktrees + Agent Teams + Ruv Stack

Built & Presented by Adventure Wave Labs

Quick StartInstallationRuv StackPluginsCommandsMigrationResources


🎬 Demo

Turbo Flow demo

A real, unedited run of the full onboarding chain — codespace_setup.shsetup.shpost-setup.shtmux-workspace.sh (11 install steps, 6-plugin verify: Claude Code, Ruflo v3.5, Dolt + Beads, GitNexus, OpenSpec) — sped up, then a real reattach to the live tmux session it creates: a tour through all four workspace windows (Claude-2, Claude-Monitor, a live htop), then the real claude CLI launched through to its actual login screen. Not a mockup, not a scripted animation — no API key on the recording box, so that's the genuine first-run auth screen. Recorded with VHS from demo/demo-a-setup.tape + demo/demo-b-tmux-claude.tape (demo/build-demo.sh reproduces it end to end).


About Adventure Wave Labs

Adventure Wave Labs

Adventure Wave Labs is the team behind Turbo Flow, a complete agentic development environment built for the Claude ecosystem. We design, build, and maintain the packaging, setup automation, plugin curation, and workflow integration that brings together orchestration, memory, codebase intelligence, and agent isolation into a single streamlined install. The orchestration core itself — the 215+ MCP tools and 60+ agents referenced throughout this README — is Ruflo v3.5, built by ruvnet ("rUv"); Turbo Flow integrates and configures it rather than building it from scratch. See the Ruv Stack Edition section below for the full breakdown of what's original to this repo versus wired in from rUv's ecosystem.


Ruv Stack Edition (v4.1.0)

Turbo Flow's orchestration core has always run on ruvnet's ("rUv") ecosystem — Ruflo v3.5 and a bundled RuVector WASM build. This edition wires in three more real ruvnet projects and gives the rest of the branded "ruv stack" an honest accounting: what's dev-relevant gets integrated, what isn't gets documented and skipped.

ModuleWhat it addsStatus
RuVector (deepened)npm install ruvector — installs the optional package AGENTDB-VS-RUVECTOR.md already documents but leaves undone by defaultDeepened
ruVLLMLocal/offline sparse-attention model tier — a $0, no-API rung below HaikuNew
MetaHarnessPer-repo branded agent-harness generator (npx metaharness)New
SkygraphBrowser all-sky radar demo (ADS-B + satellites)Optional, off by default
RuView, rUv NeuralReal ruvnet projects (WiFi sensing, EEG/BCI) — no dev-tooling integration pointExcluded
PhotonLayerNo public repo found under ruvnetNot real / roadmap only

Full breakdown, package names, and rationale: docs/RUV-STACK.md.

./scripts/setup-ruv-stack.sh            # dry run — prints the plan
./scripts/setup-ruv-stack.sh --apply    # installs RuVector+, ruVLLM, MetaHarness

Pure add-on: it never modifies devpods/setup.sh, runs independently of the main install, and writes its rv-* aliases to a separate file from the core rf-* set.


What's New in v4.0.0

Metricv3.4.1v4.0.0Change
Installation Steps1510-5 (consolidated)
Core Packages4 separate1 (Ruflo)-75%
MCP Tools175+215++23%
Agents60+60+
Plugins156-9 (redundant removed)
Cross-session MemoryNoneBeadsNew
Agent IsolationNoneGit WorktreesNew
Codebase GraphNoneGitNexusNew
UI/UX SkillYesYesKept
StatuslineYesYesUpdated to 4.0

Major Changes

  • claude-flow → Ruflo v3.5 — Single npx ruflo@latest init replaces 4 separate installs
  • Beads — Cross-session project memory via git-native JSONL
  • GitNexus — Codebase knowledge graph with MCP server and blast-radius detection
  • Native Git Worktrees — Per-agent isolation with auto PG Vector schema namespacing
  • Native Agent Teams — Anthropic's experimental multi-agent spawning
  • 6 focused plugins — 9 redundant/domain-specific plugins removed
  • OpenSpec — Spec-driven development kept
  • UI UX Pro Max — Design skill kept
  • Statusline Pro v4.0 — Updated with TF 4.0 branding

Removed (redundant with Ruflo v3.5 or out of scope)

  • claude-flow@alpha, @ruvector/cli, @ruvector/sona, @claude-flow/browser → bundled in Ruflo
  • 9 plugins: healthcare-clinical, financial-risk, legal-contracts, cognitive-kernel, hyperbolic-reasoning, quantum-optimizer, neural-coordination, prime-radiant, ruvector-upstream
  • Claudish, Agentic Jujutsu, Spec-Kit, agtrace, PAL MCP → bundled or redundant
  • HeroUI + Tailwind + TypeScript scaffold → out of scope
  • Ars Contexta, OpenClaw Secure Stack → out of scope

Architecture

+------------------------------------------------------------------+
|              TURBO FLOW v4.1.0 — Adventure Wave Labs              |
+------------------------------------------------------------------+
|  INTERFACE                                                        |
|  +---------------+  +---------------+  +---------------+          |
|  | Claude Code   |  |  Open WebUI   |  |  Statusline   |          |
|  |     CLI       |  |  (4 instances)|  |   Pro v4.0    |          |
|  +---------------+  +---------------+  +---------------+          |
+------------------------------------------------------------------+
|  ORCHESTRATION: Ruflo v3.5                                        |
|  60+ Agents | 215+ MCP Tools | Auto-activated Skills              |
|  AgentDB v3 | RuVector WASM | SONA | 3-Tier Model Routing        |
|  59 MCP Browser Tools | Observability | Gating                    |
+------------------------------------------------------------------+
|  RUV STACK (optional add-on — scripts/setup-ruv-stack.sh)         |
|  RuVector crate (deepened) | ruVLLM local tier | MetaHarness      |
|  Skygraph (opt-in, off by default)                                |
+------------------------------------------------------------------+
|  PLUGINS (6)                                                      |
|  +--------------------------------------------------------------+ |
|  | Agentic QE | Code Intel | Test Intel | Perf | Teammate | Gas | |
|  +--------------------------------------------------------------+ |
+------------------------------------------------------------------+
|  CODEBASE INTELLIGENCE: GitNexus                                  |
|  Knowledge Graph | Blast Radius Detection | MCP Server            |
+------------------------------------------------------------------+
|  MEMORY (Three-Tier)                                              |
|  +---------------+  +---------------+  +---------------+          |
|  |    Beads      |  | Native Tasks  |  |   AgentDB     |          |
|  |  project/git  |  |   session     |  |  + RuVector   |          |
|  |    JSONL      |  |  ~/.claude/   |  |  WASM accel   |          |
|  +---------------+  +---------------+  +---------------+          |
+------------------------------------------------------------------+
|  ISOLATION                                                        |
|  Git Worktrees per Agent | PG Vector Schema per Worktree          |
|  Auto GitNexus Indexing | Agent Teams (experimental)              |
+------------------------------------------------------------------+
|  SKILLS                                                           |
|  UI UX Pro Max | OpenSpec | 36+ Ruflo Auto-activated Skills       |
+------------------------------------------------------------------+
|  INFRASTRUCTURE                                                   |
|  DevPod | Codespaces | Rackspace Spot                            |
+------------------------------------------------------------------+

Quick Start

DevPod Installation

macOS
brew install loft-sh/devpod/devpod
Windows
choco install devpod
Linux
curl -L -o devpod "https://github.com/loft-sh/devpod/releases/latest/download/devpod-linux-amd64"
sudo install devpod /usr/local/bin

Launch

# DevPod (recommended)
devpod up https://github.com/marcuspat/turbo-flow --ide vscode

# Codespaces
# Push to GitHub → Open in Codespace → runs automatically

# Manual
git clone https://github.com/marcuspat/turbo-flow -b main
cd turbo-flow
chmod +x devpods/setup.sh
./devpods/setup.sh
source ~/.bashrc
turbo-status

# Optional: Ruv Stack Edition (RuVector+, ruVLLM, MetaHarness)
./scripts/setup-ruv-stack.sh --apply

What Gets Installed

The devpods/setup.sh script installs the complete stack in 10 automated steps:

Step 1: System Prerequisites

PackagePurpose
build-essentialC/C++ compiler (gcc, g++, make)
python3Python runtime
gitVersion control
curlHTTP client
jqJSON processor (required for statusline)
Node.js 20+JavaScript runtime (required by Ruflo v3.5)

Step 2: Claude Code + Ruflo v3.5

ComponentPurpose
Claude CodeAnthropic's agentic coding CLI
Ruflo v3.5Orchestration engine — replaces claude-flow@alpha
Ruflo MCPRegistered as MCP server in Claude Code
Ruflo DoctorAuto-diagnostic and fix pass

Ruflo v3.5 bundles: AgentDB v3, RuVector WASM, SONA, 215 MCP tools, 60+ agents, skills system, 3-tier model routing, 59 browser automation MCP tools, observability, gating

Step 3: Ruflo Plugins (6) + OpenSpec

PluginPurpose
Agentic QE58 QE agents — TDD, coverage, security scanning, chaos engineering
Code IntelligenceCode analysis, pattern detection, refactoring suggestions
Test IntelligenceTest generation, gap analysis, flaky test detection
Perf OptimizerPerformance profiling, bottleneck detection
Teammate PluginBridges Native Agent Teams with Ruflo swarms (21 MCP tools)
Gastown BridgeWASM-accelerated orchestration, Beads sync (20 MCP tools)
OpenSpecSpec-driven development (independent npm package)

Step 4: UI UX Pro Max Skill

ComponentPurpose
uipro-cliDesign system skill — component patterns, accessibility, responsive layouts, design tokens

Step 5: GitNexus (Codebase Knowledge Graph)

ComponentPurpose
GitNexusIndexes dependencies, call chains, execution flows
GitNexus MCPRegistered as MCP server — blast-radius detection

Step 6: Beads (Cross-Session Memory)

ComponentPurpose
@beads/bdGit-native JSONL project memory — issues, decisions, blockers

Step 7: Workspace + Agent Teams

ComponentPurpose
Directoriessrc/ tests/ docs/ scripts/ config/ plans/
Agent TeamsCLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 enabled

Step 8: Statusline Pro v4.0

3-line statusline with 15 components:

LINE 1: [Project] name | [Model] Sonnet | [Git] branch | [TF] 4.0 | [SID] abc123
LINE 2: [Tokens] 50k/200k | [Ctx] #######--- 65% | [Cache] 42% | [Cost] $1.23 | [Time] 5m
LINE 3: [+150] [-50] | [READY]

Step 9: CLAUDE.md Generation

Generates workspace context file with:

  • 3-tier memory protocol (Beads → Native Tasks → AgentDB)
  • Isolation rules (one worktree per agent)
  • Agent Teams rules (max 3 teammates, recursion depth 2)
  • Model routing tiers (Opus/Sonnet/Haiku)
  • Plugin reference
  • Cost guardrails ($15/hr)

Step 10: Aliases + Environment + MCP Registration

50+ aliases across families: rf-*, ruv-*, mem-*, bd-*, wt-*, gnx-*, aqe-*, os-*, hooks-*, neural-*, turbo-status, turbo-help

Optional: Ruv Stack Edition (scripts/setup-ruv-stack.sh)

Run after the 10 steps above, independently, any time:

StepComponentPurpose
1RuVector (deepened)npm install ruvector — the package AGENTDB-VS-RUVECTOR.md documents as optional
2ruVLLMLocal/offline sparse-attention model tier (cargo install ruvllm)
3MetaHarnessPer-repo branded agent-harness generator (npx metaharness, run on demand)
4SkygraphOptional, off by default — pass --with-skygraph to include

Adds a fourth alias family, rv-*, in its own file — never touches ~/.turboflow_aliases. See docs/RUV-STACK.md.


Plugins (6)

PluginMCP ToolsPurpose
Agentic QE1658 QE agents, TDD, coverage, security, chaos engineering
Code IntelligenceCode analysis, patterns, refactoring
Test IntelligenceTest generation, gaps, flaky tests
Perf OptimizerProfiling, bottlenecks, optimization
Teammate Plugin21Agent Teams ↔ Ruflo swarms bridge, semantic routing
Gastown Bridge20WASM orchestration, Beads sync, convoys

Removed Plugins (9)

PluginReason
healthcare-clinicalDomain-specific (HIPAA/FHIR) — not needed
financial-riskDomain-specific (PCI-DSS/SOX) — not needed
legal-contractsDomain-specific — not needed
cognitive-kernelRedundant with Ruflo's neural system
hyperbolic-reasoningRedundant with RuVector WASM hyperbolic embeddings
quantum-optimizerRedundant with Ruflo's EWC++ and RVFOptimizer
neural-coordinationRedundant with Ruflo's swarm coordination
prime-radiantNiche — mathematical interpretability
ruvector-upstreamRedundant — RuVector bundled in Ruflo v3.5

Key Commands

Status & Help
turbo-status         # Check all components
turbo-help           # Complete command reference
rf-doctor            # Ruflo health check
rf-plugins           # List installed plugins
Orchestration (Ruflo)
rf-wizard            # Interactive setup
rf-swarm             # Hierarchical swarm (8 agents max)
rf-mesh              # Mesh swarm
rf-ring              # Ring swarm
rf-star              # Star swarm
rf-spawn coder       # Spawn a coder agent
rf-daemon            # Start background workers
rf-status            # Ruflo status
Memory
bd-ready             # Check project state (session start)
bd-add               # Record issue/decision/blocker
bd-list              # List beads
ruv-remember K V     # Store in AgentDB
ruv-recall Q         # Query AgentDB
mem-search Q         # Search ruflo memory
mem-stats            # Memory statistics
Isolation
wt-add agent-1       # Create worktree for agent
wt-remove agent-1    # Clean up worktree
wt-list              # Show all worktrees
wt-clean             # Prune stale worktrees
Quality & Testing
aqe-generate         # Generate tests (Agentic QE plugin)
aqe-gate             # Quality gate
os-init              # Initialize OpenSpec in project
os                   # Run OpenSpec
Intelligence
hooks-train          # Deep pretrain on codebase
hooks-route          # Route task to optimal agent
neural-train         # Train neural patterns
neural-patterns      # View learned patterns
gnx-analyze          # Index repo into knowledge graph
gnx-serve            # Start local server for web UI
gnx-wiki             # Generate repo wiki from graph
Ruv Stack (optional add-on)
rv-vector-init         # npm install ruvector — see AGENTDB-VS-RUVECTOR.md for when to use it
rv-llm-serve           # Start local ruVLLM inference server
rv-llm-route "task"    # Route a task to the local, $0 ruVLLM tier
rv-harness-gen         # npx metaharness — scaffold a branded harness for this repo
rv-sky-demo            # Optional Skygraph demo (only if installed with --with-skygraph)

Migrating from v3.x

  1. Your old cf-* aliases are gone — use rf-* instead
  2. Slash commands (/sparc, etc.) are gone — Ruflo auto-activates skills
  3. Run bd init in your project repos to enable Beads memory
  4. Run npx gitnexus analyze in your repos to build the knowledge graph
  5. The v3/ directory preserves everything — nothing was deleted
v3.4.1v4.0.0
cf-initrf-init
cf-swarmrf-swarm
cf-doctorrf-doctor
cf-mcpAutomatic via rf-wizard
mem-searchmem-search (unchanged)
cfb-openVia Ruflo's bundled browser MCP tools
No cross-session memorybd-ready, bd-add
No isolationwt-add, wt-remove
No codebase graphgnx-analyze

Repository Structure

turbo-flow/
├── V3/                          ← archived v3.0-v3.4.1 (Claude Flow era)
├── .claude/                     ← skills, agents, settings
├── devpods/
│   ├── setup.sh                 ← main setup script
│   ├── post-setup.sh            ← post-setup verification
│   └── context/                 ← devpod context files
├── scripts/
│   ├── generate-claude-md.sh
│   └── setup-ruv-stack.sh       ← optional Ruv Stack Edition add-on
├── docs/
│   └── RUV-STACK.md             ← ruv stack breakdown: what's real, what's wired in
├── CLAUDE.md                    ← workspace context (active)
└── README.md

Post-Setup

# 1. Reload shell
source ~/.bashrc

# 2. Verify installation
turbo-status

# 3. Get help
turbo-help

# 4. Run post-setup verification (13 checks)
# ./devpods/post-setup.sh

# 5. Optional: install the Ruv Stack Edition
# ./scripts/setup-ruv-stack.sh --apply

Version History

VersionDateChanges
v4.1.0Jul 2026Ruv Stack Edition: RuVector (deepened crate), ruVLLM (local model tier), MetaHarness (per-repo harness generator), Skygraph (optional). Pure add-on — devpods/setup.sh untouched.
v4.0.0Mar 2026Ruflo Migration: Ruflo v3.5, Beads, GitNexus, Worktrees, Agent Teams, 6 plugins, UI UX Pro Max, OpenSpec
v3.4.1Feb 2025Fixes: skill install removed, plugins command, npm fallback
v3.4.0Feb 2025Complete + Plugins: 36 skills, 15 plugins
v3.3.0Feb 2025Complete installation: 41 skills, memory, MCP
v3.0.0Feb 2025Initial release with Claude Flow V3

Resources


License

MIT — Copyright (c) 2025-2026 Adventure Wave Labs


Built & Presented by Adventure Wave Labs

Turbo Flow v4.1.0 — Ruflo v3.5 + Ruv Stack Edition. 215+ MCP tools. 6 plugins. Beads. GitNexus. Worktrees. RuVector+. ruVLLM. MetaHarness. One command.

https://github.com/marcuspat/turbo-flow/blob/main/AWLabs.png

Ecosystem

Built with and powers these tools — star the ones you use:

RepoWhat it does
codescopeSingle-binary code-intelligence engine for AI agents — no cloud, no DB, no Python
secret-scanRust secret scanner — obfuscation detection, up to ~1,200 files/sec measured
SentinelDeny-by-default agentic sysadmin: Investigate → Plan → Approve → Act in Rust
spacelift-intentNatural language → cloud infra via Terraform providers, no HCL generated
netrainMatrix-style network monitor in Rust

Stay Connected

If Turbo Flow ships value for you, follow @marcuspat on GitHub — agentic tooling, Rust crates, and open-source infra drop regularly.

Follow @marcuspat

常见问题

What is turbo-flow-claude?

turbo-flow-claude is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by marcuspat. Advanced Agentic Development Environment Supporting Devpods, Rackspace Spot Instances, Github Codespaces, Google Cloud Shell, and more! Features 600+ AI agents, Claude Flow, SPARC methodology, and automatic context loading! Deploy intelligent multi-agent swarms, coordinate autonomous workflows. It has 124 GitHub stars.

Is turbo-flow-claude safe to use?

Yes. turbo-flow-claude passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.

How do I install turbo-flow-claude?

Clone the repository with "git clone https://github.com/marcuspat/turbo-flow-claude" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is turbo-flow-claude written in?

turbo-flow-claude is primarily written in Shell. It is open-source under marcuspat on GitHub, so you can review or fork the full source.

Are there alternatives to turbo-flow-claude?

Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh turbo-flow-claude against similar tools.

评论 (0)

暂无评论,成为第一个分享想法的人!

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情

claude-code

by anthropics

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

120,03119,897Shell
AI 智能体
查看详情

开发者还喜欢

基于喜欢此 Skill 的开发者投票和收藏

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

n8n

by n8n-io

12

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

201,88160,308TypeScript
MCP 服务器apisai-tools
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情
turbo-flow-claude — Claude Code AI Skill | SkillTip