Flow Nexus MCP Server

Advanced Model Context Protocol (MCP) server with E2B sandbox deployment, Claude-Flow orchestration, and real-time Supabase integration. Features 50+ tools for multi-agent swarm coordination, gamified app store, and enterprise-grade AI workflows.

Created by ruv - https://neural-trader.ruv.io

Features

• 🤖 50+ MCP Tools - Complete toolset for AI orchestration
• 🦠 Multi-Agent Swarms - Hierarchical, mesh, ring, and star topologies
• 📦 Enhanced E2B Sandboxes - Deploy with environment variables and API keys
• 🧠 Claude Code Integration - Native support for Anthropic's Claude Code
• ⚡ Real-time Supabase - Live data streaming and persistence
• 🎮 Gamified App Store - rUv credits, challenges, and achievements
• 🚀 Neural Pattern Training - WASM-accelerated AI optimization
• 🤖 DAA Autonomous Agents - Decentralized autonomous agent coordination
• ✅ 100% Success Rate - Bulletproof error handling

🆕 Enhanced Sandbox Features

• Environment Variables - Pass API keys and configuration to sandboxes
• Claude Code Template - Pre-configured Claude Code environments
• Multi-Language Support - JavaScript, Python, Bash, Claude Code commands
• Package Installation - Automatic npm/pip package installation on creation
• Startup Scripts - Run initialization commands after sandbox creation
• Runtime Configuration - Update sandbox settings and environment variables
• Custom E2B Keys - Use your own E2B API keys for sandboxes

Quick Start

Installation

npm install -g @flow-nexus/daa-wasm-mcp

Initialize Configuration

daa-wasm-mcp init

Start Server

daa-wasm-mcp start

Configuration

Environment Variables

Create a .env file with the following variables:

# MCP Server Configuration
MCP_HOST=localhost
MCP_PORT=3001
NODE_ENV=development

# Supabase Configuration
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_SERVICE_KEY=your-service-key
SUPABASE_ANON_KEY=your-anon-key

# Authentication
JWT_SECRET=your-jwt-secret-key

# Redis Configuration (optional)
REDIS_URL=redis://localhost:6379

# WASM Configuration
WASM_PATH=./dist/daa-agents.wasm
WASM_MEMORY_LIMIT=256MB
WASM_THREAD_POOL_SIZE=4
WASM_SIMD_ENABLED=true

# Agent Configuration
MAX_CURATOR_AGENTS=3
MAX_PRICING_AGENTS=2
MAX_SECURITY_AGENTS=4
MAX_RECOMMENDATION_AGENTS=2

Configuration File

Alternatively, use a JSON configuration file:

{
  "host": "localhost",
  "port": 3001,
  "supabase": {
    "url": "https://your-project.supabase.co",
    "serviceKey": "your-service-key"
  },
  "wasm": {
    "wasmPath": "./dist/daa-agents.wasm",
    "memoryLimit": "256MB",
    "simdEnabled": true,
    "threadPoolSize": 4
  },
  "auth": {
    "jwtSecret": "your-jwt-secret"
  },
  "cors": {
    "origins": ["http://localhost:5173"]
  }
}

DAA Agent Types

Curator Agent

• Purpose: Automated application quality assessment
• Functions: assess_quality, batch_review, train_model
• Configuration: Quality thresholds, auto-approval settings

Pricing Agent

• Purpose: Dynamic pricing analysis and optimization
• Functions: analyze_pricing, update_market_data, optimize_price
• Configuration: Price elasticity, market cap, volatility thresholds

Security Agent

• Purpose: Security vulnerability scanning and compliance
• Functions: scan_repository, analyze_vulnerabilities, generate_report
• Configuration: Scan depth, CVE database, parallel scan limits

Recommendation Agent

• Purpose: Personalized application recommendations
• Functions: generate_recommendations, update_user_profile, train_embedding
• Configuration: Model type, embedding dimensions, update frequency

MCP Tools

Agent Management

• daa_agent_spawn - Spawn new DAA agents
• daa_agent_list - List active agents
• daa_agent_metrics - Get agent performance metrics
• daa_agent_terminate - Terminate specific agents

Agent Operations

• daa_agent_execute - Execute functions on agents
• daa_agent_train - Train ML models within agents

Application Operations

• daa_assess_quality - Assess application quality
• daa_analyze_pricing - Analyze optimal pricing
• daa_security_scan - Perform security scans
• daa_generate_recommendations - Generate recommendations

Real-time Coordination

The server automatically responds to real-time events from Supabase:

• New Applications: Triggers quality assessment and security scanning
• rUv Transactions: Updates market data for pricing optimization
• App Installations: Updates recommendation models
• Challenge Submissions: Analyzes code quality and provides feedback

API Endpoints

Health Check

GET /health

Metrics (Prometheus format)

GET /metrics

MCP Resources

• daa://agents - List of active agents
• daa://metrics - Performance metrics
• daa://health - System health status

CLI Commands

Start Server

daa-wasm-mcp start [options]

Options:

• -p, --port <port> - Port to listen on (default: 3001)
• -h, --host <host> - Host to bind to (default: localhost)
• --config <file> - Configuration file path
• --env <file> - Environment file path

Initialize Configuration

daa-wasm-mcp init [--force]

Health Check

daa-wasm-mcp health [--url <url>]

Get Metrics

daa-wasm-mcp metrics [--url <url>] [--format json|prometheus]

Generate Test Token

daa-wasm-mcp generate-token --user-id <id> [options]

Authentication

The server uses JWT authentication with the following permissions structure:

User Levels

• Level 1-2: Basic operations (read-only, quality assessment)
• Level 3-4: Agent management (spawn, execute, train)
• Level 5+: Administrative operations (terminate agents, system metrics)

JWT Token Structure

{
  "sub": "user-id",
  "username": "username",
  "developer_level": 3,
  "aud": "claude-agents",
  "iss": "flow-cloud-app-store"
}

Rate Limiting

Default rate limits (per hour):

• daa_agent_list: 1000 requests
• daa_assess_quality: 100 requests
• daa_agent_spawn: 20 requests
• daa_agent_train: 10 requests

Configure custom limits via Redis or in-memory store.

Monitoring

Built-in Metrics

• Agent performance (tasks completed, execution time, error rate)
• System metrics (CPU, memory, uptime)
• Real-time throughput and success rates
• Anomaly detection for high error rates and low throughput

Prometheus Integration

curl http://localhost:3001/metrics

Health Monitoring

curl http://localhost:3001/health

🧪 Enhanced Sandbox Examples

Create Claude Code Sandbox

// Create sandbox with Claude Code pre-installed
const sandbox = await server.executeTool('sandbox_create', {
  template: 'claude-code',
  name: 'my-claude-workspace',
  anthropic_key: process.env.ANTHROPIC_API_KEY,
  startup_script: 'claude --version && echo "Claude Code ready!"'
});

// Execute Claude Code commands
await server.executeTool('sandbox_execute', {
  sandbox_id: sandbox.sandbox_id,
  code: 'claude "Help me debug this code"',
  language: 'claude-code',
  env_vars: {
    ANTHROPIC_API_KEY: process.env.ANTHROPIC_API_KEY
  }
});

Multi-API Development Sandbox

// Create sandbox with multiple API integrations
const sandbox = await server.executeTool('sandbox_create', {
  template: 'node',
  name: 'multi-api-workspace',
  env_vars: {
    ANTHROPIC_API_KEY: process.env.ANTHROPIC_API_KEY,
    OPENAI_API_KEY: process.env.OPENAI_API_KEY,
    GITHUB_TOKEN: process.env.GITHUB_TOKEN,
    SUPABASE_URL: process.env.SUPABASE_URL,
    SUPABASE_ANON_KEY: process.env.SUPABASE_ANON_KEY
  },
  install_packages: ['@anthropic/sdk', 'openai', '@supabase/supabase-js'],
  timeout: 7200
});

Configure Existing Sandbox

// Add environment variables to existing sandbox
await server.executeTool('sandbox_configure', {
  sandbox_id: 'existing-sandbox-id',
  env_vars: {
    NEW_API_KEY: 'your-new-key',
    DEBUG: 'true'
  },
  anthropic_key: process.env.ANTHROPIC_API_KEY,
  install_packages: ['@anthropic/claude-code'],
  run_commands: [
    'claude --version',
    'npm list -g --depth=0'
  ]
});

Python AI Development

// Python sandbox for AI/ML development
const aiSandbox = await server.executeTool('sandbox_create', {
  template: 'python',
  name: 'ai-research-env',
  env_vars: {
    ANTHROPIC_API_KEY: process.env.ANTHROPIC_API_KEY,
    HUGGINGFACE_TOKEN: process.env.HUGGINGFACE_TOKEN
  },
  install_packages: ['anthropic', 'transformers', 'torch', 'numpy'],
  startup_script: 'python --version && pip list'
});

// Execute Python AI code
await server.executeTool('sandbox_execute', {
  sandbox_id: aiSandbox.sandbox_id,
  code: `
import anthropic
import os

# Verify API key is available
api_key = os.getenv('ANTHROPIC_API_KEY')
print(f"Anthropic API configured: {'Yes' if api_key else 'No'}")

# Initialize client
client = anthropic.Anthropic(api_key=api_key)
print("Anthropic client initialized successfully!")
  `,
  language: 'python'
});

Development

Prerequisites

• Node.js 18+
• TypeScript 5+
• Supabase project
• Redis (optional, for rate limiting)

Setup

git clone <repository>
cd daa-wasm-mcp
npm install
npm run build

Testing

npm test
npm run test:watch

Development Mode

npm run dev

Docker Deployment

FROM node:18-alpine
WORKDIR /app
COPY package*.json ./
RUN npm ci --only=production
COPY . .
RUN npm run build
EXPOSE 3001
CMD ["npm", "start"]

NPX Usage

Deploy the server instantly:

npx @flow-cloud/daa-wasm-mcp init
npx @flow-cloud/daa-wasm-mcp start

Performance

• WASM SIMD: Up to 4x performance improvement for ML operations
• Multi-threading: Parallel agent execution
• Memory Efficiency: Configurable memory limits
• Auto-scaling: Dynamic agent spawning based on workload
• Connection Pooling: Efficient database connections

Security

• JWT-based authentication
• Role-based permissions
• Rate limiting protection
• Input validation with Zod schemas
• CORS configuration
• Environment-based secrets

Troubleshooting

Common Issues

1. WASM Module Not Found

   • Ensure WASM_PATH points to a valid WASM file
   • Check file permissions

2. Supabase Connection Failed

   • Verify SUPABASE_URL and SUPABASE_SERVICE_KEY
   • Check network connectivity

3. JWT Authentication Failed

   • Ensure JWT_SECRET is set
   • Verify token format and expiration

4. Rate Limit Exceeded

   • Check Redis connection
   • Adjust rate limits in configuration

Debug Mode

LOG_LEVEL=debug daa-wasm-mcp start

Contributing

1. Fork the repository
2. Create a feature branch
3. Make changes and add tests
4. Submit a pull request

License

MIT License - see LICENSE file for details

Support

• GitHub Issues: GitHub Repository Issues
• Documentation: Full Documentation
• Discord: Flow Cloud Community