🟢 LIVE 🛡️ MCP Security

DeFi Shield MCP

Protect your AI agents from malicious DeFi transactions.

DeFi Shield is an MCP server that lets Claude, Cursor, and VS Code check token approvals, address risk, and transaction safety before you sign anything.

🚀 Server is live

Connect your AI agent to the public endpoint:

https://defi-shield-mcp.lumen-models.workers.dev/mcp

What it detects

⚠️ Unlimited token approvals
🚨 Blacklisted addresses (honeypots)
❌ Dangerous function calls in calldata
🔍 Suspicious address patterns

How to use it

VS Code

Add this to your .vscode/mcp.json:

{
  "servers": {
    "defi-shield": {
      "type": "http",
      "url": "https://defi-shield-mcp.lumen-models.workers.dev/mcp"
    }
  }
}

Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "defi-shield": {
      "url": "https://defi-shield-mcp.lumen-models.workers.dev/mcp"
    }
  }
}

Cursor

Add this to your ~/.cursor/mcp.json:

{
  "mcpServers": {
    "defi-shield": {
      "url": "https://defi-shield-mcp.lumen-models.workers.dev/mcp"
    }
  }
}

Example

Ask your agent:

"Check if approving unlimited USDC to 0xdead... is safe"

{
  "risk_level": "CRITICAL",
  "risk_score": 90,
  "risks_found": [
    "⚠️ UNLIMITED APPROVAL",
    "🚨 SUSPICIOUS ADDRESS"
  ],
  "recommendation": "DO NOT SIGN this transaction."
}

Tech stack

☁️ Cloudflare Workers

📘 TypeScript

🔌 MCP SDK

⚡ 0ms cold start

💰 Free tier

🌍 Global edge