Contract Security Scanner — MCP Server

Scan any Base L2 smart contract for security risks directly from your AI assistant.

3 tools exposed:

• scan_contract — Full security scan (source verification, risky selectors, age, activity)
• batch_scan — Compare up to 5 contracts side by side
• interpret_risk — Get an actionable recommendation (SAFE / CAUTION / HIGH_RISK / DO_NOT_USE)

Risk score: 0-100. Analyzes: mint/blacklist/backdoor functions, proxy patterns, source verification, contract age, transaction activity.

Installation

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "contract-scanner": {
      "command": "node",
      "args": ["/Users/sam/Desktop/samDev/p8/mcp/server.js"]
    }
  }
}

Restart Claude Desktop. The tools appear automatically.

Cursor

Add to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global):

{
  "mcpServers": {
    "contract-scanner": {
      "command": "node",
      "args": ["/Users/sam/Desktop/samDev/p8/mcp/server.js"]
    }
  }
}

Cline (VS Code extension)

1. Open Cline settings → MCP Servers → Add server
2. Set type: stdio
3. Command: node /Users/sam/Desktop/samDev/p8/mcp/server.js

Any MCP client (generic)

The server uses stdio transport — just pipe JSON-RPC messages:

node /Users/sam/Desktop/samDev/p8/mcp/server.js

Usage examples

Once connected, just ask your AI assistant naturally:

"Scan this contract before I approve: 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"

"Compare the risk of these 3 Aave clones: 0x... 0x... 0x..."

"Is this token safe to buy? 0x4ed4e862860bed51a9570b96d89af5e1b0efefed"

What gets analyzed

Risk scoring

Technical notes

• Chain: Base L2 only (https://mainnet.base.org)
• API: BaseScan free tier (no key needed for basic checks; set BASESCAN_API_KEY env var for full source analysis)
• No wallet needed: read-only RPC calls only
• Latency: ~2-5s per contract (network dependent)

Built on Base. Agent wallet: 0x804dd2cE4aA3296831c880139040e4326df13c6e