MCP · AI editors

The security scanner
you can talk to from inside your editor.

Start a scan. Get findings. Apply AI-generated fixes. Verify with a re-scan — all without leaving Cursor, Claude Code, or Copilot Chat.

Generate API keyQuickstart

Works with your favorite AI editor

Config path: ~/.cursor/mcp.json

{
  "mcpServers": {
    "security-scanner": {
      "command": "npx",
      "args": [
        "-y",
        "@security-scanner/mcp-server"
      ],
      "env": {
        "SCANNER_API_KEY": "<YOUR_API_KEY>"
      }
    }
  }
}

The AI-Native DevSecOps Loop

  1. 1

    Scan

    Type "scan example.com" in your AI editor. MCP starts a security scan against 20 modules.

  2. 2

    Findings

    The AI editor renders the findings as a chat-renderable markdown table — severity, OWASP mapping, modules.

  3. 3

    Propose fix

    Type "fix <finding_id>". Our AI generates a minimal code patch (cached for re-use).

  4. 4

    Apply + verify

    Type "apply <finding_id>". The affected module re-runs and confirms whether the vulnerability is gone.

All tiers can use scan + status + findings. Pro adds AI fix generation. Business adds retest verification.

See full pricing →