Heimdall

Overview

Heimdall is an MCP server that scans MCP server and agent configurations to detect injection vulnerabilities, exfiltration risks, and dangerous capabilities. It brings security analysis into your AI workflow by exposing callable tools that Claude, Cursor, and other MCP clients can invoke directly, eliminating manual back-and-forth between separate tools.

The server runs locally via npx -y mcp-heimdall-scan with no authentication required. It communicates over stdio as an npm package and requires no external credentials or setup. Version v0.5.1 is released under the MIT license, with full source code available at caglarbozkurt/mcp-heimdall on GitHub.

Heimdall appears in the AI & Agents category on mcp.site and works with all standard MCP clients including Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, and Zed. If you maintain this project, claim your listing to verify ownership, display the Verified badge, and keep information up to date.

Installation

// claude_desktop_config.json
{
  "mcpServers": {
    "heimdall": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-heimdall-scan"
      ]
    }
  }
}