securityscorecard-mcp

SSC MCP Server

A community-built, comprehensive Model Context Protocol (MCP) server that integrates with the SecurityScorecard API. It runs over stdio, so it works with any MCP-compatible client — Claude Desktop, Claude Code, Cursor, VS Code, and others. It serves MCP protocol revision 2026-07-28 and stays compatible with 2025-era clients.

Published on npm as @callmarcus/securityscorecard-mcp and listed in the MCP Registry as io.github.CallMarcus/securityscorecard-mcp.

Disclaimer: This is an independent, community-built open-source project. It is not affiliated with, endorsed by, sponsored by, or associated with SecurityScorecard, Inc. in any way. It is built solely against SecurityScorecard's publicly available API documentation. "SecurityScorecard" and all related names, marks, and logos are trademarks of SecurityScorecard, Inc. and are used here for identification purposes only. You must supply your own API credentials and comply with SecurityScorecard's terms of service.

Quick Start

Prerequisites

1. Node.js 20+ - Download
2. SecurityScorecard API Token - Get from your SecurityScorecard dashboard

Option A — Install from npm (recommended)

No clone or build required. The server runs over stdio via npx, so any MCP-compatible client can launch it. npx -y always fetches the latest published version.

Most clients — Claude Desktop, Cursor, Cline, Windsurf, and others — share the same mcpServers JSON. Add this block to the client's MCP config:

{
  "mcpServers": {
    "security-scorecard": {
      "command": "npx",
      "args": ["-y", "@callmarcus/securityscorecard-mcp"],
      "env": {
        "SECURITY_SCORECARD_API_TOKEN": "your-api-token-here",
        "COMPANY_DOMAIN": "example.com"
      }
    }
  }
}

Where that config file lives:

Client	Config file
Claude Desktop (Windows)	%APPDATA%\Claude\claude_desktop_config.json
Claude Desktop (macOS)	~/Library/Application Support/Claude/claude_desktop_config.json
Cursor	~/.cursor/mcp.json (global) or .cursor/mcp.json (project)

Replace the credentials with your own, then restart the client.

Claude Code — add it from the CLI instead:

claude mcp add security-scorecard \
  --env SECURITY_SCORECARD_API_TOKEN=your-api-token-here \
  --env COMPANY_DOMAIN=example.com \
  -- npx -y @callmarcus/securityscorecard-mcp

On Windows, wrap the launcher in cmd /c: ... -- cmd /c npx -y @callmarcus/securityscorecard-mcp.

VS Code (Copilot) — uses a servers key with an explicit type, in .vscode/mcp.json:

{
  "servers": {
    "security-scorecard": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@callmarcus/securityscorecard-mcp"],
      "env": {
        "SECURITY_SCORECARD_API_TOKEN": "your-api-token-here",
        "COMPANY_DOMAIN": "example.com"
      }
    }
  }
}

Option B — Run from source (for development)

# Clone the repository
git clone https://github.com/CallMarcus/security-scorecard-mcp.git
cd security-scorecard-mcp

# Install dependencies
npm install

# Build (use build:fast to avoid memory issues)
npm run build:fast

Then point your MCP client at the local build. For clients that use the mcpServers format (Claude Desktop, Cursor, …):

{
  "mcpServers": {
    "security-scorecard": {
      "command": "node",
      "args": ["/path/to/security-scorecard-mcp/build/index.js"],
      "env": {
        "SECURITY_SCORECARD_API_TOKEN": "your-api-token-here",
        "COMPANY_DOMAIN": "example.com"
      }
    }
  }
}

Important: Replace the path and credentials with your actual values, then restart your MCP client. (For Claude Code, run claude mcp add security-scorecard --env SECURITY_SCORECARD_API_TOKEN=your-api-token-here -- node /path/to/security-scorecard-mcp/build/index.js.)

Available Tools

The server (index.js) provides 9 specialized tools:

Tool	Purpose
security_dashboard	Score, grade, and key security metrics
analyze_security_risks	Issue prioritization and risk analysis
create_improvement_plan	Actionable remediation roadmaps
discover_assets	Asset inventory with security context
analyze_email_security	SPF/DMARC/DKIM analysis
api_discovery	Search 535 API endpoints with hybrid semantic/keyword search
analyze_issue_types	Granular issue type breakdowns
validate_data_completeness	Cross-tool data verification
query_security_data	Direct API access with discovery

Response Modes

Each tool supports three response modes for token efficiency:

• minimal - Quick answers (15-50 tokens)
• standard - Overview with context (200-300 tokens)
• detailed - Comprehensive analysis (800+ tokens)

Environment Variables

Variable	Required	Description
SECURITY_SCORECARD_API_TOKEN	Yes	Your API token
COMPANY_DOMAIN	No	Default domain for queries
DEBUG_MODE	No	Set true for verbose logging

Optional rate limiting and caching:

REQUEST_CACHE_TTL_MS=300000
REQUESTS_PER_INTERVAL=5
REQUEST_INTERVAL_MS=1000

API Discovery

The server includes hybrid search (semantic + keyword) for finding SecurityScorecard API endpoints:

Use api_discovery to search for "email security"

This searches 535 indexed endpoints and returns matching paths with confidence scores, required parameters, and curl examples.

To update the API reference after changes:

npm run api:embed    # Regenerate semantic embeddings
npm run api:update   # Regenerate docs + embeddings

Development

Build Commands

npm run build:fast   # Recommended - uses esbuild (~130ms)
npm run build        # TypeScript compiler (may OOM on some systems)
npm test             # Run tests

Project Structure

src/
  index.ts               # MCP server (9 tools)
  api/client.ts          # SecurityScorecard API client
  integration/           # API discovery system
docs/api/                # Self-contained API reference
  index.jsonl            # Endpoint index (535 endpoints)
  index-embeddings.json  # Semantic search embeddings
build/                   # Compiled JavaScript

Testing

npm test             # Run test suite

Troubleshooting

Build fails with out of memory

Use the fast build instead:

npm run build:fast

"Cannot find module" errors

Reinstall dependencies:

rm -rf node_modules
npm install
npm run build:fast

Semantic search degrades to keyword-only (Windows + WSL)

Install for the platform that runs the server. Claude Desktop on Windows launches the server with Windows node, so if npm install ran under WSL the native modules (onnxruntime-node, sharp) only have linux binaries — the embeddings layer fails to load and api_discovery silently degrades to keyword-only search (results still come back, but confidence scoring is cruder). Run npm install && npm run build:fast from PowerShell or cmd in the repo directory instead — or keep two clones, one per platform.

Your client doesn't see the server

1. Double-check the config file location for your client (see Quick Start)
2. For a from-source install, verify the path to build/index.js is correct
3. Restart the client completely
4. Sanity-check that the server starts on its own: npx -y @callmarcus/securityscorecard-mcp (it should launch and wait silently on stdio)

API returns 401 Unauthorized

Your API token is invalid or expired. Get a new one from SecurityScorecard dashboard.

License

MIT

Links

• SecurityScorecard API Docs
• Model Context Protocol
• Report Issues