Mallory MCP Server

Mallory provides a robust source of cyber and threat intelligence. This MCP server exposes the Mallory API to AI agents via the malloryapi Python client, with tools for vulnerabilities, threat actors, malware, exploits, organizations, attack patterns, breaches, products, advisories, stories, mentions, search, and sources.

Once connected, your AI assistant (Cursor, Claude Desktop, or another MCP client) can look up CVEs, threat actors, malware, and more directly from Mallory — no copy-pasting from the dashboard.

Prerequisites

• Python 3.11 or higher
• A Mallory API key (mallory.ai)

Quick Start

1. Set your API key

Get an API key at mallory.ai and add it to your shell profile (~/.zshrc, ~/.bashrc, etc.):

export MALLORY_API_KEY=your_api_key_here

Reload your shell (or run source ~/.zshrc) so the variable is available.

2. Add to your AI client

Add the server to your MCP client config. Pick one of the options below.

Cursor — add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "Mallory": {
      "command": "uvx",
      "args": ["mallorymcp"]
    }
  }
}

Claude Desktop — add to claude_desktop_config.json:

{
  "mcpServers": {
    "Mallory": {
      "command": "uvx",
      "args": ["mallorymcp"]
    }
  }
}

Claude Code — run this command:

claude mcp add --transport stdio Mallory -- uvx mallorymcp

This stores the config in ~/.claude.json (local scope, current project). To share it with your team, use project scope instead:

claude mcp add --transport stdio --scope project Mallory -- uvx mallorymcp

This writes to .mcp.json in the project root, which can be committed to git.

uvx downloads and runs the package automatically — no install step needed. If you prefer to install it yourself, see Alternative: pip install below.

3. Restart your AI client and start using it

Ask your assistant to query Mallory:

• "Look up CVE-2024-1234 and summarize the risk."
• "List threat actors trending in the last 7 days."
• "Find vulnerabilities that are known to be exploited."
• "Search for intelligence on APT28."
• "What malware is associated with technique T1566?"

The assistant calls the MCP tools automatically — you don't need to invoke tool names yourself.

Note: mallorymcp is an MCP server that communicates via JSON-RPC over stdio. It's designed to be launched by your AI client, not run interactively from a terminal.

Alternative: pip install

If you prefer installing the package rather than using uvx:

pip install mallorymcp

Then reference the command directly in your config:

{
  "mcpServers": {
    "Mallory": {
      "command": "mallorymcp"
    }
  }
}

Configuration

Tools

The server exposes the following tools, backed by the Mallory API.

Vulnerabilities (7)

Threat Actors (5)

Malware (5)

Exploits (2)

Organizations (4)

Attack Patterns (4)

Breaches (3)

Products (3)

Advisories (3)

Stories (3)

Mentions (3)

Search and Sources (2)

Development

Install from source

git clone https://github.com/malloryai/mallorymcp.git
cd mallorymcp
uv sync
uv run mallorymcp

Lint

uv sync --extra lint
uv run ruff check src/ tests/
uv run ruff format src/ tests/

Project Structure

src/mallorymcp/
├── __init__.py
├── _version.py          # Auto-generated by hatch-vcs from git tags
├── app.py               # Entry point (main, stdio transport)
├── config/              # Env-based config (MALLORY_API_KEY, MALLORY_BASE_URL)
├── decorator/           # API error handling for tools
├── server/              # FastMCP server and tool loader
├── tools/               # Tool modules (one per resource area)
└── utils/               # Serialization, debug

Releasing

1. Tag a release: git tag v0.4.0 && git push --tags
2. Create a GitHub release from the tag
3. GitHub Actions builds and publishes to PyPI via trusted publisher

License

Apache 2.0.