Mimecast MCP Server

A Model Context Protocol (MCP) server for Mimecast email security. Enables AI assistants to track messages, investigate threats, manage email queues, and access threat intelligence data.

This is a Model Context Protocol (MCP) server that connects Claude (or any MCP-compatible AI) to your Mimecast environment.

Part of the MSP Claude Plugins ecosystem — a growing suite of AI integrations for the MSP stack. Built by MSPs, for MSPs.

Installation

npm install @wyre-ai/mimecast-mcp

Configuration

Copy .env.example and set the variables below.

/health and /healthz stay unauthenticated and do not read credentials. The stdio transport does not use CONDUIT_S2S_SECRET.

HTTP authentication

With MCP_TRANSPORT=http and CONDUIT_S2S_SECRET set, every /mcp request must include X-Gateway-S2S (t=<unix seconds>,v1=<hex hmac-sha256 of t=...>). A missing or invalid header returns 401.

In AUTH_MODE=gateway, /mcp also requires X-Mimecast-Client-ID and X-Mimecast-Client-Secret (X-Mimecast-Region is optional). Those headers are the only credential source for that request.

Usage

Running with Claude Desktop

Add to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "mimecast-mcp": {
      "command": "npx",
      "args": ["@wyre-ai/mimecast-mcp"],
      "env": {
        "MIMECAST_CLIENT_ID": "your-mimecast-client-id",
        "MIMECAST_CLIENT_SECRET": "your-mimecast-client-secret",
        "MIMECAST_REGION": "us"
      }
    }
  }
}

Running with Claude Code (CLI)

claude mcp add mimecast-mcp \
  -e MIMECAST_CLIENT_ID=your-value \
  -e MIMECAST_CLIENT_SECRET=your-value \
  -e MIMECAST_REGION=your-value \
  -- npx -y @wyre-ai/mimecast-mcp

Docker

The image listens on 0.0.0.0 inside the container (MCP_HTTP_HOST) and defaults to AUTH_MODE=gateway. Publish the port on loopback and pass the S2S secret. In gateway mode, Mimecast credentials arrive on each request as headers; they are not read from the environment.

docker build -t mimecast-mcp .
docker run \
  -e CONDUIT_S2S_SECRET=your-s2s-secret \
  -p 127.0.0.1:8080:8080 \
  mimecast-mcp

Compose refuses to start when CONDUIT_S2S_SECRET is unset, defaults AUTH_MODE to gateway, and publishes 127.0.0.1:8080:8080:

CONDUIT_S2S_SECRET=your-s2s-secret docker compose up

Single-tenant HTTP (AUTH_MODE=env) still requires the S2S secret, then uses MIMECAST_CLIENT_ID and MIMECAST_CLIENT_SECRET:

docker run \
  -e CONDUIT_S2S_SECRET=your-s2s-secret \
  -e AUTH_MODE=env \
  -e MIMECAST_CLIENT_ID=your-value \
  -e MIMECAST_CLIENT_SECRET=your-value \
  -e MIMECAST_REGION=us \
  -p 127.0.0.1:8080:8080 \
  mimecast-mcp

Local HTTP without a secret is development-only. It binds loopback (127.0.0.1 unless MCP_HTTP_HOST is localhost or ::1) and logs a warning. A non-loopback MCP_HTTP_HOST makes startup exit:

MCP_TRANSPORT=http MCP_ALLOW_INSECURE_DEV=1 npm run start:http

API application permissions

Create an API 2.0 application in the Mimecast Administration Console (Services → API and Platform Integrations) and assign it a custom administrator role. The client id and client secret from that application are MIMECAST_CLIENT_ID and MIMECAST_CLIENT_SECRET. API 1.0 applications (application id, application key, access key, and secret key with HMAC signing) are not accepted.

Grant only the permissions each tool calls. A permissions error from mimecast_get_queue_status or mimecast_get_threat_incidents means the role is missing the row below. An empty list from audit or TTP logs means the call was authorized and nothing matched the window.

mimecast_get_audit_events also requires a start and end. When the caller omits them, the server queries the last 7 days. Mimecast only keeps 60 days of audit history, and category filters must be codes such as account_logs or policy_logs (omit categories to search every log). Basic Administrator includes the read permissions in Mimecast's own integration guides; Threat Remediation and Held Edit are separate and must be added explicitly.

Available Domains

Messages

Message tracking and investigation

Queue

Email queue management

Threats

Threat intelligence and detection data

Interactive Message Card (MCP Apps)

mimecast_get_message_info renders as an interactive, read-only card in MCP Apps hosts (Claude Desktop/web) showing delivery status, sender/recipients, spam score, and rejection details; plain-JSON behavior is unchanged in other hosts. The card is neutral by default, brandable via window.__BRAND__ injection or MCP_BRAND_* env vars (MCP_BRAND_NAME, MCP_BRAND_LOGO_URL, MCP_BRAND_PRIMARY_COLOR, MCP_BRAND_ACCENT_COLOR, MCP_BRAND_BG, MCP_BRAND_TEXT) — no rebuild needed.

Development

# Clone the repository
git clone https://github.com/WYRE-AI/mimecast-mcp.git
cd mimecast-mcp

# Install dependencies
npm install

# Build
npm run build

# Run tests
npm test

Contributing

Contributions are welcome! Please see CONTRIBUTING.md if present, or open an issue to discuss changes.

License

Licensed under the Apache License, Version 2.0. See LICENSE for details.