Steampunk Spotter MCP Server

The Steampunk Spotter MCP server connects Steampunk Spotter scanning abilities with AI agents, so they can check the code as they write. It uses the Steampunk Spotter CLI, which is installed together with it.

Installation

Requirements:

• uv for the recommended install, or
• Python >= 3.10 for other install options (see below).

Add the following entry to your MCP configuration (see Agent setup for agent-specific instructions):

{
  "mcpServers": {
    "spotter": {
      "command": "uvx",
      "args": ["--python", ">=3.10", "steampunk-spotter-mcp@latest"]
    }
  }
}

Alternative install

To avoid uv dependency, use any python package manager to install the server into your Python environment:

$ pip install steampunk-spotter-mcp

This provides the spotter-mcp console script. Use it as the command, with the absolute path to it:

{
  "mcpServers": {
    "spotter": {
      "command": "/path/to/.venv/bin/spotter-mcp"
    }
  }
}

Authentication

Recommended: login with Spotter CLI

Log in once and the MCP server uses the stored credentials (see CLI authentication for all login options).

$ spotter login

Without the CLI installed, log in through uvx:

$ uvx --from steampunk-spotter spotter login

For on-prem, first set the endpoint in ~/.config/steampunk-spotter/spotter.json (add the key if the file already exists), then log in:

{"endpoint": "<spotter-api-url>"}

For details, see Setting the API endpoint in the CLI documentation.

Alternative: set env variables

The MCP server authenticates with these environment variables, which take precedence over the stored login:

Option 1: set env variables in your MCP config

Set them in the env block of your user-level MCP configuration. Do not put them in a project-level config that is committed to version control.

"env": {
  "SPOTTER_TOKEN": "<your-token>",
  "SPOTTER_ENDPOINT": "<spotter-api-url>"
}

Option 2: set env variables in your shell

Most agents pass their own environment to the MCP server, so you can leave env out and instead:

• export the variables in your shell profile (e.g. ~/.profile), or
• set them per project with direnv (.envrc).

Start the agent from a shell where the variables are set.

Agent setup

The examples use the recommended uvx install. Add env only if you authenticate with environment variables.

Each agent can register the server in one of two scopes:

• User: available in all your projects. Use this scope if you put credentials in env.
• Project: usually a .mcp.json file in the repository root, shared with your team through version control. Never put credentials here.

Claude Code

• User:

  $ claude mcp add --scope user spotter -- uvx --python ">=3.10" steampunk-spotter-mcp@latest
  

• Project: add the JSON entry to .mcp.json, or run the command above with --scope project.

GitHub Copilot (VS Code and Copilot CLI)

• User: add the JSON entry to ~/.copilot/mcp-config.json.
• Project: add it to .mcp.json.

Acknowledgment

This tool was created by XLAB Steampunk, IT automation specialists and leading experts in building Enterprise Ansible Collections.