Reptor

You deploy this MCP server to expose the reptor CLI’s pentest reporting and automation capabilities as a programmable service. It lets other tools or AI agents interact with SysReptor through the MCP protocol, enabling automated workflows and richer integrations while keeping the core reptor tooling intact.

How to use

Connect an MCP client to the server to access reptor’s plugins and the custom tools. You can invoke dynamically generated tools derived from reptor plugins as well as the dedicated tools like list_findings, get_finding_details, and upload_template. Operate entirely through the MCP API to automate tasks, run analyses, and retrieve findings in structured data formats. Because this server is designed for local use, you should run it within a trusted environment and avoid exposing it to untrusted networks.

How to install

Prerequisites you need before installing: Python 3.9 or newer and a Python environment manager such as uv. You also need access to the reptor CLI project, which should be available alongside this MCP server.

Steps to set up the environment and install dependencies:

Configuration

Configure the MCP server using environment variables that connect to your SysReptor instance and control secure behavior. The following variables are used by the underlying reptor integration.

REPTOR_SERVER: URL of your SysReptor instance (required).

REPTOR_TOKEN: SysReptor API token (required).

REPTOR_PROJECT_ID: Default project ID for operations (optional).

REPTOR_MCP_INSECURE: Disable SSL certificate verification for self-signed setups (optional).

REQUESTS_CA_BUNDLE: Path to a custom CA bundle file for SSL verification (optional).

REPTOR_MCP_DEBUG: Enable verbose debug logging from the MCP server (optional).

Running the server

Run the server in a way that allows programmatic access over HTTP. Use the FastMCP-based command with a streamable-http transport and expose it on a local port.

The recommended start command is:

fastmcp run mcp_server.py:mcp --transport streamable-http --port 8008

"The server will be accessible at http://localhost:8008/mcp/".

## Client connection

Configure your MCP client to connect to the server using the HTTP transport. Point the client to the MCP URL exposed by the server.

## Architecture notes

This MCP server acts as a dynamic wrapper around reptor’s CLI, exposing its capabilities as MCP tools. It uses FastMCP to provide a robust and scalable interface. Key components include a dynamic tool generator that maps reptor plugins to MCP tools and a wrapper that handles argument translation, stdin redirection, and output parsing.

## Security and best practices

There is no built-in authentication or authorization in this MCP server. It is intended for local use only. Do not expose the MCP server directly to the internet or untrusted networks. When handling sensitive project data, be mindful of data transfers to external services or AI models through clients connected to the MCP server.

## Troubleshooting and tips

If you encounter connection issues, verify that the server is running on the expected port and that your MCP client is pointed to the correct URL. Check environment variable values for SysReptor access and SSL settings. Review logs for any authentication or network errors and confirm that the reptor CLI is accessible to the MCP wrapper.

## Available tools

### list\_findings

Lists findings for a project with optional filters by status, severity, and title.

### get\_finding\_details

Retrieves the full JSON object for a specific finding by its ID.

### upload\_template

Uploads a new finding template from a JSON or TOML string.