Model Context Protocol moderate risk

Locust

Locust-MCP-Server enables AI-driven load testing by integrating Locust load tests with AI development environments via the Model Context Protocol.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwriteexec

Quickstart

Config

{
  "mcpServers": {
    "locust": {
      "command": "/Users/naveenkumar/.local/bin/uv",
      "args": [
        "--directory",
        "/Users/naveenkumar/Gits/locust-mcp-server",
        "run",
        "locust_server.py"
      ]
    }
  }
}

Exposed MCP Tools (1)

high

run_locust

Runs a Locust load test with configurable parameters like host, users, and runtime.

Executes arbitrary Python code from the specified test file and can potentially exhaust resources.

Safety Assessment

The Locust-MCP-Server is relatively safe when used with trusted test files and properly configured environment variables. However, running untrusted test files poses a significant risk due to arbitrary code execution. It's crucial to limit resource usage to prevent denial-of-service scenarios.

Requires explicit test file specification, limiting scope.
Uses a defined API with parameters for test execution.
Can be run in headless mode, reducing UI exposure.
Test parameters (users, spawn rate, runtime) are configurable.

Executes arbitrary Python code from the specified test file.
Relies on environment variables for configuration, which can be insecure if not managed properly.
Potential for resource exhaustion if test parameters are set too high.
Accesses the network based on the host specified in the test or configuration.