Model Context Protocol moderate risk

K6

Executes k6 load tests via MCP, allowing LLMs to orchestrate and analyze performance testing. Requires k6 and Python environment.

Connections & Capabilities

Connects To

GitHub

Capabilities

exec

Quickstart

Config

{
  "mcpServers": {
    "k6": {
      "command": "/path/to/bin/uv",
      "args": [
        "--directory",
        "/path/to/k6-mcp-server",
        "run",
        "k6_server.py"
      ]
    }
  }
}

Exposed MCP Tools (2)

moderate

execute_k6_test

Runs a k6 test with a specified script file, using default duration and VUs.

Executes arbitrary k6 scripts, potentially consuming resources.

moderate

execute_k6_test_with_options

Runs a k6 test with a specified script file, duration, and number of virtual users.

Executes arbitrary k6 scripts with user-defined parameters, potentially consuming excessive resources.

Safety Assessment

The k6-mcp-server is relatively safe for executing load tests if the k6 scripts are trusted and resource limits are carefully managed. Untrusted scripts or excessive resource allocation could pose risks. Ensure the k6 binary is from a trusted source.

Requires explicit script file path, limiting accidental execution.
Execution is contained within the k6 runtime.
No direct system-level access beyond k6's capabilities.
Test duration and VUs can be controlled.

Potential for resource exhaustion if VUs and duration are set too high.
k6 scripts can execute arbitrary JavaScript code, posing a risk if the scripts are malicious.
Server requires access to the k6 binary, which could be compromised.
Lack of built-in input validation could lead to unexpected behavior.