run_python

What run_python does on Code MCP Server

AI agents invoke run_python to trigger actions in Code MCP Server. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call: builds kicked off, notifications sent, workflows started.

Why run_python is rated High

This tool allows execution of arbitrary Python code within the VS Code workspace environment. While Python execution itself is not inherently destructive, the blast radius is significant because: (1) the tool can read/write files, access environment variables, make network calls, and invoke system commands depending on the Python code executed; (2) an AI agent could be tricked into running malicious or unintended…

From the tool's definition Tool name 'run_python' combined with description 'Execute Python code or script' directly indicates execution of arbitrary code. The tool is explicitly designed to run Python code whose effects depend entirely on the arguments provided.

The rule that runs run_python safely

PolicyLayer is an MCP gateway: it sits between your AI agents and Code MCP Server, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For run_python, this is the rule to start with:

{
  "version": "1",
  "default": "deny",
  "tools": {
    "run_python": {
      "limits": [
        {
          "counter": "run_python_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

Questions about run_python