Add policy enforcement for editor and project execution tools

[L1AD]

This server gives agents control over the Godot editor:

• run_project -- launches Godot projects
• create_scene -- creates scene files on disk
• execute_in_editor -- runs arbitrary commands in the editor

execute_in_editor with arbitrary input can modify project state. run_project launches executables. An agent loop creating scenes can fill disk.

Suggestion: Consider documenting support for PolicyLayer/Intercept, an open-source MCP proxy that enforces policies on tool calls before execution.

Example policy:

version: "1"
default: allow

tools:
  execute_in_editor:
    rules:
      - action: require_approval

  run_project:
    rules:
      - rate_limit: 5/hour

  create_scene:
    rules:
      - rate_limit: 10/hour

One line to set up: npx -y @policylayer/intercept init