Could OpenAI provide a visual policy tree for configuring parent and child agents in ChatGPT Work and Codex, including cloud workflows where comparable controls aren’t available? Existing custom-agent configuration already supports useful model and reasoning…