Feature Request: Optional Prompt Optimization/Review Layer Before Execution Model

What variant of Codex are you using?

CLI

What feature would you like to see?

I’d like to request a feature to add prompt checking in Codex. The idea is that when a user writes a long prompt, we could first use a lightweight model like Codex Mini to review the prompt and provide suggestions or improvements. After reviewing the feedback, if the user agrees, there would be a button to send the refined prompt to the main Codex model that will actually execute the task.

I think this is a relatively small addition, but it could be very useful. It would help ensure that the model responsible for executing the command doesn’t have to work as hard to interpret the intent, and can instead focus more directly on the execution or implementation of the request.

Additional information

_No response_