Sometimes my prompt are not very specific and lose some details which cause bad code generation. If there could be a indicator to evaluate my prompt and give recommendations (a simple poor/fair/good indicator regarding my prompt text, and few lines of suggestions).
I find using ChatGPT and letting it know you’re talking to cursor and you’re trying to achieve something, ChatGPT will often tell you what to say to Cursor and the results are pretty good.
I second this. I also often ask it which model it recommends, if it thinks MAX is worth the extra cost for the request, and I ALWAYS ask it to opine if it thinks I’m wrong about something or missed something important.