itsabhisingh07/LINKDIN_POST_GENERATOR ? reverse-engineered prompt

Build me a simple Python tool that creates polished LinkedIn posts from a short topic or idea. I want it to use a multi step AI flow where one part writes the post, another part checks it, and another part improves it until the post is good enough or it reaches a reasonable retry limit.

The post should feel professional and ready to publish on LinkedIn. Keep it within a sensible character limit, avoid emojis, and make sure the tone sounds clean and business appropriate. If the first draft does not pass the checks, rewrite it based on the feedback and keep going automatically.

I want to be able to run it locally, enter a topic, and see the final approved post clearly in the output. If possible, also show the evaluation result for each round so it is easy to understand what changed. Use the Groq API key from an env file for the model setup. Look up current docs online if you need to.