when using a word that is against policy, just sending a msg and not pointing out which word it is is impossible.
in long prompt it's just trying to guess which word is problemsitic.
and removing them automatically and not telling which word also effects the prompts in a way that completetly ruin it.
can just highlight in red the problemistic words.
also no where is it limited or warns about max amount of words per prompt. so maybe when getting to the last 20 words add a text that makes the user aware.
thanks for the access, I'll share more if needed. 🙂
Tal