Hello 
,
You could possibly save the post ID into the Data Store at the end of the scenario flow and, at the same time, perform a Data Store search before the creation of a comment.
The comment would then be created only in case the Data Store wouldn’t find a specific post ID in the database.
The above flow should then ensure ChatGPT creates only one comment per post.
I hope it helps.
Cheerio 