1. We have used the LIMIT keyword in our project in a couple of queries where it was essential (specific example; a query which returns a playlist of a specific size from a selection of choices which exceed that number). I think that as long as the LIMIT keyword doesn't compromise the query's complexity it should be OK to use it.
2. We are required to submit the crawler sources and an SQL script which can repopulate the DB offline. I don't recall having read that we need to perform any data collection on the remote servers... So I do join your question here! :)