What do you think about Quality Assurance for the source document, especially finding and removing wrong tags?
I would be happy to remove all of the bugs before the start of translation.
Concerning the question if this should be somehow integrated into CT – no, I do no think so, as there are three different systems with three different word applications involved, and the filter (1) does a pretty good job.
Can you post a screenshot of these wrong tags? Perhaps you can even attach the source file?