3.
Pretty much. 1 or 2 specials is easy, but if there's many specials, with naming diverging from what's in the database, then chances of things working out go down. You can post file names and test cases, maybe there's a way to improve things. Maybe not though.

EDIT: That's for TV specials. Movie extras are not supported at all.
4.
Loose matches typically are highlighted in red.
5.
You can process any file in "generic file mode" like so:
viewtopic.php?f=3&t=2072
This can be useful to pre-process files to make auto-detection work better.
6.
If you have movies, then extras aren't supported at all. But each extra, if not ignored, will liked be matched to the movie, as if it's a multi-part movie. With some format trickery, you could process extras even though internally filebot will have miss-identified the file (correct movie, but incorrect CD part, which we just ignore in our format).
This is somewhat related, not as copy & paste solution, but in concept:
viewtopic.php?f=8&t=10628&p=43115#p43105