1. The filter is more to protect the supply (and other devices) than the VFD.... they're noisy buggers.
2, Correct
3. Closed-loop makes no difference. Whether you're driving high or low is largely irrelevant (the two pins are the anode/cathode of an internal LED used for isolation - so switching either works). What this gives you though is flexibility - for example an open-collector (open drain) drive is designed for low-side switching (i.e. take the + to supply, and switch on the -). The axbb manual describes the non-isolated outputs as being able to push/pull 20mA - so can source (+ve) or sink (-ve) - so can drive either input.
4. Your call. It's better to cut to length but not if you risk damaging anything.
5. Yes. Unequivocally. Do I?, I might.
6. I only glanced at your schematic - it feels a safe design but more than I'd do for a personal machine. I can't knock safe design but I can duck an errant shock 50% of the time.
7. It's a good idea. Otherwise a stray wire in the spindle could short to the enclosure and zap you. Of course your protective earth will then ensure that you brick your VFD before your blow the protective devices.
8. No. Unconnected, the LED opto-isolator remains inert and the drivers enabled.
9. It's a good idea, but realistically will achieve little. You'll be generating pulses at maybe 100kHz at around 5mA, that will generate some EMF but I'd hazard a guess that you're not going to impact much. The driver inputs going through an opto isolator will be pretty tolerant of any induced emf.
10. Yes, in fact I have insulated ferrules that accept two cores (i.e. a wide rectangular receptacle rather than round), Use common sense - if the cores fit comfortably then there's nothing wrong. Test each crimp.
11. I'd go for the permanent supply, otherwise you can end up with pull-lows on outputs when you don't intend this (remember, you still have a 0V supply to the AXBB, it's capable of hard-driving low). But remember I've not really looked at your schematic.