In a word: crucial.

Pro-Q 2 and Pro-L are two of my most-used plugins. All of FabFilter's stuff is super high quality

I use a free plug-in called A1StereoControl

+1 for A1. The key is to only ever use small amount, don't get carried away or yes it will start to sound hollow, and god forbid it get collapsed to mono on a different system it will sound like everything has disappeared.

Yeah that's a pretty brutal scenario to work with. You'd probably want to just go for full absorption across the board (as opposed to a mix with diffusion) to eliminate all the odd reflections but the trade-off is you're going to end up with a very dead sounding room.

As a mastering engineer here is how I prefer to have mixes prepared before receiving them:

No limiting, clipping, or EQ on the master whatsoever. Compression is fine, and encouraged if it's integral to the overall tone of your track but only if it's done properly and not smashed or pumping all over the place.

Short answer: compression can be ok, nothing else. If you're unsure about your compression then take it off.

Format-wise: .wav or .aiff uncompressed. 24-bit or 32-bit float is preferred (especially if intending to do a 'Mastered for iTunes' release) but 16-bit isn't the end of the world. Sample-rate should be whatever rate it was mixed at, never upsampled.

Other often overlooked things:

"Tips & Tails" - aka the lead in and fade out of your track. If you've started your track with a big transient (i.e. a kick or a crash cymbal) right on beat 1 of bar 1, some DAWs will cut it off slightly when it renders, so check that it hasn't been cut off, or provide a bit of silence at the beginning of the track to be safe. Similarly at the very end of the track make sure you haven't cut off the reverb tail or anything like that, give it a little extra room to breathe because it can easily be edited and faded back down in mastering.

Snap, Crackle, Pop - not necessarily distortion or clipping, but watch out for the occasional click or pop from a buffer error, or from a piece of audio/sample that has not been cut at a zero-crossing. A good mastering engineer will notice these and have tools to remove/repair them as transparently as possible on the master, but fixing it in the mix will ALWAYS be have better results for these things.

I think people got used to all kinds of "stereo knobs" so much they forgot the wonders of plain simple panning.

hahaha 100% true. I'm coming from the mastering perspective, so by that point panning is more or less out of the question, but in the mix panning is your best friend.

I run my main parametric EQ in mid/side so the image can be tweaked throughout the frequency spectrum. Similar to your multiband method but with the benefit of being able to adjust the filter slopes as well.

As much as people say never ever clip on the master, that's not entirely true. Clipping is a tool like anything else, there's a time to use it for a desired result, and if abused has some pretty serious consequences in terms of distortion and artifacts.

That said, clipping is an extremely routine process for me when mastering, whether it be with a plugin or clipping my converters going from my analog chain back into the digital domain. A limiter will (usually) push transients down in order to prevent distortion, and the result is loss of impact. Clipping on the other hand tends to retain a better perception of those original transients (I say perception, because they're not there anymore, but they've been cut off rather than pushed down) at the cost of distortion. Most of my masters will have small amounts of gain using both methods to get a balance.

Some tracks just don't have enough transient response to begin with and would sound completely flat with a limiter, so in those cases hard clipping can be better, if the instrumentation is capable of masking small amounts of distortion. You just need to be extra careful, because the tiniest bit too far and it's basically ruined.