Sibilants and plosives can sound bad enough in your mix, but they will be highly exaggerated in the mastering process, especially due to the excessive limiting most people demand in order for their music to be loud. And Lord knows, being loud is what's important these days above all else. Yes, the mastering engineer can help these to some degree, but the mastering lab is not the proper place for the these to be addressed. The proper place is while the vocals are being cut. Of course the tracking engineer can resort to de-essers and notch filters, but before any of that is used they should first work with mic selection. Many condenser microphones, which is what most recordists will grab first for vocals, have a certain amount of a presence peak. Sometimes this is due to the nature of the design of the mic, but often it's a trick being played by the manufacturer, especially in cheap condensers, to give the the false impression that the mic has a wonderful 'air' and brilliance, when instead it's just an exaggerated top end that you could simulate with eq. So if you use one of these mics on a vocalist who already has exaggerated sibilance in their voice, you have a mess on your hands. The lazy recordist, if they're aware of the problem, will just throw a de-esser and/or notch filter on the track and be done with it. But as i've said before, you should strive to use as little processing as possible in every step of your recording and mixing. So taking a little time to get up and down and trying different mics to find the one that at least doesn't exaggerate the problem is the proper thing to do. Ribbon mics most often have a creamier, smoother top end than condensers, and also don't ignore quality dynamics like the Shure SM-7 (used on many or most of Michael Jackson's vocals, for instance) or the Sennheiser 441 (used on TomPetty and Stevie Nicks, for instance). Then even if you do still have to use the processors, you're using them less.
The plosives are caused, as I said, by aspiration hitting the diaphragm of the mic, so that is a simpler matter to fix by just working with the placement of the mic first so the breath blast doesn't hit the diaphragm, and then using the good ol' pantyhose pop filter if that doesn't completely solve it.
If the problems didn't get addressed during tracking, the next preferable fix would be during the mix. If it is just pure sibilance, and you are working with a ProTools type digital recording system, it's best to take the time to find the sibilant in the vocal track, make an edit on either side of it, and then lower its volume. But the problem is that sibilants have such extreme energy that in some cases they will also cause a splatter of high frequency distortion from either overdriving the mic preamp or the a/d converter or both. In that case you have a much longer duration splat than than the actual sibilant itself was, and you'll have to resort to some de-essing or 'spot' eq'ing.
If you are working with a vocalist who has extreme sibilance (I certainly hope my wife's friend doesn't sing or do voice work. If so, I pity the poor engineer), then the best thing to do (after mic selection, which won't solve all of the problem in these cases, but will help) is to add a little de-essing and/or filtering while you're cutting the vocal, and then a little more while you're mixing. This will be MUCH more transparent than trying to fix it all in one step or the other.
As I said at the top, your mastering engineer will have de-essers and filters to add to the mix to at least help control these problems, but that's the worst solution, as he will not be applying these to just the vocal but instead across the entire mix. And if the vocal has had a considerable amount of reverb added to it, the mastering engineer will be able to kill some of the initial attack of the sibilant or plosive, but the trail in the reverb will still be there, which might draw even more attention to problem than just leaving it alone. If it had been properly addressed during tracking and mixing, the reverb trails will be clean.
