In regards to the tolerances I get the feeling with the PCBs it could be the PCBs themselves, I've noticed when it comes to both plates and PCBs some designers prefer to use a reduced spec so their nominal hole values are less than the switches nominal speced values but within tolerance to help make things more snug on Cherrys, the problem being when you add tolerances onto the reduced nominal value they could go under spec, pair that with Gateron using a tight fit PCB pin vs. Cherrys loose fit (I'm speculating on that but it seems to be the case) and you have issues with Gateron but not Cherry even though both are within spec.
The cap catching problem is tough to say what part is the issue because I've never seen a spec on that part of the switch but again it could possibly be the cap makers tolerance instead of the switch, the reason I tend to think this is I have seen a post where one cap would work fine on any switch but another from the same set would catch on any switch suggesting a variance in the caps not the switches. Gaterons have a visibly smaller, if only slightly (~0.3mm), clearance on the LED side of the switch but the above example tends to suggest their tolerances are fairly consistent even though they seem to have tighter nominal fits all round. I would be interested to hear from others with catching caps if they catch on all switches or if they catch on some but not others as that would help narrow down if its a cap tolerance issue paired with tighter fitting switch dimensions or if it is in fact looser tolerances on the switches themselves.
So far all the caps I've tried (granted not a lot) don't catch and the PCBs I've designed with measurements taken directly from the Cherry/Gateron (they are the same) metric spec fit tightly. I'm still very happy with the switches and if I do run into these issues I'm glad its a fairly simple fix but if they are actually issues with the switches tolerances it would be a shame as it would be a good reason for mass producers to overlook using them.
My 2c
Edit: After digging a bit more it looks like the PCB pins on the Gaterons have a ~1.5mm core with three ~0.15mm "wings", the hole specs are 1.7-0.05mm (Cherry) 1.7+/-0.05mm (Gateron) which would suggest the Gaterons "wings" are in fact designed to be press fit onto the PCB being the pins have a maximum cross-section of 1.65mm and a interference circumference of 1.8mm requiring a 0.025-0.075mm compression on the wings. I don't have any PCB mount Cherrys to compare.
For reference the stem clearance on Cherrys are ~6.65mm while on Gaterons it's ~6.35mm so with what I regard as a standard stem (5.5mm) there shouldn't be any issues but thicker ones at 6mm or larger might have issues depending on how much wobble comes into play.
I would say my measurements are reasonably accurate but there may be some error due to the sharp edges of my callipers cutting into the softer plastic and in the case of the stem clearance the thin walls may have flexed a small amount (I was as careful as I could be and took measurements from several switches for comparison).