Hand 1

Overpairs don't make great bluffraising combos because they only block flushes, which are close to indifferent to call vs a theoretically balanced shoving range -the only viable exception is the nut flush because it's not losing to higher flushes. You really want to block boats (which are auto-calls) and unblock strong not nutted hands, such as medium flushes and straights, which are the ones we're trying to fold out sometimes when we bluff-raise.

This is a spot where it's very easy to overbluff if you start to use all the AcTx, AcAx combos, and where players are going to overcall against people they perceive as "bluff-happy". In general population at midstakes (500-1KNL) doesn't like to bluff against a very strong perceived range (here we represent straights, flushes and boats, very rarely trips), especially when the range is uncapped and they represent few value hands. I expect this spot to be extremely underbluffed, because people don't like to consider folding flushes in this SPR and will consider herocalling also combos of 98 with a club and 7x, hence a way strong players can capitalize some EV is to include thin shoves with nut flushes.

Hand 2

The card removal is already included in Pio's calculation (it's reflected in the EV of our hand). In general it's a spot where the size chosen by my opponent is not very coherent with the ranges in play (we're not capped, we have sets, straights and flushes) and it looks like a size that wants to maximize against a strong part of our range (2p+), plus it's really hard to find natural bluffs for that size -the only combo is A9 off with a diamond, which is a super low frequency raise on the flop against my sizing. Jamming some Q9s with bdoor on the flop, without blocking the flush, is a suicidal play, as it heavily blocks our folding range and unblocks our nutted region.

To answer your question: I'm always playing against my opponent's range, but sometimes I would fold all bluffcatchers, even the premium ones, if I expect them to be significantly underbluffing. I wouldn't fold value-catcher hands like a straight, as I expect villain's shoving range to be straight +, targeting sets, two pairs and pairs with a flush blocker. This hand is very useful to show how theoretical EVs don't matter much in spot where some actions are very rare in equilibrium (the preferred size is the small bet), as EVs can shift massively by just nodelocking tiny fractions of few combos. It's also a spot where some opponents might decide to turn JdTx into a bluff and calling 9d9x would be a disaster against that.