There is a common line of thought amongst DMs (of more than just D&D) to basically not bother to ask for skill checks when failing doesn't matter. Unless the room is filling with water, if you fail to pick the lock you can just pick the lock again. Why waste everyone's time? I agree with this line of thought, and it's interesting how some of what you've expressed falls into this.
On the other hand, something like a tripwire does have an immediate price of failure. Maybe Gale sneezed when your rogue was delicately trying to take the trap apart

I will note there is an ability for high level thieves to automatically roll 10s on skill checks they are proficient in if they choose, which is a way of smoothing over the "I'm far too experienced to make this mistake" scenario.
As for the dialog checks (like that book) I have actually found (to my disappointment) that a lot of these don't actually lead to more content at all.
For example, Underdark spoiler:
The friendly Mindflayer you meet in the Underdark makes you a potion to try to help you with the tadpole, and there are a series of checks you make against the tadpole under the effects of the potion. I had similarly terrible luck despite using lots of buffs to get through these 2 checks. Only to find out that battling through them seemed to make zero difference to the outcome.
It's hard not being sure you're going to miss something, but I think in general if you treat it like a D&D game where you have to role play and adapt to what happens you can get a lot of fun out of it. Sure, maybe on another play through you really want to go down a different path and you gotta try really hard for that other outcome, but so far I haven't seen many of those (decided by the dice). That being said, keep up your feedback, I think there is a tight balance between giving us choices and taking those choices away with a bad roll, so that's something Larian is going to have to iterate on.