Are the players informed, in advance, about this "logical conclusion given setting assumptions"? Or are they expected to work it out for themselves?Let's say you try to bribe a guard. But this is an important post and the person the guard is employed by is paranoid enough that once a month they have a cleric come in and cast Zone of Truth which just has the guards walk through the zone and answer "Have you upheld your employment agreement". It's a simple yes or no question so the cleric could be done in a short period of time as they quickly shuffle through dozens of guards for one casting of the spell.
The guard may not be particularly loyal, but he knows if he takes the bribe he won't be able to answer the question in a way that preserves his employment and potentially his life. In a world where Zone of Truth is not a particularly high level spell I think this could be fairly common practice. So yes, bribing a guard may be off the table in many situations because it's a logical conclusion given setting assumptions that they couldn't get away with taking a bribe. It has nothing to do with limiting options, it's about considering how magic could alter the world.
If the GM's head, and notes, are full of this sort of thing, and the players aren't told about it, and so their action declarations keep bumping into them, where is the player freedom and control that is at the core of sandbox play?