This seems to me like it breaks down into two separate problems/questions: 1) how to provide sufficient feedback to the player when using these kinds of remote effects so that the puzzle feels fair, and 2) whether to design the game such that it’s possible to make it unwinnable. This second question isn’t specific to using remote triggers, of course, but it’s true that the design problem gets trickier when you start throwing in timers and the ability to affect things at a distance.
Flipping the order here, it to a certain extent depends on the audience you’re targeting but my sense is that the “mainstream IF” answer would be to prevent the game from being made unwinnable unless you have a really, really good reason for it – say, multiple playthroughs being part of the intended experience in a time travel-themed game, that has tools built in to make the boring repeated parts less annoying. I’d defer to others who are closer to the more old-school text-adventure side of things speak to whether that assumption would hold if that were the intended audience, but even still, I have to say it’s probably good design to prevent the game from being made unwinnable so long as there’s a logical reason to keep things on track (like, in your example, whoever designed the missile launch system was presumably smart enough to include a failsafe that aborts if the silo door is closed!) Regardless, if there are particular puzzles or setups that you’re having a hard time figuring out how to keep from locking the player out of victory, I’m sure folks here would be happy to help brainstorm solutions for those if you share the details!
As to the first question, I think this largely depends on genre and the complexity of the puzzles – like, if you have five different remote triggers affecting three different devices across a map with more than a dozen rooms, you’ll probably want to include quite a lot of feedback, whereas if you just have one switch that will cause a door three rooms down to visibly open, you probably don’t need to do much more than indicate that the switch seems to have done something.
I will say that as a player, having to hoof it around a map just to see what a switch or button did is often irritating, all the more so if there’s no indication why the system was built that way without a monitor or sign explaining things, or a voice-over announcer initiating a countdown or something like that – it can feel artificial. Obviously games like Myst and Riven included a lot of puzzles like this, so there’s an audience for them, but personally I never really got on with them, so you might have better luck getting thoughts from folks who do like these kinds of designs.
One final thought that maybe could help ground some of the complexities here is that it might be helpful to think through why you want to separate things out such that remote action is needed. If it’s just to obfuscate cause and effect, that might not be a good reason, but it might be needed to set up a timing puzzle or a navigation puzzle or something like that (say, you need to get from one side of the missile base to the other to launch the rocket after you opened the door, but there’s an enemy who’ll come out of hiding to close it if you haven’t yet dealt with them). In those kinds of situations, the added complexity feels worth it, but it also makes the “true” puzzle a little more concrete and easier both to design, and for the player to understand.
Hope this noodling is somewhat helpful!