With that said, you could definitely design a game in such a way that mandatory damage-taking doesn't break the established rules, but it would have to be very different to the typical action game.
Aevin's example is also a good example where it's okay, like the toxic floor thing, but I feel it's fair, and only fair if there is a buffer in place, such as the healing before you enter the room.
-----
But ya, my takeaway from all this is that I was right about the "no dmg idea" being the golden standard in almost ever possible way, with the rare exceptions being RPGs and some very rare examples provided in this topic. But I also believe that many people are right by saying "It's only a good start" as relying on it as the only standard to fair design is clearly nonsense. You will almost certainly create terrible unfair challenges that only you think is fair.
For example, you can no-damage a boss that doesn't telegraph anything and you're forced to memorize the entire thing. Does this make the boss fair because you can eventually figure out its pattern and take no damage? No, absolutely not. The boss is still bullshit because it's designed for you to beat it based entirely off memory as your only strategy. Just because the quest developer can "no dmg" it doesn't mean the quest designer can use this as his excuse to greenlight the room, boss, or challenge, it could still be a bullshit room.
My experience with Mario Maker and discussions within this topic has made me realize there is another gold standard, you could say it's one step up from the "no dmg" is essential to fairness rule, and that is very good solid communication.
That's the bottom line. If the developer cannot communicate to the player what the developer wants the player to do to face that challenge, the developer has failed in every way in making a fair challenge.
And what's unfortunate is that communication is not often a strong suit of many game developers because we're all kinda on the autistic spectrum in one way or another. lol (That's a joke, don't take me too seriously).
One thing that fails me personally, and I'm starting to realize it (and beta testers really help me with this area) is that I get into the same pattern of playing my quest over and over again. I do it my specific way, specific route, with the exact same strategies, and find myself completely blind to the overall design that other people will experience. I will not see how this corner is put in a bullshit spot because I personally never go near that corner. I will never see how a statue is put in a bullshit spot because I never go near that statue. But people playing it blind adds a lot more chaos to the "what to expect" realm of design, and holy shit, there is an infinite potential for a well designed dungeon to go horribly wrong if not every single one of these precautions are put into place.
Which makes me realize, to make things fair, you need a higher vision. A higher way of looking at your rooms, you actually have to stare at this room and analyze every object and ask yourself "what does this object communicate to the random player", and this is my new discovery on why game design is so insanely hard. Because I have never thought for the life of me that I'd need to pay attention to every single object and think how every single person could responed to that object.
At this rate, for me to get my quest right, and I mean right to the quality that I'd hope to achieve, it would take me at least 50 years to finish my quest. Of course, I'm going to have to settle to just let some bullshit slide, but hope I can just make it all work in the end.