Let me begin by making it clear this is not my invention, I encountered the method in City of Mist but I doubt it debuted there. But it is a nifty method.
The problem I encounter from time to time is that my players don't latch onto my descriptions of the scene, not using things in it to grant themselves advantages (bonuses, extra effect etc). Am I perfect? No. Could I do better? Yes.
Or I can take my fuzzy descriptions and make them mechanical by introducing them as Scene Tags. Market square during market day would get Crowded-Market-1 and during a festival Packed-Festival-Market-2 indicating that there is a lot of people there and also how much advantage one would get by incorporating it into ones action. Or disadvantage depending. Trying to pickpocket someone? Take a bonus. Following someone? They easily get lost in the crowd - penalty.
How dark is the night? Moonlight-Night-1 or Moon-Behind-Heavy-Clouds-2?
Traveling through a mountain pass and how deep is the snow? Ankle-Deep-1 or Up-To-The-Dwarf's-Beard-2? What about that Foul-Voice-In-The-Wind-4?
I play pretty much only online so tossing an index card onto the table with the Scene Tag on it is kinda tricky. Instead, depending on how much effort I've put into the VTT, I either write it in big bold letters on the scene image/map. Or I put down a virtual index card, essentially a small graphical element to bring attention to it (see post image).
One more thing, how much is a Tag-1 compared to a Tag-2 worth? This all depends on your system. City of Mist gives +1 for a Tag-1, +2 for Tag-2 etc. So for pretty much any other PbtA/2d6 systems the same works. For D&D (and other d20 systems) a scheme of +2, Advantage, Advantage and +2. I've lost much of my familiarity with d20 systems due not having ran something recently. So someone (everyone?) else probably have better ideas. In dicepool systems an extra dice for each tag level is appropriate.
That is the basics of it. But what if the players want to create a Panic!-At-The-Market-3? I'll write about that some other time.
I agree with you, somewhat, that PbtAs do away with situational modifiers. If we consider a clean or barebones (pure?) PbtA. But that isn't so true anymore considering the whole family. City of Mist (mentioned) above is a PbtA that is all about situational modifiers and how to exploit them. Lots of floating numbers from all those tags. Ironsworn have a bunch of add+N among its moves and even more among the assets.
So there are PbtA that do embrace situational modifiers and introducing Scene Tags to them wouldn't alter things that much.