That's an interesting comparison. In a way, there is a relationship between the Keyed Map dungeon crawl and the "skilled play" (aka sniff and listen/pixel audit) methods. Both involve a predefined setting or determination of what is there and it's up to the players to navigate that predefined location whether it's a dungeon or a trapped door or a searchable room. That's pretty much D&D and dungeon crawling at the beginning.
Yeah, lets see... In Dungeon World a trap would be a hard move, possibly the GM might reason "well, Discern Realities already indicated traps as a likely hazard. Now the party seems to be just traipsing down this hallway, handed to me on a silver platter, BAM!" At that point the player might describe their Defy Danger as "I attempt to carefully note any tripwires, pressure plates, or similar triggers." That would probably invoke a +WIS check, the Thief might well instead describe dodging the consequences +DEX. So, you certainly CAN get a narrative that reads "and the Thief spotted the trap before setting it off!" I might even let them use Trap Expert instead of Defy Danger even though they technically didn't explicitly trigger that move.
In the case of, say, TB2, well its actually pretty D&D-esque in general shape. The GM created a trap as one of the obstacles, how it will manifest is a bit dependent on the fiction to that point, but somewhat like DW, the character could be just moving along, and the GM might say something like "You feel a bit of pressure on your left foot as you start to pick it up." OK, given whatever character attributes this PC has they will have to form a dice pool to deal with that fiction consistently. Or the GM might simply say "you note the presence of a wire crossing the path."
The development of task rolls or tests to accomplish these things, whether searching for traps, looting rooms, or even navigating mazes or racing the enemy to the wizard's tower came later, in various stages, to streamline play or model the PCs rather than the players' skills or ability to negotiate with the DM or, in more recent games, provide an avenue for players to narrate the outcome rather than rely on a DM-accessed key.
Right, so the first problem was with traps and locks and such, where the GM is certainly not expert enough, nor unbiased enough, to judge the outcome of a run-in with a trap, or a lock picking attempt. Thus the thief was endowed with 'skills' that would abstract away the actual "did I set it off" etc. This is essentially the same role that attack rolls play in combat. It does, kind of, work, actually not too badly for something like a trap. The DMG even indicates a kind of 'level of success' determination heuristic (IE you may set it off, you may just fail to disarm it, or you may succeed in disarming it). At the time Greyhawk was published there were maybe 2 RPGs on the market, so its not surprising this model was used.
Modern games like PbtAs certainly have seriously changed things, but there was a kind of evolutionary sequence.
First was Greyhawk, pure success/fail
Second were games that allowed some resource to be spent to improve the odds (or succeed outright)
Third were ideas like 'fail forward' and 'retained results' (IE as long as you keep doing it, don't roll again)
Finally we get to intent and/or 'if you do it, you do it' kind of setups, usually with levels of success, etc.