Manbearcat
Legend
That is a problem solved for a long time, which is just using mechanics to describe the problems at hand, its what statblocks have been used for, and are still used for. Using mechanics to detail problems seperate from the actions/resolution of it, leads to having a toolbox to manage problems themselves, letting solutions be open ended. If you design around the idea that the players must do X to do something, or have to do X things, you will always either end up doing way too much, or railing them into a handful of solutions, this is not good.
Also i know thats not the OP's intent, but when your playing around with mechanics that are designed with this play in mind, its always the end result you get, my point is, skill challenges are simply not good. Its why whenever I see most people using them, It always works better in theory then in practice, and requires a lot of on boarding, because it is something designed with a specific engagement of the players in mind.
"That strength is that skill challenges centre the fiction in the process of action declaration and resolution."
If you're designing your challenges around what your players might do, you're just setting yourself up for wasted prep, or railroads.
This is such awful design, its always easier to design your challenges around describing the environment and let your players figure out how to solve it, there is no need to force them into doing stuff a X amount of times, or designing around the player's actions themselves, they can solve the issue in 100 ways, just use the mechanics to describe the problems.
Given that I’ve run 1000s of hours of play that employs the sort of game tech that are Skill Challenges, I can say for that your contention above “that they’re awful design” and “they don’t work” is just fundamentally not true.
I think the first issue is when you say “if you’re designing your challenges.”
This sort of game tech is there in large part to resolve noncombat challenges without prep (and/or to insure play against GM’s with preconceptions of play or “designs upon play” that they attempt to impose during play)! You’re not designing anything! You’re just running the game and rolling with whatever direction it goes!
The other reasons for things like Skill Challenges or Dogs in the Vineyard/Cortex+/Blades in the Dark Clicks etc Closed Scene Resolution (and dozens of other games) is for exactly what I’ve written above:
1) To ensure the opposite of railroading! When a GM has an encoded, table-facing scene budget with defined Win Con/Loss Con parameters and inviolate principles and procedures that govern resolution…the GM doesn’t get to just decide “this scene is over/still online” by fiat. System will answer that question by binding the GM’s play and overtly signaling that to the table.
2) To create dynamic situations that evolve over time to create premise-coherent fiction and interesting and consequential decision-points are und the gamestate (this last point feeds back into point (1) above about denying arbitrary GM fiat over “scene online/scene over” because you always know for certain what the status of the gamestate is).