OK, so your earlier posts about skill challenges being bad RPGing with low agency turns into I prefer GURPs and D&D 3E to anything else.
That is the least charitable reading of my position possible, and unfair. My criticisms of skill challenges are very specific, and I am happy to concede that there exist entirely different drives/desires and things to extract enjoyment from tabletop RPGs than mine, and that skill challenges are serving a specific purpose, particularly in pacing and structure of non-combat challenges.
I am pointing to a specific flaw, to which a reasonable response is "expression of mechanical agency is not particularly important in my non-combat RPG scenes." If you don't think it's important that the mechanics of non-combat resolution require what I'm calling "gameplay" then my criticism is not relevant to you. I want to be able to make good decisions when faced with a challenge, to feel clever about having found a better way to resolve it with less risk, less expenditure, or, alternately, to fail to do so and figure out what I could have done better when I review the situation later.
After I play something like a game of Netrunner, there's often a conversation wherein we discuss the lines of play we tried, analyze whether some risks were worth it, and talk over any lines of play we missed that might have changed the outcome. I'm suggesting that a game of D&D (say a reasonable sub goal, like breaking into a castle vault) that uses a skill challenge model of action resolution would not be amenable to such conversations after the fact, because there would be little difference mechanically between various lines of play.
The second is not true for much RPGing. When I asked my friends earlier this year to play a session of White Plume Mountain - which they did - the objectives were not player determined. And I'm not the first GM in the history of AD&D play (or other versions of D&D) to have suggested a module in this way! Gygax gives an example of play in his DMG in which the GM, not the players, sets the objective of play.
I'm not sure that's meaningfully not a player set goal. I am admittedly not familiar with the module, but I think it quite reasonable to assume the characters played in that scenario will adopt the modules goals as their own. What I'm saying is that unlike a board game, the players can pick goals other than "accumulate the most victory points" or perhaps "survive for 6 rounds" from a cooperative game.
I do want to point out that I'm talking significantly more granularly than a module as a goal/victory condition. A goal might be as grand as "dethrone the illegitimate monarch" or as fine as "cross this river," both of which are things a player might want to do, and both of which I would like to be able to take optimal actions toward achieving.
4e DMG p 72:
A skill challenge can serve as an encounter in and of itself, or it can be combined with monsters as part of a combat encounter.
I don't believe this is relevant to the point at hand. Unless I'm misunderstanding the example, the success state for the skill challenge was a diplomatic solution and temporary truce to deal with a devil threat and the failure state was continued conflict between the gnolls?
The PC's action in this case changed the failure state of the skill challenge, something outside the listed rules, but again, I think the correct choice and a reasonable extrapolation of the fictional state. The quote you're pointing to seems to be suggesting something more like say, attacking a necromancer's minions while disrupting his magic circle, which I don't think is a contentious example of a skill challenge in a combat scenario.
4e DMG2 p 80:
A good skill challenge also allows for other actions outside the framework of the skill system to contribute to the party's success. Spending money, using powers or action points, combat encounters, rituals, and the simple passage of time can contribute successes or do any of the things that secondary skill checks can accomplish.
So it seems to me that what
@Manbearcat is doing is exactly what the rulebooks tell a 4e GM to do.
I think what you're suggesting here, to interpret the situation back into the skill challenge framework, is that the combat encounter created a success in the skill challenge outside of the usual skill checks? I admittedly was done with 4e before grabbing a copy of the DMG2, but I think the point you're making here is that you could retain a skill challenge framework, interpreting actions/events that are resolved outside of the skill system as successes (and implied, but not stated, also failures)?
If I've got that right, I do agree that potentially adds more agency to the framework, because you have more choices than striving to pick your best skill to roll. I do worry that you're kind of leaving it wide open to how effective any course of action is, which may make it impossible for a PC to make an effective choice, but definitely it's an improvement.
I am less persuaded, however, that this is preferable to not using a skill challenge framework. Instead you could adjudicate each action individually.
What differentiates a RPG from a boardgame is primarily that fiction matters to resolution, and that (subject to some fairly avant-garde exceptions) most of the participants engage the game via controlling an imagined person within the fiction, and declaring actions for that person.
We can agree to disagree on what makes TTRPGs good and still like them.
This claim is highly contentious.
If the GM is allowed to introduce elements into the fiction more-or-less at will, and is entitled to extrapolate consequences from the fiction as they think makes sense (as per your discussion of
@Manbearcat's gnolls), then it is the GM, not the players, who decide when any scene or conflict is resolved. Because the players can decide to make particular skill checks with discrete outcomes as much as they like, but it is the GM who decides what follows from that.
I don't generally think that's true for most skill checks. Obviously, the GM creates the world at large, but if you specify action resolution, you're not doing anything fundamental to the fiction you already described when you complete any given action.
Once you've described a wall as existing, the fiction doesn't change once a player decides to climb it, or to walk around it, or to smash it. Perhaps you have to elucidate some part of it you hadn't specified ahead of time, say, if they sneak into a keep and you suddenly need to figure out where the guards are located, but that's just providing further clarity to a fictional state that is presumed to already exist.
More to the point, what I'm saying about "picking victory conditions" is precisely the means by which players decide when a conflict is resolved. They know what their goals are, when they're achieved or can no longer be achieved after we check back in on the fictional situation after resolving each action, they can set new ones.