If you strip them of their narrative context, skill challenges are a terrible game; select your highest skill and try to roll well.
If you strip the roll to find a secret door of it's narrative context, it's a terrible game: try and roll a 1 on a 1d6.
But skill challenges have no meaning outside of their narrative context. The 4e DMG even says as much (p 72):
More so than perhaps any other kind of encounter, a skill challenge is defined by its context in an adventure.
It goes on:
Define the goal of the challenge and what obstacles the characters face to accomplish that goal. The goal has everything to do with the overall story of the adventure. . . .
Give as much attention to the setting of the skill challenge as you do to the setting of the rest of the adventure. You might not need a detailed map full of interesting terrain for a skill challenge, but an interesting setting helps set the tone for the encounter.
If the challenge involves any kind of interaction with nonplayer characters or monsters, detail those characters as well. In a complex social encounter, have a clear picture of the motivations, goals, and interests of the NPCs involved so you can tie them to character skill checks.
I have never seen a compelling argument for what benefit this serves over a specified action system
Here's an account of the
difference between HeroWars (an early version of a closed scene resolution framework) and RuneQuest.
And
here is an argument
in favour of conflict over task resolution:
In task resolution, what's at stake is the task itself. "I crack the safe!" "Why?" "Hopefully to get the dirt on the supervillain!" What's at stake is: do you crack the safe?
In conflict resolution, what's at stake is why you're doing the task. "I crack the safe!" "Why?" "Hopefully to get the dirt on the supervillain!" What's at stake is: do you get the dirt on the supervillain?
Which is important to the resolution rules: opening the safe, or getting the dirt? That's how you tell whether it's task resolution or conflict resolution.
Task resolution is succeed/fail. Conflict resolution is win/lose. You can succeed but lose, fail but win.
In conventional rpgs, success=winning and failure=losing only provided the GM constantly maintains that relationship - by (eg) making the safe contain the relevant piece of information after you've cracked it. It's possible and common for a GM to break the relationship instead, turning a string of successes into a loss, or a failure at a key moment into a win anyway.
Let's assume that we haven't yet established what's in the safe.
"I crack the safe!" "Why?" "Hopefully to get the dirt on the supervillain!"
It's task resolution. Roll: Success!
"You crack the safe, but there's no dirt in there, just a bunch of in-order papers."
"I crack the safe!" "Why?" "Hopefully to get the dirt on the supervillain!"
It's task resolution. Roll: Failure!
"The safe's too tough, but as you're turning away from it, you see a piece of paper in the wastebasket..."
(Those examples show how, using task resolution, the GM can break success=winning, failure=losing.)
That's, if you ask me, the big problem with task resolution: whether you succeed or fail, the GM's the one who actually resolves the conflict. The dice don't, the rules don't; you're depending on the GM's mood and your relationship and all those unreliable social things the rules are supposed to even out.
Task resolution, in short, puts the GM in a position of priviledged authorship. Task resolution will undermine your collaboration.
Whether you roll for each flash of the blade or only for the whole fight is a whole nother issue: scale, not task vs. conflict. This is sometimes confusing for people; you say "conflict resolution" and they think you mean "resolve the whole scene with one roll." No, actually you can conflict-resolve a single blow, or task-resolve the whole fight in one roll:
"I slash at his face, like ha!" "Why?" "To force him off-balance!"
Conflict Resolution: do you force him off-balance?
Roll: Loss!
"He ducks side to side, like fwip fwip! He keeps his feet and grins."
"I fight him!" "Why?" "To get past him to the ship before it sails!"
Task Resolution: do you win the fight (that is, do you fight him successfully)?
Roll: Success!
"You beat him! You disarm him and kick his butt!"
(Unresolved, left up to the GM: do you get to the ship before it sails?)
(Those examples show small-scale conflict resolution vs. large-scale task resolution.)
Something I haven't examined: in a conventional rpg, does task resolution + consequence mechanics = conflict resolution? "Roll to hit" is task resolution, but is "Roll to hit, roll damage" conflict resolution?
Whether or not you find that argument
compelling is probably a matter of your taste.