You have a very direct approach that I'd like to use as an example. I want to start with: I agree with you. And then continue with afterwards with: How do we improve on the mechanics of a skill challenge to make it more in-line with our other sub-system that we do allow to decide life and death.
---
Combat is going away from the whole RP / skills approach to a mechanically heavy minigame where you roll dice for life and death situations. In other words, mechanically similar to a skill challenge in some ways.
But when you replace "skill challenge" with "combat", you get a description that doesn't fit most tables:
So why is it that one set of constrained mechanical effects are used extensively for life and death, and others is an "Absolutely NO"? I think it has to do with the options that a player has in combat, vs. the very limited scope in the mechanics of skill challenge the OP posted. But that's just on viewpoint.
I'd be interested in your (and everyone's) viewpoint on how to mechanically handle out-of-combat challenges to a degree where we are comfortable with the same range of outcomes we would get from combat.
In my ideal world, skill challenges would have more than two dimensions and several defined potential end states. The players would have a set number of actions they can take to try to move from their initial position to as close to their desired outcome as possible. Each successful action moves their state towards their goal; each unsuccessful action moves their state towards a different potential outcome, as appropriate to the situation. An action can be used to attempt to repair a previous failure.
Based upon the amount of planning and preparation undertaken, the DM assigns a number of steps the players may take before the situation is considered resolved and based on the initial situation the number of steps required in any one direction for complete success.
For example, the PCs are planning a major heist. Their ideal outcome is they get all of the treasure, the owners are unaware of the theft, the PCs leave no clues behind to identify the perpetrators, the authorities are unaware of the theft.
Based upon the starting circumstance, the DM works out how many steps will be required to navigate the challenge. The owners will detect the players unless at least 6 steps to avoid their attention, the authorities will become aware something is going if less than 3 steps are made towards them, and 4 steps are necessary to avoid leaving enough evidence to finger the them afterwards. Getting the whole haul will take 5 steps.
The DM determines the players have prepared for 20 steps. The PCs don't have a lot of room to repair failures and still achieve complete success!
The nice thing about this challenge format is it supports mixed success results. The PCs get the treasure, but leave some evidence behind and the authorities are partially alerted, for example.