Still in the last stages of sick. Spent another day playing Mass Effect 2 and it still holds up as magnificent. Don't want to drill into it yet, as it's new enough to be very spoiler sensitive, so instead i want to talk a little about skill challenges and color.
If you're unfamiliar with the term, color is how things are described in a non mechanical sense. it is separate from mechanical effect but, ideally, closely related. When I make an attack, the mechanical description is "I roll to hit, if successful, I roll damage and apply the number rolled with appropriate modifiers". The color may be "I make a heavy overhand swing, smashing through his defenses" or "I make a number of feints to the left, then strike a blow the to the opening in his guard on the right."
The difference between those two elements of color is quite broad, and that points to the heart of the issue. Depending upon the game you're playing, the mechanical description may have a similar level of variance - for example, if the game has rules for "all out" attacks or feints, then there will also be some mechanical differentiation. There are also implicit mechanical differentiations from things like the weapon used, stats, skill and such.
Ideally the mechanical and color descriptions fit together like hand in glove, perfectly matched for one another. In reality, that is very hard to accomplish, and it tends to depend on one or the other being primary, and then expecting the other to match. To illustrate, consider that all out overhand smash.
If I am playing a game which is "color first", I describe what I'm doing and the GM interprets it to express it mechanically. In this case, for example, the GM might decide to increase the opponents defense slightly (representing the inaccuracy of my attacks) but increase my damage if I succeed. Or he might consider how the other guy fights and, based on whether this is a good or bad tactic against him, grant me some bonus or penalty. This has a lot of benefits for player seamlessness of experience, but it also puts a lot of onus on the GM, and is only as fair as the GM herself. Also, this is hard to do well: the GM needs to have mastered the system and be able to think quickly. Some systems are better suited to this than others.
If I'm playing "mechanics first" then I make the decision of which mechanic I intend to use, such as the Power Attack feat, and I then describe my action in a fashion defined by which mechanics I'm using. If I'm not engaging a mechanic then I might embellish the description with "pure color", but that color still needs to fit within the general framework of action. The exact boundaries of that framework tend to be social ones, laid down by the genre sense and general sensibilities of the table.
Most games waver back and forth along this line in play, but the most common default is "mechanics first, if there's no appropaite mechanic, try color first." Mechanics really drive action, often in ways we're not even consciously aware of. And this is where we come back to 4e.
4e D&D is very much mechanics first. Admirably, it supports color first play with page 42, but that's very far from its focus - most of the time you are applying a specific mechanic, usually in the form of an action or power. This works very well for it, and it is supported well on the small scale (in combat) as well as on the larger scale (rituals and such). But not for skill challenges.
The mechanics of a skill challenge are very flat. Roll, add, accrue a success or failure. Nothing distinguishes between a Survival roll and an Arcana roll excepting the GMs say-so. It is because of this that it is absolutely essential that the color of a skill challenge be vibrant enough to make these rolls make sense - the players should be picking the skills they'll use because they're obvious, not because the GM has to explicitly ask for them.
But that's a pretty serious gearshift from the standard mode of play for 4e, and not every GM will have an easy time shifting gears like that, but thankfully 4e provides some easy examples for how to address this.
You need to treat every action in a skill challenge as a power use.
That may sound weird on the face of it, but the logic is simple. Rolling the dice for an action should be as interesting and have as much impact as it would in a fight, and the reason actions in a fight work so well is because something happens. If you call for a player to make a skill roll where nothing is going to happen (except perhaps an offscreen checkmark) then that calls into question why they should roll in the first place. The challenge to the GM who wants to create an engaging skill challenge is to offer lots of explicit actions to be taken, ideally with specific mechanical and color impacts, as if each action was a well designed power.
The prospect of coming up with that many "powers" for a skill challenge may seem daunting at first. It can be overwhelming unless you take it as an opportunity to think about what makes this skill challenge awesome and fun (rather than just a placeholder). If you can't think of ways this skill challenge can be exciting, then maybe you should consider skipping it.
1 - A good example of this is jumping 30 feet in the air and then smashing down on your foe with a mighty blow, like something out of anime. In some games that's entirely appropriate (such as Exalted, where you might even get rewarded for that) but not in others. Especially not in games where you later come across a 20 foot high wall and wonder why you can't just jump over it.
2 - Stunt systems, like Exalted or 4E's page 42 are an interesting way to blur the line, though in different ways. Exalted's stunt dice are rewarded for providing better color to your mechanical decisions, whereas page 42 is basically "the rules" for color-first 4E play. Needless to say, I love this stuff.
3 - 4e Powers actually make the application of color much harder. Their own baked in color (the descriptive text) is often only tangentially related to the mechanical effect. This is intentional, and it is a double edged sword: on one hand it has made the actual play experience much smoother, but it has explicitly also removed much of the "logic" from the setting. Fireball doesn't light things on fire, Lightning can't be grounded etc. The GM can set up exceptions, of course, but if they work all the time, it's a change to mechanics that can throw things for a loop. The GM might be able throw around a few +2 bonuses for doing something clever, but that's ultimately bland
4 - Skill challenges work great as "color first" activities, especially if the GM never mentions that there's a skill challenge going on, rather just tracks players actions and successes towards an end.
5 - It's far from perfect, but I'll point to The Siege of Fallcrest has some good example of making skill use explicit and impactful.
6 - This is a problem for 4e, but it's even more of a problem for more freeform games like SOTC. If there is only one currency for dealing with all responses to a problem then there is no difference between having exactly the right tool for the job and just being kind of prepared. On one hand, this rewards players without them having them obsess over inventory management and tracking minutiae, but on the other hand it can feel disappointing to players who really have lined things up.