First, 2 Realizations:
1) SP is also used for currency, but to heck with it, I'll keep using it for the time being. This whole thing is going to need a big language cleanup by the time it's done.
2) Rather than 1d6/1d8/1d0 damage progression, it probably makes more sense to use the light, medium and heavy damage progressions from page 42. I'll still use the previous model for illustration because I can remember it, but at the table, I'd totally go with the other approach.
All right, so with the basic concept in place, let's start with how to actually build a situation. Assuming a default monster HP level of 24 + 8 per Level, it's pretty easy to do a quick XP/budget conversion. A single 7th level monster is equal to a situation with 80sp. That math is pretty easy. What exactly to do with those 80 points is a little more interesting. It's effectively a budget with which to create problems, or to reflect player action. For example, breaking into the enemy camp might be an 80 sp challenge, or you might break it down into 4 20sp guards who need to be overcome.
Now, one fun part of this is that, in theory, SP can be converted 1:1 with hit points. That is to say, if I'm running a fight with an elite opponent with 250 hit points, I could take 50 of those points and turn them into SP and use them to make the fight more interesting. In strictly literal practice, there are some problems with this: monster hit points are not actually consistent with this and the budgeting will not always be intuitive. Much the same way traps can sometimes be turned against enemies, there are going to be time it makes sense to keep hitting the situation rather than the monster.
And here, right here, is the critical decision point. Those are real roadblocks, and not every player would be comfortable going past them, and as such, it's entirely reasonable to say this model won't work for you. No harm no foul. But if you can tolerate the loosey-goosey, GM improvisation this demands then stick with me. You can do some cool stuff with this.
First, don't worry too literally about monster hit points. If you bring a monster in as its own encounter element then you know what the rules are for that. It has a certain number of hit points and costs a certain amount of XP. No problem. But when you throw in a monster as part of a situation, its hit points are just a part of the situation. To illustrate, consider the 20sp guard. We can "beat" him by doing 20 points of damage (progress) with stealth, but if it turns into a fight, he'll have 20hp (or more aptly, 20 hp minus any reduction in sp from earlier skill rolls).
This rolls into the second thing: If you have a monster that also has a challenge component, don't worry too strictly about distinguishing it's HP budget from it's SP budget because - and this is the kicker - really they're the same thing. Consider, for example, the group being attacked by a goblin horde (treated as one creature). The players can fight - just keep killing goblins until they stop - but maybe they want to try to scare them off (Intimidation). If a player wants to engage the "situation", then he might use intimidation to reduce SP, which comes out of the horde's HP budget. Something similar might happen when you try to reason with someone you're fighting, trying to convince him not to fight.
This interoperability with hit points would not work on its own, but there's another important element of challenges - the difficulty. How hard is it to do this stuff? To answer that, I ask the much more important question: What are you going to do?
There's an instinct to answer "Use this skill!" but that overlooks something important. The goal here is to open up _actions_, which skills represent and support, but do not define. That may not be immediately clear, but pull up to a higher level. When faced with a challenge, there are a few ways to approach it, but in the abstract, they break down into four different approaches:
First, you can try to circumvent it. You can elude it, go around or otherwise avoid engaging it on its terms.
Second, you can try to manipulate, control or wrestle with it. You meet it head on and try to bring it to heel.
Third, you can try to understand it. Study it, and use that knowledge against it.
Fourth, you can hit it. Hard. Possibly repeatedly.
Fifth, you can suck it up.
Sixth, you can run away.
Now, I want to set aside #5 and #6 right away. Both of these might call for skill rolls, but neither actually helps get past a challenge, so they're outside of the scope. #6 might be a _different_ challenge, but that's it own thing. #5 is the desperate hope that the other guy will get tired punching you. There are situations where it's appropriate, such as when the challenge is on a time limit or otherwise constrained, but in that case, it's under the auspices of what bad things the challenge is doing, and that's another topic. So, in short, they're off the table here.
#4 is easy. D&D already handles that really well.
#1-3 is the trick. You will generally use a skill to do one of these things, but there's no 100% correlation between which skill is used what way. There are some logical limitations, but they're situational. When you need to travel through enemy territory, Survival might be used to get around enemies (circumvent) or it might be used to try to find the best route (understand). Viewing actions through this lens of what they accomplish makes them more versatile and interesting, and definitely I smore satisfying than the "skill first" approach where a player says "I use Survival!" and upon being asked "Do do what?" has to scramble for an answer. 
Experienced 4e hackers probably have already seen where I'm going, but I'll lay it out here. The 4 approaches correspond to the four defenses used by 4e. Thus:
1 - Circumvent - Reflex
2 - Manipulate - Fortitude
3 - Understand - Will
4 - Smash - Armor
While this creates a little bit of a shift in how to handle some situations (Stealth being an attack vs. reflex rather than a roll vs. Perception) it streamlines things a lot and, more importantly, makes it VERY easy to extrapolate a challenge from a monster (or to fold a monster into a challenge).
So, for example, If the Warlord of the Orcs is hunting for you, the GM could create the hunt as a situation, drawing from his HP to budget it (or just creating it), and set all the "difficulties" of the hunt (which is to say, its defenses) using the Warlord's stats. In effect, by evading the hunt, they're fighting him by proxy.
Now obviously, there's more to cover. The big one is, of course, the bad things that challenges can do to you in return - without that they're pretty toothless. But that, I think, is something for next week.
1 - Yes, this effectively "weaponizes" skills (which is part of why I want to sync up with the page 42 damage expressions. A strict mind could view these as repeatable stunts). This is an idea I'm fond of from its success in SOTC fights as a way to help all players feel able to contribute at all times 9and ot encourage them to find interesting ways to apply their skills). This is maybe a little less relevant to 4e where there's less of an idea of a "non-combat" character, but it can still help.
2 - This may immediately raise some confusion of how to handle fights where a lot of HP and SP damage has been done, but 4e actually solves that VERY elegantly for us. If you take someone down to zero HP (or SP), you decide what happens to them. Often this is just used to decide "Dead or Unconscious", but it's a MUCH more powerful tool than that. The guy who takes his last HP could decide his enemy listened to reason and surrendered. , just as the guy who takes the last SP could decide he distracted the guy enough for him to get stabbed (or that he had just gotten through, just as he got killed, for maximum angst). The limiter on this is not a function of rules, but of what makes sense to the table, as it should be.
3- Important note: In all three cases, the activity moves towards action. That is, just studying something isn't going to make progress against a challenge (unless studying IS the challenge) - the knowledge needs to be gained and applied. This need not be _strictly_ applied with each roll, but it must be part of the trend. That is, if you sneak past the guard and get him to down to zero, you should be taking him out (in whatever manner you see fit). If all you're doing is going past him and not actually impacting the situation, that's just a roll, not a challenge.