Reading over the skill challenge. I personally think its less the mechanics of the challenge itself and more the presentation of how its used (though the core math needs to be solid, one of 4e's biggest issues was that the initial system was straight up bad, players had huge failure rates with even standard challenges. That was eventually addressed but it left a bad taste in many mouths).
In the example, the DM is calling for specific checks, which to me is the right approach. 4e's system often allowed the players to "figure out" the check they wanted to use, which was slow, fiddly, narratively awkward, and just helped players game the system. In this one, you have specific checks (athletics and stealth) which is a cleaner execution.
The second problem with many 4e challenges was players could feel like they were a detriment rather than a benefit. The way the math worked out, it was often better to have fewer players (who had the better skills) do all the work while the rest of the team went off somewhere. many social and research challenges were especially problematic in this regard.
This model partly solves this with the notion that the enemy is rolling to, and we have to beat "them" rather than a static threshold. Further, I think challenges work best when they are thrust on the players (in this case you have to stealth or be discovered) rather than voluntary challenges the players can take (we do a challenge to go and research that ancient relic). Aka challenges work best under an immediate time pressure.
I also like the "reset" after each round. This helps make the math less swingy and creates a sense of continuous achievement (or failure), rather than rolling rolling rolling until finally the magic number is hit.
Overall, I like the skeleton here, its got some good bones.