Welp, I tried it...
I was hoping that after I got OpenNI working it might give me something more than what freenect gives, but nope. Oh well, I attempted it. The NiTE libraries would be super useful, but they've all but vanished and it looks like OpenNI's usefulness for skeleton tracking is almost 0 at the moment. There really doesn't seem to be a good incentive to choose it over say, OpenCV (especially since OpenCV has built in features for image processing, analysis, and vision algorithms).
I'm still glad I have it, as I now have more options available just in case I find it does something I don't know about right now. In the mean time I'm looking into [Baak et al. 2011] which was used by SkelTrack, a project which I'm wary of because it hasn't been updated in year. The easiest/best option right now would be to use SkelTrack but we shall see.