So stepping outside the box is just too much for developers?
No, it's just that Kinect wouldn't quite work with it. It is good but not good enough to be utilized in more complex games. It would require a lot of improvements and it would be a pain in the ass to make all features of the game work with it.
Take an FPS game for example. How would that work? Would you go around running on your living room to move your character? Would it be easy and would it recognize well motions like pulling the trigger or reloading the gun? The kinect camera has really low resolution and it can't recognize the smallest movements like your fingers or the quick movement realized by your hand when "reloading" the gun.
Or an RPG like TES. Imagine how would you for instance open the inventory? Or walk? Or grab objects, or cast a spell, talk to a character, activate something, and a bunch of other things. It's too much complexity and data for the Kinect to process.