As far as "proper" goes, this is one place where there's pluralism. There are many good ways to do it, and many bad ways to do it, but no one perfect way applicable to all games and designs (or design sensibilities). Whichever way you choose to implement this depends on your (good) judgement as the game designer and programmer.
I personally don't like game-logical movements to be determined by the animation data, and prefer discrete states to my main character.
So I would not move my character upward in Spine, I would just apply a velocity or force to the RigidBody in Unity, and just animate the tucking in Spine. Then, I would just switch between two collider shapes: the full height and the tucked shape.
But letting your animation determine logic is something some designers definitely do.
You should also try checking the Unity tutorials. I recall there being a topic there touching on a jump animation causing the character controller collider to shrink. You could probably pattern your implementation after that if you think it makes sense for your setup.