The other half of what useBlobShadow starts: not a smudge under a thing, but the shape of the thing thrown onto what is behind it. One light in a scene opts in with castShadow, and from then on the engine draws every caster into a depth picture taken from where that light stands.
The fox casts the pose it is in, because the bones are applied again while the map is drawn. The floor opts out of casting and still receives, since being in the map is about casting and never about receiving. And only that one light's share is taken away, which is what keeps a shadow from reading as a black smudge.