There is no way to do what is asked without coding a system from scratch.
Yes you have the pose snapshot, that gives you the bone positions.
Yes you have a skeletal mesh with placed vertex.
Yes you could use a shader to copy the vertex order from the skeletal mesh to the static mesh to pose it.
No. Its not fast.
More vertex = more processing.
Since you already have the skeletal mesh posed and calculated, doing anything else after that is a waste of computation that will just increase the cost of everything else in order to maybe free up some 60byte of data used to keep the skeletal mesh posed…
Its probably not even worth the time to look into how to copy the vertex buffer between meshes in terms of end result.
Yes, you could in theory replace anything and keep it on screen without a skeletal mesh at a much reduced cost (because skeletal meshes have a limit).
The computation power of doing it is however very likely to break the gameplay with stutter for most systems.
So… smoke and mirrors.
Use a pre-made end pose. At the end of which, delete the skeletal mesh, replace it with a static mesh @ the same transform.
Won’t work with ragdolling. Will work with anything else.