I've thought of it too. I think one possible way is to do it with three people. Of course this involves some commitment from a person that's not technically playing but maybe they could learn something from it.
Anyway you would have two people face each other in real space. One person is actually playing and the other person is just a puppet. The puppet is being controlled by the other person skyping. The one skyping in just makes the call as the puppet follows his orders.
Otherwise, yeah, just do it like you just said.