I agree the
implementation isn't sound, but there's nothing wrong with the
concept of a shaft passer.
As an example of why it is conceptually feasible, remember that you could computationally (or mechanically) determine when a shaft is closing in and open a gap in that particular support at the right time, maintaining ability to pass torque through all other supports.