What puzzles me in your statement, and I'm reading between the lines here, is that you seem to imply that any tendency to twist while increasing spoke tension would automatically result in a yielded spoke.
Is this the case or am I reading too much into it?
The maximum stress in the bladed portion of the spoke will be the same whether you use a tool that holds the blade, or you just let the spoke twist. The torque applied is governed by the resistance of the threads, and if this torque is resisted by any part of the (weak) bladed part of the spoke the same stress will occur. The only difference will be in the angle of the twist... and I guess the location, since if you are holding the spoke, you know exactly where the highest stress occurs.
So... if the spoke is yielding without the tool, it will also yield with the tool, and if you exceed the yield stress, the maximum amount of plastic deformation will be the same. Maybe confining the yield to a shorter section of spoke is better than letting it occur anywhere in a large span... not sure about that. At least if you confine it to a small length the spoke will still appear to be straight rather than noticably twisted... but the magnitude of plastic deformation and weakening (if any) will be the same.
The only aero spokes I've used were Wheelsmith's, and they are thicker in the middle (1.3mm vs .9mm for CX-Rays). The spoke twist didn't seem excessive without holding them, but I don't know what happens to CX-Rays. I had assumed that normally the yield stress is not exceeded... but maybe I'm wrong... maybe these are taken beyond yield all the time, and the tool is necessary to make sure that the permanent twist is a short one rather than a long one. If this is true, but the spokes are not failing, then I guess it doesn't hurt anything.
Holding the spoke at the round part would eliminate the over-stress completely, though.