i think that's a little simplistic.
going up-wind ... you're right, for a given propeller disc, there's a maximum vehicle speed. initially the blades would be set at an angle of incidence to the wind to maximise torque (like a wind turbine). as the blades start turning they create an in-plane velocity which combines with the wind to create an apparent wind vector, and the blades will now rotate to an incidence that maximises thrust (to get the vehicle moving). now the apparent wind is the vector sum of the vehicle's velocity, the blade's turning velocity, and the wing; and speed will be limited by the thrust produced by the propeller and the aerodynamic drag and othe losses. presumably the vehicle would start against brakes (the the wind would be pushing the vehicle backwards), and the vehicle will only move forward once the thrust produced exceeds the aerodynamic drag.
going down-wind ... the same procedure would happen ... initially the blades produce torque to turn, to rotate the apparent wind; and then produce thrust to move the vehicle. here though, the windage is helping the vehicle, so "drag" is initally +ve (a thrust driving the vehicle), then 0, and then -ve (retarding the vehicle like drag normally does).
thinking about the blades ... either the profile is dependent on the vehicle direction (up- or down-wind) which sounds silly (you'd have to change propeller depending on which way you were going) or the disc rotation would change (CW for up-wind, CCW for down-wind); i think this'd allow the same blade to maximise torque (initially) wih the same settings. mind you the propeller incidance control would be moving (in pitch) the blades in different directions but thats quite easily done. So the transmission would need to be able to take either rotation and produce fwd speed ... again QED.
Quando Omni Flunkus Moritati