Dave, I'm not sure how useful this idea might be.
You could use a small clamshell type Gurney flap on the trailing edge of each rotor blade. The "clamshell" could have a capacity for infinitely variable deployment from 0 to 90 degrees, and when fully extended, the 2 Gurney flaps (making up the "clamshell") would be deployed 90 degrees equally above and below the trailing edge.
This would have the effect of increasing or decreasing the drag of the rotor blade, with the variable deployment providing direct control over the amount of increase in drag. If the flaps (and mechanism) were small and light enough, they could even be cycled opened and closed during each rotorblade revolution to apply drag (and torque) where desired.
I can think of two possible means of inducing yaw with this idea. In one means, if we use the K-Max as an example aircraft, you could simply open the clamshells on both blades of a single rotorhead (without any rotational variance) to increase the drag on that rotorhead, to induce yaw. I think this would work well under power, but perhaps not too well during autorotation.
Another means would be to vary the opening of the clamshell so that only an advancing rotor blade on one rotorhead experienced increased drag. Say the clamshell started to open at 45 degrees after the advancing rotorblade passed over the rear of the fuselage, and closed again 45 degrees before passing over the nose of the fuselage. This would have the effect of increasing the drag on one side of the rotor disk, thus causing the aircraft to yaw towards the dragging side. I noticed on the K-Max that the advancing blade of each rotorhead passes over the fuselage, which would reduce the moment arm of an arrangment like this, but it might still work. I also think this arrangement might work better during autorotation.
Perhaps some cyclic control could also be built into this mechanism as well, since the K-Max already uses a flap type "tab" for cyclic control. This also might be a lot simpler than blade twisting.
I have no idea of the possible failure modes of an arrangement like this, but maybe the idea is useful.
PS - I also think this idea might work well with rigid rotorblades.
(edited to correct typo)