RAX and RAX-2 and possibly other cubesats launched by the University of Michigan (I can't remember at the moment) used fixed neodymium based magnets on their Z+ axis. Hysteresis was added to dump any residual momentum after P-Pod ejection. On the RAX missions, the magnets then were used to orient the satellite vertically over the poles where the science missions were conducted.
I haven't kept close tabs on the smallsat industry lately, but I would estimate that it is totally within the realm of doable, but with its risks. Cubesats are often built out of aluminum. While there may be other ferromagnetic materials used, and current loops in the electronics, I think a well chosen servomotor could over come most issues. The biggest risk in my mind would be dealing with vibration on launch. Moving parts tend to have higher risk.
Yes, there are rules about magnets (and just about everything else) on secondary payloads. The primary payload usually has many times more risk invested so they control almost everything about secondary payloads. The launch provider will have a say as well. It's almost always a case by case basis. The largest risk takers are probably any other secondary payloads on board. For example, M-Cubed (another Michigan launched sat) was believed to have stuck to another cubesatstuck to another cubesat using passive magnetic control, rendering both missions a loss.
Disclaimer: I worked on the missions mentioned here, as well as a few others.