Ah you have been handed the poison chalice of 'just measuring'.

First, if you are looking to see if you need to be adding filtering, it will not be just mains frequencies you are interested in, so the normal meter on the AC setting will probably not do - a scope or even a measuring receiver or a spectrum analyser may be more use as the switching spikes will extend to some MHz spectrally speaking (due to microsecond risetimes.)
Measuring common mode current for EMC purposes is commonly done with a clamp meter with a ferrite core and a dummy load and the instruments on the secondary. Measuring common mode voltages does require some thought as to 'relative to what exactly '  and in a professionally setting normally involves a LISN. (more info)

The limits you will be looking at will not be expressed in volts and milliamps, but as power in  a given spectral bandwidth and this can be hard to deduce

This sort of thing ( many thanks to Tim for putting his course notes on-line)

Mike

Good answer Mike! I however have a question, "why do you want to do these measurements?". As Mike says they are difficult and will probably not give you any useful information. If you are hoping to get a rating for C1,2,3, or a current in N2, the answer is more easily calculated than measured. The EMC result from proper measurement with a measuring receiver is dependent on many things, particularly the type and length of the motor cable, its enclosure, perhaps conduit if long, and the effectiveness of the Earth path at higher frequencies than the switching frequency ( a few kHz). The 50Hz leakage will depend on the mains filter and general capacitance to Earth of everything. I would not expect to meet 3.5mA with a VFD of any significant size, and there is zero reason to worry about it as connection via an RCD would not be required, or if the system is TT the value could be set much more than 30ma. So why?