what you want is a three phase power analyzer.. the ones with 3 voltage probes, neutral, and four clamp on amp meters, one for each phase and for neutral. some use rogowski coils instead of current transformers.
in anycase they are expensive and i'm supprised someone hasn't programmed their own yet.
basically an oscope isn't a good choice for this task, the reason being that the 8 bit adc front end isn't really enough, a 16 bit sound card would get you more information (0-400hz is pretty slow, oscopes are designed for 0-400Mhz)
if you want to cheap out buy a 300$ oscope that is known for a simple, easy to use, and bug free FFT math function.. but that won't tell you much.. it will tell you if one coil isn't the same as the other, but that's about it.
also most cheap oscopes can't give you information on the math output which is unfortunate.
so you won't be able to use an oscope as a watt meter.--i think most scopes over 800$ will do this.. i'm sure others will point out that rigol's latest 400$ scope can be hacked.. and i think it can be used as a watt meter.
As to making an amp meter, you can use any cheap clamp on amp meter as a current probe for motor rewind work..
You can literally buy a 8$ ac amp meter on ebay, open the case up and solder two wires from the sense coil into a bnc cable.. make sure you have a load resistor.
hopefully its something like an even 1000 turns or 2000 turns, not all of them are.
anyhow if you can get voltage in one channel and current into the other you can use the math function to multiply the two together, but again most cheap oscopes won't be able to tell you the average, rms, of that resulting math waveform. it would also be nice to be able to perform an FFT of that resulting math waveform.
so basically, those dedicated 3 phase power meters are designed to do all that and tell you what you want to know.. which is power on all three phases, amps, current and their harmonics.. and you want to know the backwards rotating harmonics, which have to be calculated not measured afaik..
you could perform these measurements with an oscope but you will literally be taking screen shots and counting squares under the math waveform.
also the 8 bit adc on most oscopes will quickly grind your gears because the best you'll get is 1 part in 100 for any rms voltage measurement.