It's interesting how they make the current shunts for the cheap DMMs, just a thick copper wire that they appear to "calibrate" by pinching it with some wire cutters until the right voltage drop occurs with the current value.
A real current shunt would be best, but a few wires paralleled together probably would work on the cheap.
If the shunt was set up to give full scale of 200mV at 100A, it would draw 2W from the signal at full load; if that's a problem then maybe just amplify the signal to be able to use a smaller resistance for the shunt, so there's less insertion loss.