You can’t use historical mean because during transformation Box-Cox your data are modified by mathematical transformation. During Box-Cox tranformation you need to transform your USL and LSL as well. I hope its helps..
I agree with Daves. There is no possibility to get Cpk =0,8 with 100% conforming and Cpk=2 with 100% nonconforming. These indicators are depends on specification limits, so situation describd above is impossible (in my opinion).