[confidential transactions] would require a hard fork. The output value would need to be replaced with a EC points. You also need range proof support.
Actually you can soft-fork CT. https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2016-January/012194.html
The main challenge is size. I believe I have a way to reduce the range proof from 2.5kB per proof to 2kB per proof but it is still large.
Mimble Wimble is interesting also in making an aggregatable CT which allows the bloat to be more than reclaimed at least as far as catchup goes. (More total bandwidth used, but less bandwidth to catchup as catchup becomes proportional to UTXO size + a smaller overhead per historic transaction). Maybe the historic transaction overhead can be removed.