Confusing (wrong?) band names for subset with snappy

It seems you have worked it out, almost.
There is an issue with using the SLSTR data from the command line.
I’ve noted it here:
https://senbox.atlassian.net/browse/SIIITBX-238

You should explicitly define which reader you want, as there three available. Ashish mentioned this already.

That Rad2Refl is complaining about the input product, might be caused by a SLSTR product in 1KM resolution.