The problem here is that the compiler doesn't know that b := make(byte, n) implies that len(b) == cap(b) == n. We already have an issue for that somewhere. I investigated this a little bit, but the only solution would be to introduce a new SSA op OpSliceBuild (or something like that), because otherwise the make call is lowered far before prove is able to see it.
EDIT: this is the issue I was thinking of, with also Keith agreeing that a new op is the way to go (#24660 (comment)).
Actually, the "call" part of this issue was already solved by a CL last year that made each make call have OpSliceMake wrapper in the SSA form, and prove has already been modified to handle this (87e2b34, CL196784).
What's left here is the +1 part of the argument to make, that confuses prove.