The following is documented in https://github.com/openjdk/valhalla/pull/32 as a known issue that has been deliberately deferred to a later iteration:

"Need to verify that attributes from V are not carried over inadvertantly to V$ref.class"

There is already a defect report mentioning this behavior. See:

https://mail.openjdk.java.net/pipermail/amber-dev/2020-May/005999.html