A few KW tests in v1 folder marked "acceptable" violate spec for minimum plaintext length about wycheproof HOT 2 CLOSED

NIST does indeed not define KEY_WRAP for short inputs. However RFC 3394 does that in Section 2. The difference leads to ambiguous definitions. Some comparisons are here:
https://github.com/google/wycheproof/blob/master/doc/key_wrap.md

Also when a test vector has ambiguous definitions (i.e. is "acceptable") then there should be a flag, which is described in the header of the test vector file. Some libraries do in fact implement the behavior from RFC 3394. In these cases it is important to test the correctness of the wrapped key, since short keys are a rather strange exception.

from wycheproof.

Thanks Daniel, your explanation makes sense. Closing the issue i created for KW and KWP would make sense, for the same reason you cite.

from wycheproof.