Please add "in an enum-head or opaque-enum-declaration" after "enum-key". I think we may have duplicated this previously as a misguided and unclear attempt to address both cases separately.

Hmm, could we refactor opaque-enum-declaration to use enum-head-name, and reference that here? That would avoid confusion over whether this applies if there is an intervening attribute-specifier-seq.

Was this decltype rule for opaque-enum-declarations present before (maybe implied by pre-existing wording)? I see we had such a rule for enum-heads.

Well, we have (unchanged from before):

If an \grammarterm{opaque-enum-declaration} contains
a \grammarterm{nested-name-specifier},
the declaration shall be an explicit specialization\iref{temp.expl.spec}.

and we have

11.3 [dcl.meaning]
The nested-name-specifier of a qualified declarator-id shall not begin with a decltype-specifier.

That seems close enough to me. (it seems we don't want redeclarations to use decltype to nominate the entity being redeclared. Which is totally reasonable.)

I agree that we don't want to allow decltype here, and indeed current implementations don't distinguish between the opaque enum declaration case and the regular enum case, but if this is a normative change (as it appears to be), we should fix that portion of this via the CWG issue process.

Fixed by limiting the prohibition to the enum-head case.