changing from "NSTR constrained" to "STR constrained" does not make things any clearer, because the alternative phrasing could easily be interpreted to be meant as "STR operation is the source of the constraint" as this is a common interpretation of such phrasing without additional words to more definitively express the intent, such as "constrained by" or "is constrained"
i.e. the verb "constrained" is often accompanied by some descriptive phrase which indicates the source of the constraint, but it can be accompanied by wording that indicates what thing is being constrained and so, without some preposition or verb, it is difficult to tell what the relationship is
it's easier to just drop the word "constrained" which is what I have done
2)
while NSTR is used as an adjective for STA and because of this, you seem to be objecting to its use to describe a relationship among links, the use of NSTR for STA is insufficient to describe behavior
i.e. while a STA with no NSTR constraints can be labeled as an STR STA, and by direct comparison, a STA that is not STR is most conveniently called an NSTR STA, there are degrees of NSTR and the behavior of one STA with respect to another cannot be based only on STR STA v NSTR STA, but rather, on the particular links which are subject to NSTR - so in the behavioral section, we need to refer to links, not STAs
to keep the wording minimized, we must have a term which can be used to identify those links to which the behavior is applicable, i.e. the links that are mutually subject to NSTR
hence the term NSTR link set
so the use of NSTR to modify both MLD or STA and to modify link is appropriate
furthermore, there is ample precedent for having a single adjective which is applicable to STA and to other items as well, for example:
VHT STA
VHT PPDU
QOS STA
QOS frame
etc.