The way I remember it, and I may be mistaken, that km > ngw was superseded by km > nk because Tolkien was unhappy with the former. That would not invalidate attested words using the former route, IMHO.
Whether Quenya would typically make compounds like your example, I don't know. Perhaps it would just stick to separate words.

Also, it seems that NIK actually yielded a suffix (-inke) rather than a prefix, so perhaps *mapaitinke would answer your needs?

this question?
http://eldamo.org/content/words/word-2143444883.htmlcover your last

+Tamas Ferencz Thank you: I didn't know that such material existed in Eldamo: it's very useful and indeed covers my last (third) question. Regarding km, if metathesis occurs first (km > mk) then that page also gives mk > ŋk and ng > ŋg, providing input for my first and second questions. Metathesis is a stub entry referencing PE19 which I don't have but there must also be other relevant sources. Do you recall where you may have heard about the supersession km > (ngw >> nk)? I'm interested.

As for other approaches to my toy compound, I agree that there are several viable alternatives. As for direct prefixion of √NIK there is at least ✶niktil > nihtil from VT47:26 although ​-til is a suffix there. VT48:18 gives the original source material on √NIK and VT48:15 references a prefix piki-​ and prefixed adjectives pitya >> nitya in compounds (WJ and PM are original source references), so perhaps *niki-​ could be an updated prefix and nitya is attested as the first element in a compound (Nityafinwe). Certainly the suffix or separate words work fine. In this case I'm mainly trying to more deeply understand how such compounding works for potential future non-toy situations.

I think it was in one of the later VTs (49?) where I read about Tolkien moving away from km > ngw. I am sure someone will correct me soon enough. I can't access my VTs at the moment.

OK I have found my VT after all; the note I was thinking of is in VT49:46 concerning the root YAN, and does not specifically mention km, but tm > nw, and a 'general dissatisfaction ... with the development o consonants in contact with nasals".
So km > ngw might be OK. Not sure.

+Tamas Ferencz I agree: it seems inconclusive. As Tolkien says of this topic on page 54, "What a muddle" :)

Two notes:

1) The phonetic notes in Eldamo are very much in a “rough draft” state and should not be relied on without verification against original sources.

2) The rejection of km > ngw is mentioned in multiple places, notably PE17/44 and PE19/85. The new development appears on PE19/85 - km > kw. Often, though, km would undergo metathesis to mk > nk.

The most notable example of the phonetic change is Q. tengwa

eldamo.org - Eldamo : Quenya : tengwa

The original primitive form of this word was tekma, but after the revised phonology, Tolkien invented a new primitive form tenwa.

+Robert Reynolds If you're interested in PE19, I have a copy to spare (I accidentally bought two back in the day).

+Matt Dinse I am interested and have PMed you. Thank you!