詮索好き	        4,047	91.7%	
せんさく好き	364	8.3%

Most of the kokugo definitions for -好き terms include "また、その人". The original glosses were probably OK.
As for part of speech, some are nouns only, others are 名・形動. Meikyo has 「━の／な 連中」as an example for 酒好き, so adj-no is at least sometimes valid. Needs to be handled case by case.

@@ -8,0 +9 @@
+<ke_inf>&sK;</ke_inf>
@@ -17,0 +19,5 @@
+</sense>
+<sense>
+<pos>&n;</pos>
+<gloss>nosy person</gloss>
+<gloss>busybody</gloss>

ルミナス (詮索好きは身を誤る  Curiosity killed the cat.), Reverso
Tanaka: 私の隣人はいわゆる詮索好きなのです。 My neighbor is what we call a nosy person.

It's a noun, and it's the condition not the person.. In cases like this we don't usually add adj-no.

@@ -13,0 +14 @@
+<pos>&n;</pos>
@@ -15,4 +16,2 @@
-<pos>&adj-no;</pos>
-<pos>&n;</pos>
-<gloss>nosy person</gloss>
-<gloss>busybody</gloss>
+<gloss>curiosity</gloss>
+<gloss>inquisitiveness</gloss>

詮索好き	4047	  
詮索好きな	1638	  
詮索好きの	776	  
詮索好きで	452	  
詮索好きな人	211

Search: 詮索好きな隣人

I still haven't wrapped my head around what exactly happens when you drop が and change すき　to ずき.  I know in principal that you generally gain the ability to use の as a connecting form(where 好き is usually な), but that な also remains common.  This could be because the ずき form is now a noun, and の is how you connect nouns together. (I would think this would support almost all of these entries being [adj-na,adj-no,n]).

It certainly seems like most of these remain [adj-na] whatever the dictionary might otherwise suggest...  We aren't handling these consistently. Here's the first 1/3rd of our entries from an advanced search "ends-with-好き", "ends-with-ずき":
[n] 下手の横好き, 人好き, 子供好き, 客好き
[n,adj-na] 世話好き, 出好き
[n,adj-no] 女好き
[adj-na] 争い好き, お客好き, 交際好き
[adj-na,adj-no] いたずら好き
[n,adj-na,adj-no] きれい好き

These distinctions don't generally seem to be supported by the ngrams.  To take one example of an entry listed currently as [n]:
人好き	61235	  
人好きな	22042	 <- な is 1/3rd of all occurrences
人好きの	5929	  

One more (currently [n,adj-no])
女好き	85862	  
女好きの	14583	  
女好きな	10827	  <- な only marginally less common than の

I'm afraid I don't understand the grammatical nuances well enough to say what *should* be the case, but it is my impression that (at least grammatically speaking) most of the 好き/ずき-suffixed forms should behave the same in most cases.  But our entries are all over the map.

Does anyone have a better understanding here than I do?

We only have about 30 entries of this kind.  I'm willing to print all 30 on a sheet of paper and start surveying Japanese people ;).

@@ -13,0 +14,2 @@
+<pos>&adj-na;</pos>
+<pos>&adj-no;</pos>

Entry created