The name of one of the authors of the Hebrew book "Mabruk : minhage ha-hatunah ha-'Arvit be-Yisra'el" [diacritics omitted throughout], appears in 3 scripts:
- Kusai Hag' Yahya [in Hebrew, on the t.p.]
- Kussai Haj-Yehia [in rom.] and
- Qusayy Hajj Yahya [in Arabic, on the t.p. verso]

RI22.3C refers to "Names of persons with surnames written in the Hebraic  alphabet".
The next sentence states: "Follow this order of preference for Hebrew and Yiddish persons with surnames".

Who is a "Hebrew person"?

The name of one of the authors appears  in  three scripts  in prominent sources:  Hebrew, roman,  and Arabic.   The  name  is  clearly Arabic  (so, presumably, the  author  is  "an  Arab," whatever that  means  in  this particular  case).  Which form of the name should become the heading?

The presumption in AACR2 is that a person writes in  his own  language, unless there's evidence to the contrary.
Such evidence  might  be,   for  example,  a  statement  that  the  work  is a  translation,  or  knowledge  learned  from   a  reference  source  (encyclopedia, bibliographic  database,  etc.)  that  he  usually  writes in some other language than the  one  in front of you.  If  there's no  such evidence, the author of  a  book  in  Hebrew  is  presumed to be a "Hebrew author" (or, in the words of RI 22.3C, a  "Hebrew person"), and  should  be established as  such  per  rule 22.1B.

In the above mentioned case, where the work is not a translation and we  know  of  no other books by  the author, he should be established  (paradoxically) neither from the Arabic-script form of  his  name nor the Hebrew one, but from the roman one, per RI 22.3C.  The "ethnicity" of the author is not relevant to  the choice of form.

------------------------------------------------------

In order to establish an author who wrote in Judeo-Arabic, CPSO was asked how strictly we need to interpret the section of RI22.3C on "Names of Persons with Surnames Written in the Hebraic Alphabet."  As you know, the section begins by referring to "Hebrew and Yiddish persons."

According to CPSO, we should interpret this much more liberally than first appears, and apply the provisions of the rule to persons with surnames written in the Hebraic alphabet even if they're not "Hebrew or Yiddish."