The problem is that the expected metrics of these half brackets should batch those used for standard [brackets].
Of course kerning may be used ontextually to cancel the advance width (this can occur in fact with many spacing characters, see for example ‹T.› or ‹L·› or ‹L’› where the dot or apostrophe contextually become zero-width, depending on font style), but given the height of ideographic characters in the ideographic box, and of Hangul, Kanas, and other East-Asian scripts, I doubt that this use as zero-width marks added around these scripts match very well. My opinion then is that the use as East Asian mora marks is as East-Asian specific zero-width diacritics that are not unifiable with the standard spacing behavior of the half-bracket punctuation, and that don't match very well in their real semantics as punctuation marks (but this last part of the statement is questionable and may be contradicted). Note that some scripts are using non-spacing diacritics to denote punctuations rather than to denote a semantic or phonetic modification of characters (they are not part of the basic orthography of words to which these diacritics are added, and are then ignorable in plain-text searches, at least for the 3 first collation levels).
