Wraiths do not appear to be particularly vengeful (as revenants go) so a connection to wroth < Proto-Germanic *wraiþaz is not as semantically clear as it might seem.
Phonologically we run into issues too. Wroth is the expected reflex of this word in English so we'd have to look for a borrowing (we also can't rely on an e-grade from otherwise unattested Proto-Germanic *wreiþaz as this would give a form like *writh).
The (early) Old Norse reflex of *wraiþaz is vreiðr which may seem tempting, but note that it has a voiced fricative. Old English does show final devoicing of some final fricatives, but as far as I can tell this occurred well before the Old English period and only affected neuter a-stems (cf wife < Proto-Germanic *wīβą). Additionally Old Norse ei is generally preserved as ei in English spelling (with the pronunciation /iː/, cf dialectal bein < Old Norse beinn).
The fact the word wraith first appears in Middle Scots might complicate the vocalisation though. Ai /ei/ typically comes from Early Scots /aiː/ continuing Old English /aiː/, which is not how we'd expect Old Norse ei to be borrowed. It's possible that a relatively late borrowing via Norn might avoid these issues with the vowels, but I don't know the detail; and it would still leave the issue with the voiced fricative.
Other Germanic cognates of wroth all have the wrong vowel and are unlikely to have had contact with Scots.
So a connection between wraith and wroth is difficult, albeit not necessarily much moreso than vorðr or arrach.