That looks pretty thorough to me, and though I can't see anything wrong with it, there are several misleading things.
The examples of corpus/corpora and opus/opera are irrelevant. Those are from stems corpos- and opos- and have plural ending -a just like every other neuter. The adjectives are corpor-al but vir-al (not viror-al).
Any mention of vocative or accusative by analogy with masculine is irrelevant: neuters always had nom/voc/acc identical. The accusative and vocative of virus _must be_ virus, whatever its declension.
The Greek pelagus and cetus aren't relevant, because the Romans fitted Greek words into Latin declensions when they fitted easily, but kept the Greek endings otherwise: hence Greek plurals pelag-e, cet-e. The Greek for poison was ios, cognate with virus. (A quick look at the Greek dictionary tells us it's masculine, with no plural given, so that doesn't help.)
Fourth declension? My trusty Latin grammar (it had bloody better be trusty) says the neuter ending was -u, as in cornu = horn, gen.sg. -us, plural -ua as in cornua. As the word is virus, not viru, that seems to rule out virua.
I think it's a rule that neuter plurals always end in -a, regardless of the declension (sometimes -ia as in mare/maria, animal/animalia, and sometimes -ua as in cornua, but always final a).
