Why do Greek textbooks and paradigm references disagree on pluperfect endings, and how do I determine which are more standard for Attic vs Hellenistic?

If you want to go digging about this kind of thing, go digging in a German grammar. Dig in something that spends 300 pages on the different variants of verb ending.

Kühner–Blass, §213.5.

The original Pluperfect Active endings in the singular were -ea, -eas, -ee(n), which contract in Attic regularly to –ē, -ēs, -ein.

The variants –ein, -eis, -ein involved remodelling of the 1sg and 2sg endings after the 3sg ending –ein, and the middle aorist –ēn, -ēs, -ē. This first shows up in Isocrates and Demosthenes—so during the Classical period in Attic; that’s why you’re seeing both taught in grammars. The –ei– diphthong spreads to the Plural in “later” authors (that is, in the Koine: Aristotle, Plutarch); those are the endings you hesitate to consider “dubious” in details.

It’s hard for me to say which should be considered standard. A historically-oriented approach will go with the older endings, so those with the etas. And grammars of Classical languages tend to be historically-oriented. That’s what Smyth lists in its summary table (§383); the variant endings in Demosthenes are mentioned in passing in §701. For that matter, I’d be surprised if the teaching of Koine features the plurals in –ei– prominently.