The author of the anonymous prologue to the Φ recension of the SibOr writes:
Ἔδοξε . . . κἀμὲ τοὺς ἐπιλεγομένους Σιβυλλιακοὺς χρησμοὺς σποράδην εὑρισκομένους καὶ συγκεχυμένην τὴν τούτων ἀνάγνωσιν καὶ ἐπίγνωσιν ἔχοντας εἰς μίαν συνάφειαν καὶ ἁρμονίαν ἐκθέσθαι τοῦ λόγου, ὡς ἂν εὐσύνοπτοι τοῖς ἀναγιγνώσκουσιν ὄντες τὴν ἐξ αὐτῶν ὠφέλειαν τούτοις ἐπιβραβεύσωσιν.
"Having found them scattered, and reading and knowledge of them in a state of confusion, I was determined to publish the collected Sibylline Oracles in a unified and orderly volume, so that, now that they are readily available to readers, they might get the benefit of them" (SibOr Prologue 8-13).
The unity that the 6th-century CE χρησμολόγος/"oracle-collector" imposed on his σποράδην . . . καὶ συγκεχυμένην/"scattered and confused" materials would seem to consist of little more than his εἰς μίαν συνάφειαν καὶ ἁρμονίαν ἐκθέσθαι τοῦ λόγου/"publishing them in a unified and orderly volume", i.e., (more or less) just as he found them, only now between the two covers of a single codex. At least, (most of) what he archived doesn't appear to be substantially different from what we know Lactantius had been reading nearly two centuries earlier. As Buitenwerf notes, "Each time Lactantius announces that he is going to quote a different Sibyl [than the one(s) he has just previously quoted], the oracle he quotes can indeed be found in another book of the extant [Φ] collection" (Book III, p. 82). Kudos to Φ's nameless Byzantine Christian editor, then, for (evidently) keeping his own finger out of the pie.
Whatever unity the Φ recension of the SibOr might have beyond that, however, it's not generic. Five of its eight constituent λόγοι/"books" (SibOr 3, 4, 5, 7, and 8) could, I suppose, be labeled "oracle collections", but I'm still not convinced that that's a very accurate or helpful descriptor for all of these texts. Some are more coherent and cohesive (and narrative) than others (SibOr 7, little more than a series of not necessarily related fragments, comes closest, IMHO, to deserving the appellation "oracle collection"), and each is at least as different from the others as any five early Jewish and Christian apocalypses, or any five modern American novels, would be from one other. If these are oracle collections, then, like the apocalypse, or the novel, the oracle collection would have to be regarded as an inherently mixed genre, by which I mean both (1) that its incorporation of a variety of genres is basic to its own generic identity, and (2) that any two given examples of the genre may or may not look anything at all like one other. One book in the collection, SibOr 6, is not an oracle collection, but an early Christian hymn. Finally, books 1-2 belong together as a single work, but it isn't an oracle collection either: it's a didactic poem masquerading as a historical review apocalypse disguised as a Sibylline prophecy.
The Φ recension of the SibOr is often mistaken for an oracle collection, but I would argue that it's not . . . in spite of the fact that it was clearly so regarded by its late ancient/early medieval readers . . . rather, it's an anthology that includes oracle collections alongside texts belonging to other genres. (By the same token, are gospels biographies or not? It all depends on what you mean by "biography". It's a similar kind of question about generic classification, here.) Dactylic hexameter verse, figurations of the Sibyl, and archive fever are, inter alia, what hold the gathered materials together, not literary genre.