The first criterion I found particularly helpful was the distinction of whether or not a given text was intended to replace a prior "biblical" text. A standard copy of a "biblical" book would be intended to further the transmission of the prior work and replace it as a newer copy. On the other hand, the nature of some works is that they are not intended to replace the prior work, but rather to complement the prior work and be read alongside of it. This criterion, to me, seems to be the determining factor in whether a book should be considered "biblical" or "para-biblical." It may not always be easy to tell the intention of a work (or its reception in a community), but it is nevertheless important.
The second criterion I found helpful was the distinction of whether or not a scribe felt free to invent new material. While some scribes carefully copied without alteration the text before them, others felt free alter the text with information culled from the text itself. For instance, some may have chosen explicitly to harmonize passages assumed to be harmonious in the prior work. Others may have rearranged material from the prior work to emphasize certain points or group thematically similar passages. Still others may have explicated clarifying information interpreted to be implicit in the text. But in each of these cases, the scribes were acting on a relatively conservative impulse, that, while permitting some changes, restricted those changes to material available (explicitly or implicitly) from the prior work. On the other hand, some scribes may have felt free to add their own inovations to the text that could not be derived from the prior text itself. They may have incorporated additional stories familiar through oral or written traditions or even composed them themselves. They may have reworked the prior text into an entirely new literary structure with its own rhetorical aims independent of those of the prior text. Each of these scribes would have engaged in textual innovation, adding material not accessible from the prior text itself.
Based on these two primary criteria, I think we can find a helpful paradigm for sorting out "biblical" texts from other types.
(copy) (e.g., excerpt texts)
Innovative "biblical" "para-biblical"
(new edition) (e.g., rewritten Scripture)
This paradigm allows us to discern four distinct categories.
Conservative Substitutionary - These are simple copies of "biblical" books which are intended to replace other copies and do not add significant information beyond what is inferable from the prior text. They are clearly to be considered "biblical."
Innovative Substitutionary - These are copies of "biblical" books which are intended to replace other copies but also add significant information beyond what is inferable from the prior text. They are to be considered "biblical," but their innovations identify them as new editions of the prior "biblical" books. Perhaps this could provide a relatively simple definition for the otherwise difficult to define "variant editions" of "biblical" texts?
Conservative Complementary - These are new compositions intended to be read alongside the prior "biblical" texts, but which do not add significant information beyond what is inferable from the prior text. Examples might be excerpt texts, which select passages from "biblical" texts without significant modification and rearrange them for theological or referential purposes.
Innovative Complementary - These are new compositions intended to be read alongside the prior "biblical" texts and add significant information beyond what is inferable from the prior text. Examples would be those works commonly included in the category of "Rewritten Scripture" (e.g., Jubilees, Genesis Apocryphon, etc.), which are intended to be read alongside their prior "biblical" books and provide additional information.
While these criteria (as with all criteria) are bound to encounter complicating factors, I suspect they accurately describe the differences between the different classes of literature. I further suspect that these qualitative criteria lead to more helpful distinctions than a sliding scale or spectrum based primarily on quantitative criteria.