How about this?
I. A true religion will show that its claims conform to the facts, as opposed to just stating them. (Showing versus saying.) Such showing will conform to the laws of logic (deductive, inductive or abductive).
II. An empty religion will just repeat its claims, without being able (or perhaps even trying) to show their factual correspondence. A religion that is based on claims that are (deliberately or not) in principle indefeasible (un falsifiable) is an empty religion. (By in principle indefeasible, I mean that no imaginable test could falsify, or verify, its claims—whether or not such a test is currently available.)
III. A false religion will have one or more core claims that have been falsified.
—Note that the above places a stronger burden on true religion (verificationism) than on false religion (falsificationism). I think that one could release this burden by allowing for cases of “equal abductivity”—that is, where the particular religion can be said to fit the known facts as well as any alternatives, even if it has not been completely verified, as long as it has not been falsified; I think this might, theoretically anyway, put a similar burden on religion as on the physical sciences. However, it is also subject to a Humean “extraordinary claims” critique. Therefore, I will add a fourth category:
IV. A valid religion is one whose claims are in principle defeasible, do not violate logic (deductive, inductive or abductive), and have not been shown to be false.