A gap is a natural phenomenon that current science cannot explain. Let X be a gap. A God-of-the-gaps claim with respect to X – denoted by GG(X) – is a claim of the following form: All other things being equal X increases the likelihood of God's existence.
Now, I take it that GG(X) is epistemically credible if and only if
1. X is a scientifically substantial gap (e.g., the origin of life),
2. X is a theologically significant gap (idem),
3. The gap has been persistent for a significant amount of time (idem),
4. Increase in likelihood is seen as modest given that science closed many earlier gaps,
5. Increase in likelihood is not considered as being on itself an argument for God’s existence,
6. It is acknowledged that future science may close the gap and thus from that moment on (i.e., not earlier) undo the increase in likelihood.
Let GGC be the collection of all epistemically credible God-of-the-gaps claims. Let C be a cumulative case for the existence of God that is epistemically credible and that does not contain any claim in GGC. Typically, cumulative case C includes but is not necessarily limited to teleological (e.g., fine tuning), cosmological (e.g., Kalam) and ontological arguments for God’s existence.
I take it that adding GGC to C increases the epistemic strength of C, that is to say, renders God's existence more likely than before GGC was added. In this sense, and in this sense alone, God-of-the-gaps reasoning seems acceptable. God-of-the-gaps reasoning is thus not always flawed.