On its own we can't judge if this is a workable way to find vulns, as we don't know how many false positives you'd get if you ran it on all the code. (iow might be https://en.wikipedia.org/wiki/Base_rate_fallacy)