I recall reading (but cannot find at the moment) a different study looking at sunlike stars with known exoplanets that did not involve the kepler data. It found that there was a preponderance of super-flare prone stars amongst sunlike stars with hot-jupiter-style large close in exoplanets, adding fuel to the idea that magnetic interactions cause a subset of them. Might not account for all of course even though the Kepler data would only actually see a tiny fraction of the hot jupiters that exist in its field of view.

Stars as gravitational sinks that allow strong energy release just from objects falling onto them is an interesting concept. I presume that what makes such an event extremely rare is the very small amount of angular momentum you need to actually hit one and the fact that light pressure and sublimation will tend to destroy or deflect most objects long before they come close to anything like an impact?