Well the danger threshold would theoretically be 4.9v (better leave a little margin though). Seems not too likely, since V increases as the log of solar intensity. The theory suggests that you'd have to about double the solar output to pick up that extra 0.2v. Doubling the solar output would mean that I'm no longer worried about my project.
The extra power from more sunlight mostly comes from increased current, which is not our concern since we are interested in OC voltage.
I suppose a possible no-cost software countermeasure would be to make the system stay on for a while if the input voltage indicates an unexpectedly full charge. Ensure that the thing is always guaranteed sufficiently discharged to be safe until the next wake up. It seems like a low probability scenario.
I guess I'll have to order it and do some testing. The specs seem to indicate that the OC voltage is independent of input light level.
I found a couple nice sources:
http://www.pveducation.org/pvcdrom/solar-cell-operation/effect-of-light-intensity