Nitrous oxide is fairly safe in comparison with other rocket oxidisers commonly used (e.g. hydrogen peroxide or dinitrogen tetroxide) but it can spontaneously decompose, like a monopropellant. This usually requires a localised high temperature to trigger the reaction but can also be caused by impurities. Any oxidiser system has to be kept clean of grease though nitrous is far more tolerant of impurities than, for example, high pressure oxygen or high test peroxide (HTP).
Alternatively, it could have been an over-pressure failure of a pressure vessel but this seems less likely due to the nature of damage reported.
A destructive (though non-fatal) nitrous transfer accident happened at Eindhoven some time back when a worn pump bearing overheated and caused spontaneous decomposition of nitrous in a bowser, so this sort of thing isn't unheard of. It is, however, very rare. Much amateur and commercial work has been carried out on nitrous oxide hybrids in recent times with no one suffering anything worse than minor cold-burns during liquid transfer.
During a hot test-firing of the rocket engine everyone would have been under cover and all operations carried out remotely (one always assumes the worst will happen during live firings). In comparison, propellant transfer is not a particularly dangerous process so there would have been more people around and no one would have been expecting anything like this to happen, especially given the benign track record of nitrous.
My thoughts go out to those guys and their families and I hope this doesn't set Scaled back too far in their programme.
B.