No experience with keeping them myself, but, for boxfish specifically, these guys aren't toxic when they die, they're only toxic when alive and stressed (they have to be alive to produce the toxin, and they only produce it when stressed). The toxin they produce is a potent ichthyotoxin called Ostracitoxin or Pahutoxin (ichtythoxin meaning it's a toxin that primarily effects fish, though this toxin has been demonstrated to slowly affect a wide variety of inverts too). In small quantities, the effects may be mild as long as the toxin is promptly removed, but the effects of it on fish are irreversible (meaning the fish - if they heal from it at all - will recover over a long period of time, and they will only recover if the damage is mild and the toxin is no longer present in their environment).
When the toxin is present even at 5ppm in the water (the equivalent - if my math is right - of ~3.4ml of the toxin in a 180 gallon tank), 50% of the following species of fish died within the following times*:
Abudefduf abdominalis - 6 minutes
Acanthurus sandvicensis - 8.5 min
Kuhlia sandvicensis - 10 min
Mugil cephalus - 12.5 min
Mollienesia litpinna - 15 min
Bathygobius fuscus - 30 min
Given that the damage is irreversible and lethal even at relatively small doses (and that boxfish themselves aren't immune to the toxin, though they are more resistant to it than other fish), I'd guess it's probably not an overblown risk (though it is something you could likely try to prepare for by running carbon and having a water change and QT ready at all times).
That said, I don't know how fast these guys produce the toxin, but as long as the fish doesn't get too stressed, it should theoretically never produce enough toxin to cause an issue (though I'd constantly run carbon on the tank just to be safe).
* The study I pulled the data from:
Donald A. Thomson, Toxic Stress Secretions of the Boxfish Ostracion meleagris Shaw, Copeia, Vol. 1969, No. 2 (Jun. 3, 1969), pp. 335-352
www.jstor.org