Cyanogen bromide is not a nice reagent. It’s not quite on my list of things that I refuse to use, but it’s definitely well up on the list of the ones I’d rather find an alternative to. The stuff is very toxic and very volatile, and reactive as can be.
But it’s not the worst thing in its family. A good candidate for that would be cyanogen azide, which you get by reacting the bromide with good old sodium azide. Good old sodium azide, which is no mean poison itself, will do that with just about any bromide that’s capable of being displaced at all. Azide is one of the Nucleophiles of the Gods, like thiolate anions – if your leaving group doesn’t leave when those things barge in, you need to adjust your thoughts about it. Cyanogen bromide (or chloride) doesn't stand a chance.
Cyanogen azide is trouble right from its empirical formula: CN4, not one hydrogen atom to its name. A molecular weight of 68 means that you’re dealing with a small, lively compound, but when the stuff is 82 per cent nitrogen, you can be sure that it’s yearning to be smaller and livelier still. That’s a common theme in explosives, this longing to return to the gaseous state, and nitrogen-nitrogen bonds are especially known for that spiritual tendency.
There were scattered reports of the compound in the older German and French literature, but since these referred to the isolation of crystalline compounds which did not necessarily blow the lab windows out, they were clearly mistaken. F. D. Marsh at DuPont made the real thing in the 1960s (first report here, follow-up after eight no-doubt-exciting years here). It's a clear oil, not that many people have seen it that state, or at least not for long. Marsh's papers are, most appropriately, well marbled with warnings about how to handle the stuff. It's described as "a colorless oil which detonates with great violence when subjected to mild mechanical, thermal, or electrical shock", and apologies are made for the fact that most of its properties have been determined in dilute solution. For example, its boiling point, the 1972 paper notes dryly, has not been determined. (The person who determined it would have to communicate the data from the afterworld, for one thing).
The experimental section notes several things that the careless researcher might not have thought about. For one thing, you don't want to make more than a 5% solution in nonpolar solvents. Anything higher and you run the risk of having the pure stuff suddenly come out of solution and oil out on the bottom of the flask, and you certainly don't want that. You also don't want to make a solution in anything that's significantly more volatile than the azide, because then the solvent can evaporate on you, making a more concentrated stock below, and you don't want that, either. Finally, you don't want to put any of these solutions in the freezer - a particularly timely warning, since that's one of the first things many people might be tempted to do - because that'll also concentrate the azide as the solvent freezes. And you don't want that. Do you?
Actually, the careless researcher shouldn't even work with cyanogen azide, or anything like it, but you never can tell what fools will get up to. The compound has around a hundred references in the literature, a good percentage of which are theoretical and computational. Most of the others are physical chemistry, studying its decomposition and reactive properties. You do run into a few papers that actually use it as a reagent in synthesis, but I believe that those can be counted on the fingers, which is a good opportunity to remind oneself why they're all still attached.
In fact, the reason I got to thinking about this wonderful little reagent was a recent paper in Angewandte Chemie, which details the preparation of horrible compounds like the one shown. But what does the experimental section spend the most time warning you about? The cyanogen azide used to make them. Enough said.