Have they not shown the naked RNA genome of many viruses to be infectious in and of itself?
		
		
	 
RNA is a terribly fragile molecule. Of course, in the most insidious of retroviruses, we know it to be very dangerous. This is why we are covered in RNAses, and thankfully, tiny amounts of RNA will essentially degrade to uselessness in a manner of seconds when exposed to simple room temps. 
Engineering a protein coat to protect RNA, and delivery mechanism in some sort of weaponized form (as I think you are suggesting) is beyond our scope. ...not to mention engineering a stable, infectious, RNA sequence. 
But I'm no DoD-funded virologist, so I could be totally wrong. :\
My assumption is that research like this tends to be highly regulated. With the type of restrictions placed on studies involving controlled narcotics, I would assume that the availability of data that would allow for such dangerous engineering would be even more guarded.