"A tiny bit in a hollowed out bullet can't destroy more than a bullet sized hole to begin with..." How do you figure that? Sure you're not going to end up with grey goo but you're still going to end up in a world of pain if the disassembler nanites have been programmed to consume organic material as well. Likely though not at a 1d10/2 per round though. The pages for reference are pp. 328-329.

The rules are not very clear on inventive applications of nanoswarms and nanoware. You're going to have to make a lot of this stuff up if you want to be creative. [s]We don't know how long nanoswarms last for, how big nanoswarms are, how long it takes to produce a nanoswarm, etc. If we had more information, we could come up with better estimates.[/s] Edit: However, I imagine that a transhuman nanoswarm delivered by a bullet would be less effective than a Titan nanoswarm. Edit 2: Found some of the information I was looking for. I thought there was less information available than there actually was. See pages (core rulebook, p. 328) for nanoswarms, and (core rulebook, p. 341) for the buzzer spray weapon (it shoots nanoswarms). I still would like to know how big a swarm would be if they clumped together as a ball.

Things of interest. The bio-defense swarm (gatecrashers, p. 157) is used by gate crashers to purge an 10 meter radius area of alien mater. If an alien creature were to walk into the area, it will inflict 1d10 / 2 damage per turn. It lasts 2 weeks without a hive. This swarm can also destroy armor, inflicting its damage until it gets through, at which point it starts damaging the host. The buzzer (core rulebook, p. 341) is a gun that is equipped with a specialized nanoswarm hive. It will literally shoot a nanoswarm at a target as a spray weapon. It can produce 1 swarm per hour. Reading (core rulebook, p. 328) provides some of the information I thought was missing. I still would like to know how much room a compacted swarm (if they were to clump together like a ball) would occupy.

Simple physics argument: a disassembler swarm is essentially a very odd corrosive chemical (with very big molecules) that uses its own stored chemical energy to break down chemical bonds. The swarm stores energy in some chemical form (whether nanotech springs or high energy molecular bonds - carbon nanotube springs can store 7.2 MJ per kg, fat 37 MJ/kg, gasoline 47.2 MJ/kg - the later two requires oxygen to react). This is means that unless the swarm can refill energy or is very selective in what it can break down it will not be able to break down a much larger mass of matter than itself. So the bullet might take a few grams of armour and body with it, but it will not disintegrate a kilogram chunk. Note that this is a disassembler swarm made to disassemble anything. Suppose you instead just want to disassemble protein. Then you will hydrolyse peptide bonds, which in theory should net you some 8–16 kilojoule/mol, but most likely you will need to supply some energy to mince them up (and move around, identify proteins etc). Let's say you pay 100 kJ/mol to cut the bonds. Collagen (tropo-collagen) has a molecular weight of around 300,000, so 1 kg contains 2e21 proteins, each with about 1000 amino acids. So the energy to dissolve one kg of protein will be 333 kJ. A 70 kg body is 67% water, and 75% of the dry weight is protein = 35 kg. So to liquefy a person you need around 12 MJ. (that energy will of course end up as heat - a rapid disassembler will also tend to boil you) Assuming a 40 MJ/kg nanofuel in the disassembler swarm, this gives 1 g of protein disassembler bullet (40 kJ) the ability to munch up 120 gram of protein. Or, assuming the above dry weight thing, about 266 g of body. Ouch. That is a decent chunk. Still not total disintegration, but given that the disassemblers will no doubt be carried around in the bloodstream and poke holes in lots of places (in the heart, in the liver, in the brain...) it is a nasty weapon. Summing up, here is my take on how I would handle it ruleswise: the nanoswarm subtracts its DUR from the victim's DUR until it runs out. Say 1d10 points per round or so. Processes that remove it - nanodefenses, bleeding, cutting off infected limbs, can reduce the swarm DUR before it does more damage. Fast nanoswarms might do more damage per round but have less overall DUR for damage purposes (since they tend to cauterize their environment), while slow and insidious nanoswarms may be slower but have larger DUR (since they only break down particular choice molecules, like DNA or ribosomes).