Sarin, like insecticides such as the neonicotinoids that have just been temporarily banned by the European Union, works by attacking the nervous system, specifically by inhibiting a molecule called cholinesterase, which clears actylcholine from nerve cells. Acetylcholine is a neurotransmitter, and when cells can't clear it neurons can't refresh and the muscles stay contracted and the animal can't move or breathe. Death is usually due to asphyxia. These and similar molecules can be used on humans and as insecticides, because the neurotransmitter mechanism is evolutionarily deep and shared among animals as diverse as us and insects.
These nerve agents are very powerful, and target only a single molecule (and hence the gene that codes for it) with devastating results. This might seem quite strange, because when we look for genetic causation of many if not most devastating problems, we find that the effects vary among victims and are quite heterogeneous, usually because many causal factors have individually small effects.
Devastating effects may seem to be more likely to be simple, compared to gradual, accumulating severity over time as happens with diseases like cancer, diabetes, hypertension, or even normal traits like body weight and stature. Indeed, these conditions can at least sometimes be due to single major genetic disruptions. But that is not the rule.
In the case of nerve agents, it is likely that background genetic variation modifies the effect on individuals to some extent, but if not major then the organism will still perish and the details won't be important. Indeed there are some instances of resistance to such agents, and the genetic mechanisms is at least sometimes known.
Unfortunately, neurotransmission is rather simple and very important (which is presumably why was so widely conserved during animal evolution). If all you want is a blunt killing instrument, which is the case in insecticides--and, sadly, in warfare--it seems that there are places in the pathways to interrupt that are biologically difficult to resist.
Just think how the world might be if there were generally such effective targets for disease problems. Experience so far has not been all that encouraging, even if there are exceptions. At least, if genetic causation is too complex for such simply-targeted therapies to be common as ways to help our urge to cure, it would be nice if the same complications applied to our urge to kill.