I think we've all heard the axiom about dosage being everything.
This can be proven wrong if there is even just one substance where the indivisible part of it would be fatal.
Firstly, dosage clearly isn't everything since the substance being dosed is pretty damn important as well. Secondly, this wouldn't be proven wrong by finding something that could be fatal with one molecule. All that would prove is that one particular substance is so deadly that even the minimum dose possible is still fatal. The does would still be important for things like the speed of fatality and the possibility of an antidote.
Finally, no. It is simply not possible for a single molecule to kill someone. There are a huge number of cells in the human body, and killing one or two will not even be noticeable, let alone fatal. Just look at the example of botulinum toxin Gravy provided. About 10
-7g/kg for a fatal dose. That makes something like 10
20 molecules. Single atoms and molecules are just not relevant when it comes to living systems.
Of course, as already mentioned, if you include self-replicating systems then introducing a single one could result in it reproducing and building up to a fatal dose. However, I wouldn't count this as a single molecule being fatal, since death won't occur until there are a hell of a lot more than one of it.
Edit: I suppose if you want to be really picky there are some possibilties. For example, ionic substances exist as crystals. Although various flaws usually mean the crystal we see are made up of many different small crystals, it is entirely possible for a single macroscopic crystal to be a single molecule. So you could technically die from having a single molecule of salt dropped on your head. The same sort of logic can apply to various other things as well, carbon, for example, can form arbitrarily long chains as far as we know. However, most people probably wouldn't count this sort of thing as an answer to the OP.