I wonder if, as another freeper described it on another thread, many of these chemicals are “binary” - that is, 2 stable compounds with a long shelf life are stored separately, then mixed when the resultant agent is desired. The images on another thread of 2-sided containers lends to that possibility.
So, the gas could have a short shelf life, but the compounds that, when mixed, create the toxic gas are more stable.
I'll bet you're exactly right. The Wikipedia article did mention agents being separately mixed, but I didn't connect it with longer shelf life.