As far as instrumentation error, the stall being triggered at a 6-degree angle of attack (AoA), suggests momentary updraft. I wonder how granular those data are. The stall initially occurred at (going from memory here, I might be off a couple thousand) 38,000 feet. At that altitude, it doesn't take a lot of AoA to generate a stall condition.
In the "cookie corner", indeed not, and in the context of a big thunderstorm possibly throwing ice out of its anvil, a stall warning could throw off many a green pilot. Also, at such an altitude, at the rather nose high attitudes consistently mentioned in the BEA's description presumably can't be sustained without exacerbating if not creating a deep stall.
HF