This is clearly Excel's problem not xlwings, I'm mostly reporting it here for future reference for others who run into the same issue. The Excel crash information points to freeing memory in code involving AppleEvents, possibly heap corruption or other memory-related bugs.
Reading different ranges doesn't always trigger the crash, for example with the same spreadsheet reading B26:C73 and then reading D26:L73 both succeed without crashing Excel. Not a Python traceback, Excel itself crashes. Causes Excel to crash if cell C26 contains #DIV/0.
