According to the story, it was when she earthed herself by attempting to turn the water supply off is when she was electrocuted. That may well have been correctly bonded? As NickD has said, testing may not have pick up this fault, particularly in dry conditions. As it was due to hidden metal grid it was not necessarily an extraneous conductive part. It was due to water leakage that extended the metal grid works potential beyond the plasterboard / wall.
As regards RCD protection, its not the be all and end all. She may well have only passed a current of 20 -25mA, not enough to trip such a
device but enough to kill.
I'm in no way supporting the electricians as I don't know all the facts, but with the same reasoning, nor am I going to blame them at this stage.
Maybe and just maybe, if the regs were not relaxed regards supplementary bonding in kitchens, this may not have occurred?