How would an attacker have used this vulnerability? It seems to me that they'd already have to be able to inject code into the chrome context, to be able to point it towards content they control. Maybe I misunderstand how this works.