Threat model. To mount our main attack where we capture video without any external indication to the victim, we assume that an attacker is able to run native code on the victims computer as an unprivileged user. Further, we assume the code is unencumbered by defenses such as Apples App Sandbox which is used for applications downloaded from the Mac App Store but by little else. This assumption is quite mild...
...
We stress that our main result disabling the iSight LED only applies to the first generation internal iSight webcams and we make no claims of security or insecurity of later models...
Please provide proper attribution and a working link which goes directly to the published material in your post. In fact, always provide proper attribution, including a working link to the site of original publication each time you post any published material.
Thanks.
Thank you for pointing that out. I read through the entire paper except the citations. Amazing work though it is at best a Trojan in that they do have to get the user to install and run the malicious App so that part of it is activated in a VirtualBox in a virtual OS that is not OSX. One other caveat seems to be, reading between the lines, is that user had to have administrator privileges. . . and the attacker had to have it too because they mentioned the necessity to use SUDO. Had the victim user been operating as a Standard User as is the recommended practice, this would not have worked. Couple of other points. . . G5 computers cannot access the App Store, which they mention is a prerequisite for this to work, although there are other modalities to get the iSeeYou app on target G5. Biggest is the VirtualBox necessity to be running. . . That is a killer and sort of takes us back to the preparing the machine in advance to be invaded before it can be. How many Mac users are going to be running the appropriate guest OS under VirtualBox which has full root privileges (that's actually how the hardware reprogramming of the iSight camera EPROM is accomplished)?