Seems more ethical and positive experiment to have the students to work with historical data, you can find a metric on how easy it is to slip a bug into the kernel, maybe they could identify problematic trends like time of day/year when bugs can slip in, or time before a release.
Do we really need someone to prove that a malicious backdoor can be added? since real bugs happen then it is clear that the system is not perfect,
Imagine this would be allowed, then all contributions would need to be checked not for the regular kind of problems but for all those stealthy tricks, so instead of working on the kernel you would need to attempt to find on how each commit is truing to trick you , or maybe this is fine on it's own but is part of a bigger plan.
Do we really need someone to prove that a malicious backdoor can be added? since real bugs happen then it is clear that the system is not perfect,
Imagine this would be allowed, then all contributions would need to be checked not for the regular kind of problems but for all those stealthy tricks, so instead of working on the kernel you would need to attempt to find on how each commit is truing to trick you , or maybe this is fine on it's own but is part of a bigger plan.