Yes, the fundamental problem is that this sort of thing has around the top of the threat list for academic computing facilities for 30 years or so (originally t

Yes, the fundamental problem is that this sort of thing has around the top of the threat list for academic computing facilities for 30 years or so (originally typically coming in to the UK from CERN). It's just that this is larger scale, possibly more automated (filching SSH keys), and has a higher profile. Despite that, the systems are normally not managed to counter the threat, running with known privilege escalations either through unpatched OS vulnerabilities or through something like the batch system. Don't trust them with anything sensitive, including credentials like typed passwords or SSH forwarding, yet people do. I have an existence proof that it doesn't have to be like that for HPC systems, even if you're not allowed system time -- in which case live patching of login node kernels is specifically necessary. Incidentally, if attackers were looking for sensitive research results from this, I think it would have to be targeted with detailed knowledge about what specific researchers were doing; after all, it's difficult enough for a typical researcher to keep track of their own stuff, and it mostly won't have look-at-me names.