Sort of command that would pull a download that is self executed to the host machine?
That’s worded a bit fucky, if I need to elaborate, please chime in.
Sort of command that would pull a download that is self executed to the host machine?
That’s worded a bit fucky, if I need to elaborate, please chime in.
Sounds like you are trying to setup a keylogger of sorts…
Not quite, PC gets hacked, on hacked machine someone does something like cd, but on that PC cd has been set up as an alias for some sort of command that downloads a malicious executable to the hackers machine and executed it.
That executable very well could be a keylogger, but doesn’t necessarily have to be. It could be be rm -rf --no-preserve-root / or a reverse shell or whatever really.
I imagine cd would be a terrible choice to alias given how much it’s used, but maybe something else more obscure could be used that is frequently used when bots/attackers are rummaging through files for stuff to steal.
I mean, I’m sure it’s possible, it’s just a matter of how to get the honeypot/script on the system and give the downloaded file executable rights.
Is there anything that is specific to ssh that would allow for this? Like a command that would allow something to execute back to the other machine similar to downloading? I’m not well versed just kind of a napkin idea I thought of.
you can’t know what system architecture is on the ssh client side. It could be a mac, a windows or linux box, android or ios, beos, amigaos, etc.
So even if you could get ssh to pass malicious code how would you know what code to pass?
What you’re asking for is nigh impossible i think.
And when you say “can’t know” do you mean it would be impossible to tell strictly through SSH?
the ssh protocol doesn’t provide any mechanism for your server to identity the client host architecture, nor does it provide any mechanism i’m aware of to send any payloads back to the host client for arbitrary execution.
The thing is ssh is a protocol and it’s implementation is not going to be identical across clients.
So if you found a bug in an ssh client that allowed for this your hypothetical attack vector would only work on intruders using that broken client. ssh is one of the most scrutinized protocols there are and you should be a hero if you found this bug. what you’re asking for, if it were as easy as you might hope, well ssh wouldn’t be the ubiquitous security connection protocol if it were simple like that right?
Very true. Thanks for the education. SSH to me is just magic portal that lets me talk to my server in my closet lmao
it’s a fun thought experiment as it’s a very complex protocol
I kind of figured it would be a shot in the dark, some scripting could definitely be done to assess that, and even run code per major OS depending on some automated recon.
Let’s say you’ve got that figured out, and the user is running putty on windows as an administrator. Is there anything that could take advantage of that fact?
I feel like this would be way easier/more feasible to run a script on your own machine as a defensive measure like OC mentioned early, but just more asking our of curiosity. I’m not skilled enough to even imagine what to do with this or write it, but I am fascinated by security stuff.