Shell “stabilization” refers to the process of making a remote shell behave like a normal local shell — so, allowing interactive programs to work properly, ensuring that input is not echoed inappropriately, etc. In practice, this generally involves creating a second connection from within the “unstable” shell, and then using that (keeping the first connection around just so you can restart the “stabilized” shell if you accidentally exit/kill it).
A common method of stabilizing netcat shells is to use Python:
- Start an instance of Bash connected to an actual PTTY:
env TERM=xterm python -c 'import pty; pty.spawn("/bin/bash")' - Suspend the reverse shell.
- Use
stty raw -echo; fgto switch to raw keycode transmission (so that things like arrow keys get pushed to our remote shell), turn off terminal echo (to prevent seeing commands twice), and foreground the reverse shell.
Note that the stty command can be canceled using reset (after closing the reverse shell). Since echo is turned off, typing this won’t be visible. Trust the force!
The rlwrap package will handle almost all of this for you.
rlwrap -cAr nc -lvnp $PORTOr just use socat!
IMPORTANT
In none of these cases will the reverse shell pick up on your terminal size, so you’ll need to manually specify it using
stty rowsandstty cols.