I don't classify it as a problem, it's just an interesting idea. When I use RSA authentication with ssh-agent, and I use more multiscreens, I used to have run more instances of ssh-agent. I said there's nothing wrong with that, but there may be more convenient just to connect to the pipe of already-running ssh-agent from another multiscreen than typing always my RSA passphrase. I know that there may be a security hole with some pseudotty's or serial lines, that may use my login and password (which is simplier to obtain than a RSA passphrase) and also connect to a pipe, but limiting this to specified group of terminals could solve this problem. Any ideas ? mates <[EMAIL PROTECTED]>
