[119003] in Cypherpunks
Scheme to defeat all keyboard sniffers
daemon@ATHENA.MIT.EDU (Gary Jeffers)
Tue Oct 12 14:26:34 1999
Message-ID: <19991012181229.40347.qmail@hotmail.com>
From: "Gary Jeffers" <jeffersgary@hotmail.com>
To: cypherpunks@cyberpass.net
Date: Tue, 12 Oct 1999 13:12:28 CDT
Mime-Version: 1.0
Content-Type: text/plain; format=flowed
Reply-To: "Gary Jeffers" <jeffersgary@hotmail.com>
Correction on last post this subject:
It would be sufficent to just change one cipher character in order to
defeat a keyboard sniffer. My original post was correct. This, of
course, assumes that only the keyboard buffer was sniffed. A monitor
sniffer would be a more unlikely attack but THE ANTISNIFF as now
designed would not defeat it.
If a monitor sniffer were installed, I would think it would use
Interrupt 10h, Function 08h (Intel type machines) "Read character
and attribute at cursor position" to sniff.
The above type monitor sniffer could be defeated by using "icon"
characters rather than the standard write to screen characters.
The above scheme could, in turn, be defeated by a sniffer that
recorded or examined pixels. But that would require the attacker to
either copied very large numbers of screen data (recorded) or
did a huge amount of computation (examining).
Generally, monitor sniffers would attack THE ANTISNIFF (as
currently designed) by doing a frequency analysis of characters
of data abstracted by correlating key presses with static and
varible (cipher type) characters.
Yours Truly,
Gary Jeffers
Expert: A guy who knows his subject from b to z. They are every-
where!
BEAT STATE!!!!
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
