Re: HTTP based trojans

["s.wun" <s.wun () thales-is com hk>]

I think this so-called flow-based IDS is about analyse each end-to-end
connection based on what protocol the connection is using. For example, if
protocol is 6, it should follow standard TCP communication standard,
anything other than that will be regarded as Potential hack. That's why in
http connection, it detected communication is not belong to http, so it
should be able to raise alarm.

One can create this kind of analyse with simple programming, not neccessary
to purchase StealthWatch if we understand the principle of it.

sam
----- Original Message -----
From: <AQBARROS () BKB com br>
To: <focus-ids () securityfocus com>
Sent: Wednesday, November 06, 2002 8:56 PM
Subject: RES: HTTP based trojans


Good question! It's just what I want to know, but it seems that my question
did not raise a discussion.

People has been using HTTP based trojans for some years, but only after the
Sensepost Black Hat presentation about Setiri it has become a major point of
discussion.

I didn't see nobody sharing ideas about detecting (or even blocking) this
stuff. I can imagine a couple of Snort rules to try to detect it, based on
filenames and paths, like cmd.exe, \winnt, etc, but it would find a lot of
false positives and wouldn´t be effective on cases using SSL. So, perhaps
the point is on HIDS; But how can we detect the abnormal behaviour if the
trojan is getting out through a IE window? Which adverse effects there will
be if we block the use of invisible IE windows?

Regards,

Augusto


-----Mensagem original-----
De: s.wun [mailto:s.wun () thales-is com hk]
Enviada em: quarta-feira, 6 de novembro de 2002 0:27
Para: AQBARROS () BKB com br; focus-ids () securityfocus com
Assunto: Re: HTTP based trojans


Hi,

What other open-source tool do you use to detect this attack?

Sam.
----- Original Message -----
From: <AQBARROS () BKB com br>
To: <focus-ids () securityfocus com>
Sent: Thursday, October 31, 2002 8:46 PM
Subject: HTTP based trojans


> As I saw on the last messages about detecting trojans through flow-based
> analysis, I thought if someone already made anything to detect trojans
that
> use Internet Explorer controls to communicate with the client, even on
> networks that allow only proxied (even authenticated) http connections.
Did
> anyone try to do such kind of thing?
>
> Regards,
>
>
> Augusto.