bugtraq - RE: Possible XSS in User-Agent

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date: Mon, 25 Apr 2005 15:22:12 -0400
From: "Scovetta, Michael V" <Michael.Scovetta@...com>
To: "Nicolas Montoza" <xonico@...il.com>,
	<bugtraq@...urityfocus.com>, <full-disclosure@...ts.netsys.com>,
	<news@...uriteam.com>, <bugs@...uritytracker.com>,
	<submissions@...ketstormsecurity.org>, <vuln@...unia.com>
Subject: RE: Possible XSS in User-Agent

Nicolas,

This is no more an XSS than creating and misusing an arbitrary header:

GET / HTTP/1.1
FooBar: <script>alert();</script>

...
<%=request.getHeader("FooBar")%>

The headers are **always** modifiable by an advanced user, a script-kiddie with a utility, a clever firewall, a proxy server, etc. Just because information resides in the header doesn't mean that it doesn't need to be validated, just like everything else.

This is not a "bug" in any language, browser, or protocol. As per RFC 2616, the User-Agent field is defined as:

       User-Agent     = "User-Agent" ":" 1*( product | comment )

   Example:

       User-Agent: CERN-LineMode/2.15 libwww/2.17b3

So in other words, don't just echo input from the browser back to the user. This is no more than a trivial example of "Why You Should Validate Your Input". This could be considered an XSS vector into applications that fail to validate, but no more so than any other header (or GET/POST variable).

Regards,

Michael Scovetta
Computer Associates
Senior Application Developer

-----Original Message-----
From: Nicolas Montoza [mailto:xonico@...il.com] 
Sent: Monday, April 25, 2005 12:05 AM
To: bugtraq@...urityfocus.com; full-disclosure@...ts.netsys.com; news@...uriteam.com; bugs@...uritytracker.com; submissions@...ketstormsecurity.org; vuln@...unia.com
Subject: Possible XSS in User-Agent

Analyzing User Agent does not make filters of anyone type, being able
to inject xss or HTML.

POC
===

let us suppose that the page we visit has the navigator´s check

You are sailing with Mozila Firefox....

In php, this simply is

<? echo $HTTP_USER_AGENT ?>

then we install any kind of soft which allows us to modify the user
agent,  in mozila _firefox you could use this plugin

https://addons.update.mozilla.org/extensions/moreinfo.php?id=59

Example:

USER AGENT: <h1>Soulblack</h1>
USER AGENT:<script>alert('SoulBlack')</script>

it works correctly :).

The logfile of apache ;

127.0.0.1 - - [23/Jan/2006:14:54:02 +0000] "GET /favicon.ico HTTP/1.1"
404 283 "-" "<script>alert('SoulBLack')</script>" "-"

the tests were made with php and apache.

The bug could be in php, or in the protocol , we have not even probe
in another language like asp , etc ...
if the bug resides in the protocol, the model of control of user agent
could  be   [a-z][0-9] .

Any suggest or comment?

POC created by Soulblack Group.
www.soulblack.com.ar

--
 SoulBlack - Security Research
 http://www.soulblack.com.ar