logo       
Google Custom Search
    AddThis Social Bookmark Button
-->

Re: XSS in info2ww, and some questions.: msg#00004

Subject: Re: XSS in info2ww, and some questions.
On Wed, Feb 23, 2005 at 10:19:11PM +0100, Uwe Hermann wrote:
> Hi everyone,

Hi!

> My question is whether anyone knows of an automated tool (similar to
> rats or flawfinder) which can check if a web application is vulnerable
> to (at least some specific forms of) XSS or not.

I don't know of any. When I do web application audits I either do it by 
analysing the application logic (from the outside) or through a code review 
(which is usually best)


> In my case, I would like to know if the patch I applied to info2www
> really fixed the XSS problems or not, and if it fixed _all_ problems.

You can ask here for other people to review your info2www release (with the 
patch) and try to find XSS there.

> On a related note, I intend to do some auditing of packages (Debian, as
> well as non-Debian) myself in the nearer future. What is the usual
> procedure you use when finding security issues and what do you do first:
> Inform the Security Team? Report a bug? Try to find a patch? Publish
> the issue on Bugtraq etc?

I believe the procedure is documented in the website 
(www.debian.org/security/audit) in any case. I (personally) take the 
following steps (and in this order):

1- try to produce a patch (if the fix is simple)

2- report the issue to the security team (if the bug is present in a 
released version of Debian), provide a detailed analysis and a patch

3- report the bug to the maintainer and to upstream (in all cases), provide 
the analysis and the patch.

BTW, I usually don't provide attack samples (or POC) if the bug is
self-evident. I find that a waste of time. If there is a bug in the code
and is self-evident, there should be no need to justify a fix by producing
a POC. If the implications of the bug are not all to clear, and I have 
sufficient time (which is not usually the case), I might try to produce a 
sample exploit in points 2) and 3). 

IF the Security Team says that the issue can go "public" or IF the bug only 
affects an unreleased version of Debian file a bug in the BTS with the 
proper priority so the package is not pushed to testing and, if it's in 
testing, so its removed from there before release if a fix is not produced.

As for Bugtraq et al, I don't usually post there, I leave that to the DSAs 
that the Security Team posts when a new package is published with a fix in 
a released Debian version.

> Apropos Bugtraq: Which mailing lists or other forums, websites etc. do
> you use to publish security issues which also affect non-Debian systems?

The Security Team uses vendor-sec, which is a closed, invitation-only, 
mailing list for Linux vendors (so it's not a public mailing list). If you 
wish to publish security issues which affect non-Debian systems you can 
either ask the Security Team to forward that to vendor-sec or publish to 
Bugtraq or full disclosure. I actually favor the first.

I actually don't have time to track if all vendors are vulnerable when I 
find something that is and don't contact the different distributions that 
might repackage upstream's work. I leave that up to the upstream maintainer 
and to the Debian Security Team. As for Bugtraq or full-disclosure, I don't 
send anything there myself, as I've already said.

In any case, I do this because I'm looking for low-hanging fruit (stuff 
that is easy to find and anyone can audit himself). If I were to find a 
remote buffer overflow in, say, OpenSSH, I would probably do stuff 
differently.

Regards

Javier

Attachment: signature.asc
Description: Digital signature

<Prev in Thread] Current Thread [Next in Thread>