Re: Filezilla's silent caching of user's credentials

[Jeffrey Walton <noloader () gmail com>]

Hi Ryan,

No inline comments. Sorry (I wanted to reorder topics).

> I just wanted to gauge the FD community on this issue, because
> with enough backing and explanation from the security community
> as to why this is a problem, this issue may finally be resolved (it's
> been doing this for years now)
This is an alarming trend in open source software, and diametrically
opposed to the claims of "more eyes equates to more secure"", "open
source software is more secure", and "open source software fixes bugs
faster than other software models".

Is also blows away the theory of "Darwinian Software Evolution": good,
robust, secure software thrives and lesser software dies. Filezilla
and the Python example below are "proofs by counter example". It
appears the model in use is greatly influenced by popularity, which
makes it more appropriate for politicians (who tend to lie for a
living) ;)

> "I do not see any harm in storing credentials as long as the rest
> of your system is properly secure as it should be."
> Source:(http://forum.filezilla-project.org/viewtopic.php?f=3&t=17932)
That should earn the project a Pwnie Award nomination for lamest
vendor response (http://pwnies.com/).

> To me this is not only concerning, but also completely un-acceptable.
Agreed.

Other recent similar examples of this sort of response by open source
projects include "Python ssl handling could be better...", where the
Python Standard Library did not (still does not?) verify the hostname
in the certificate with CN or SubAlt name
(http://seclists.org/fulldisclosure/2010/Sep/381). The python bug was
filed in 2007 (http://bugs.python.org/issue1589).

Jeff

On Thu, Oct 7, 2010 at 11:10 PM, Ryan Sears <rdsears () mtu edu> wrote:
> Hi all,
>
> As some of you may or may not be aware, the popular (and IMHO one of the best) FTP/SCP program Filezilla caches your 
> credentials for every host you connect to, without either warning or ability to change this without editing an XML 
> file. There have been quite a few bug and features requests filed, and they all get closed or rejected within a week 
> or so. I also posted something in the developer forum inquiring about this, and received this response:
>
> "I do not see any harm in storing credentials as long as the rest of your system is properly secure as it should be."
>
> Source:(http://forum.filezilla-project.org/viewtopic.php?f=3&t=17932)
>
> To me this is not only concerning, but also completely un-acceptable. The passwords all get stored in PLAIN TEXT 
> within your %appdata% directory in an XML file. This is particularly dangerous in multi-user environments with local 
> profiles, because as we all know physical access to a computer means it's elementary at best to acquire information 
> off it. Permissions only work if your operating system chooses to respect them, not to mention how simple it is *even 
> today* to maliciously get around windows networks using pass-the-hash along with network token manipulation 
> techniques.
>
> There has even been a bug filed that draws out great ways to psudo-mitigate this using built-in windows API calls, 
> but it doesn't seem to really be going anywhere. This really concerns me because a number of my coworkers and friends 
> were un-aware of this behavior, and I didn't even know about it until I'd been using it for a year or so. All I 
> really want to see is at the very least just some warning that Filezilla does this.
>
> Filezilla bug report:(http://trac.filezilla-project.org/ticket/5530)
>
> My feelings have been said a lot more eloquently than I could ever hope to in that bug report:
>
> "Whoever keeps closing this issue and/or dismissing its importance understands neither security nor logical argument. 
> I apologize for the slam, but it is undeniably true. Making the same mistake over and over does not make it any less 
> of a mistake. The fact that a critical deficiency has existed for years does not make it any less critical a 
> deficiency. Similarly, the fact that there are others (pidgin) who indulge in the same faulty reasoning does not make 
> the reasoning any more sound." ~btrower
>
> While it's true you can mitigate this behavior, why should it even be enabled by default? The total lapse in security 
> for such a feature-rich, robust piece of software is quite disturbing, and I don't understand how the developers 
> don't think this is an issue.
>
> I just wanted to gauge the FD community on this issue, because with enough backing and explanation from the security 
> community as to why this is a problem, this issue may finally be resolved (it's been doing this for years now).
>
> Regards,
> Ryan Sears

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/