[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Millions of PDF invisibly embedded with your internal disk paths

(Fixing rejected post)

Meh.   I replied to something similar off-list.

"Leaking" a pdf with 'e:\nethome\joe_kitten_lover' doesn't remotely "prove" anything.  If I create a user called MayIMommaDogFaceToTheBannanPatch and "leaked" a pdf, it doesn't mean Steve Martin was culpable.  This is a non-issue, no matter how much you might want to create some fanciful "bonsai kitten" theory to get Joe in trouble, dawg.


From: WebDawg [mailto:webdawg@xxxxxxxxx] 
Sent: Thursday, December 03, 2009 1:58 PM
To: Pavel Machek
Cc: Patrick Webster; Thor (Hammer of God); bugtraq@xxxxxxxxxxxxxxxxx
Subject: Re: Millions of PDF invisibly embedded with your internal disk paths

While the risk may not be large it is still information that should not be leaked.  Leaky computers should always be plugged.
On Thu, Dec 3, 2009 at 4:01 AM, Pavel Machek <pavel@xxxxxx> wrote:

> I agree. Discovering the local path may be considered a risk, but in
> most cases the risk is nil.
Often, risk is not big, agreed.

> Considering that, perhaps for the PDF format specifically this could
> be an issue, under the assumption that consumers use PDF
> /specifically/ to prevent data leakage.
Exactly. Imagine someone posting (anonymously) copy of EvilCorp's
internal web pages, that prove EvilCorp is planning  to produce bonsai
kitten, as .pdf. If the pdf contains 'e:\nethome\joe_kitten_lover'
.. then, well, Joe has a problem.

(It would be bad if that .pdf contained username/hostname, too; I
could imagine even timestamps being problematic.)

(And yes, similar problems are elsewhere. Exif contains way too much
information, if you try to leak pictures of bonsai kitten from digital

(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html