Learn, hack!

URL

https://media.blackhat.com/bh-ad-10/Ridley/BlackHat-AD-2010-Ridley-Escaping-The-Sandbox-slides.pdf

File name

BlackHat-AD-2010-Ridley-Escaping-The-Sandbox-slides.pdf

File size

2.6 MB

MD5

73813396163191236950083786eca50f

SHA1

9dc1a91c02acdbc05e378199c53f4cd7c3063ad9

As many have predicted, 2010 will be the “Year of the Sandbox”. We will probably see many Commercial Off-The- Shelf (COTS) products using these sand-boxing technologies in the very near future starting this year. This presentation will discuss and demonstrate practical techniques for the evasion and escape of “Sand-boxing” technologies. Many techniques have been discussed but only vaguely at popular security conferences. Very little *actual* code and demonstrations have been performed. This presentation will consist mostly of demonstrations and review of actual code. I believe that most technical security talks these days don’t need to be longer than 20 minutes, so I only want to use my time to talk about real things and demonstrate real tools. I will demonstrate tools and techniques using Chromium and custom written “sandbox” examples. Some such subversion techniques discussed will be: * Injecting Interpreters into Sandboxes to test from the inside out * Using Kernel Mode debuggers to assist you (token exchange, IO, handle creation, IPC) windbg scripts incl. * Token Sniping/Stealing (whatever you call it) * Token inspection tools (includes a .h’d and dll’d version of Matt Conover’s dumptoken.c modified to include more Native API helpers) * Handle Sniping/Stealing (whatever you call it) * User32 Messaging tricks (no, not just SetWindowsHook ;-) None of these above techniques in this talk will be without example code or demonstrations! In addition to the above, this presentation will try to “fill in the gaps” where there seems to be a lot of vagaries around tokens and DACLs. Additionally I will talk about some of the practical considerations that makes deploying a sandbox with COTS products impractical on WindowsXP. There will be some other “goodies” that were also discovered in the course of this research such as: how to detect kernel mode debuggers from userspace, how userspace debugging works under the hood, (yet) undisclosed Chrome bugs, etc. I will also talk a bit about some areas of interest I wish to focus on in the future regarding these topics.