The talk was about how (some) doors and padlocks can be easily opened. The presentation was full of videos and explanatory schemes. For the doors the following parts can be attacked:
- hinge removal – to fix, use jam pins
- the latch
- the inside handle
- key boxes
- edge baps – request to exit sensors
- the bottom gap
- the door frame
The presentation contained the following topics:
- memory anti-forencics
- the goal is to inhibit the acquisition and analysis
- for windows, removing PE header from disk (once the executable is loaded in memory).
- for windows, zero the header from disk (once the executable is loaded in memory).
- for linux, remove the EMF header
- for linux, zero the header (memeset)
- android anti-forestics
- use encryption to protect.
- power down the device.
- leverage device sensors; foe ex: if the phone is moving, then shut down the device
- fun with sd cards
- demo of the SDTool tool that modifies the SD card firmware to write/or not the card or in memory.
The sql injection vulnerability is dead due to the massive use of the ORM frameworks, the same for the XSS injections due to the mvc, templates and default HTML So, as a hacker you must find new vulnerabilities; here are 5 (esoteric) vulnerabilities:
- aggressive input decoding; nosql injection using ruby on rails and MongoDB
- call me to verify my identity; try to hack the phone activation procedure for a 2 FA functionality.
- password reset implementation feature; try to hack the password reset feature for a 2 FA
- hack around the usage of the Paypal IPN protoco
- just missed that one; it happen even to the best of us.
The idea is to retrieve passwords stored by the browsers (in RAM) by scrapping the RAM content. Do to this a plug-in to Volatility framework was created (the plug-in will be available soon on GitHub).
The best browser is Chrome with 67% chances to expose the passwords; the worst browser is Firefox with 81% changes to expose the passwords.
Vendors response to this findings; Microsoft created a CVE and the path will be pushed in October/November, Google and Mozilla are denying that’s a real issue.