In November of 2005, Matan Gillon discovered a vulnerability in Internet Explorer in the way it handled the CSS import directive (http://www.hacker.co.il/security/ie/css_import.html). He proved the danger of the IE vulnerability by attacking Google Desktop. This proof of concept proved a powerful demonstration of the browser exploit and was successful for the same reason that our research into Google Desktop's security was successful - because a link is maintained between the public web site and the local web server. If Google had disabled this link, or given the user the option to remove the link, they would not have been vulnerable. Matan Gillon should receive credit for recognizing the danger of this intersection. We firmly believe that the integration between public web server and the local web server is a very dangerous thing. While Google have put a band-aid on the immediate issue, it becomes a matter of time before another vector of exploit (either through the browser or web application) allows similar access. Best regards, Yair Amit Security Team Watchfire (Israel) Ltd. -----Original Message----- From: Yair Amit [mailto:yairam_ng@xxxxxxxxxxxxx] Sent: Wednesday, February 21, 2007 4:03 PM To: bugtraq@xxxxxxxxxxxxxxxxx Subject: Overtaking Google Desktop Hello, A new research from Watchfire has revealed a serious vulnerability in Google Desktop. The attack, which is fully presented in a new Watchfire research paper released today (available at http://www.watchfire.com/resources/Overtaking-Google-Desktop.pdf), can allow a malicious individual to achieve not only remote, persistent access to sensitive data, but in some cases full system control as well. Google Desktop is a popular freeware desktop search tool which offers powerful indexing abilities along with an easy to use interface. In many cases, Google Desktop manages highly sensitive information. Therefore, the impact of a security breach in it is far-reaching. Google Desktop contains several protection mechanisms to secure its indexed data against remote intruders. In this paper, we present a step-by-step attack flow that circumvents Google Desktop's protection mechanisms and allows a malicious attack to take place against Google Desktop users. The attack is composed of web-application security flaws found in Google Desktop along with exploitation of Google Desktop's tight integration with the Google.com website. The paper shows that it is possible to achieve a remote and persistent access to sensitive data on attacked systems. In addition, under certain conditions, it is also possible to covertly inject and execute malicious applications on attacked systems, using Google Desktop's own features. The full paper can be found in the following link: http://www.watchfire.com/resources/Overtaking-Google-Desktop.pdf A demonstration of the attack flow can be found at the same page or at the following link: http://download.watchfire.com/googledesktopdemo/index.htm Note: ----- The Google Desktop security flaw was coordinated with the Google Security Team. Google has been responsive and recently issued a patch which mitigates the risk of the attack. We highly recommend all Google Desktop users to make sure they have an updated version installed on their system. This vulnerability was discovered by me with the cooperation of Danny Allan and Adi Sharabani. Best regards, Yair Amit Security Team Watchfire (Israel) Ltd.
