Microsoft is planning an update to the cross-site scripting filter in Internet Explorer 8 to address an attack scenario revealed at the Black Hat Europe security conference.
At the conference, security researchers David Lindsay and Eduardo Vela Nava revealed details of how the filter detects attacks, as well as its strengths and weaknesses. The weakness included ways the filters could be abused and bypassed to enable cross-site scripting (XSS) on sites that would otherwise be safe, such as Twitter and Google.
“To begin with, the neutering mechanism (of Microsoft’s XSS filter) can be abused by an attacker to block benign content on a page,” the duo wrote in a research paper. “For example, embedded JavaScript can be blocked from executing by “faking” a XSS attack. This is possible since the string <script will trigger one of the filters, if present in a request and in the response. This string naturally occurs on most webpages, so if an attacker appends a dummy GET parameter like &foo=<script then Internet Explorer 8 will see this on the outbound request and in the response body and thus trigger the neutering mechanism.”
“An attacker may exploit this behavior in order to prevent client-side security functionality from working,” the paper continues.
While Microsoft patched an attack scenario in January that involved modified HTTP responses, the security bulletin did not address the SCRIPT tag attack scenario, explained David Ross of the Microsoft Security Response Center engineering team.
“While the issue identified and addressed in MS10-002 was identified to exist on high-profile web sites, thus far real-world examples of the SCRIPT tag neutering attack scenario have been hard to come by,” he added. “In the case of the Internet Explorer XSS Filter, researchers found scenarios that are generally applicable across XSS filtering technologies in all currently shipping browsers with this technology built-in…Overall we maintain that it’s important to use a browser with an XSS Filter, as the benefits of protection from a large class of attacks outweigh the potential risks from vulnerabilities in most cases.”
Microsoft plans to issue an update to address the situation in June.
You can view Lindsay and Vela Nava’s research paper here (PDF).