For almost two years, Microsoft's Internet Explorer browser has been vulnerable to attacks that steal digital security tokens and other sensitive data, a security researcher said recently.
Researcher Chris Evans said he alerted Microsoft to the information disclosure vulnerability in IE in December 2008. As of October 21, it remained unfixed, making his disclosure a “600-day” vulnerability, he quipped.
The bug resides in the IE mechanism for handling Javascript and runtime errors. In some cases, cross-origin content can be echoed back to attackers, allowing them to retrieve sensitive javascript variables. Once upon a time, this proof of concept exploited the vulnerability to steal a security token Google Reader uses to prevent XSRF, or cross-site request forgery, attacks. It has since been neutered by changes Google made, but when it worked, it forced the user to subscribe to a goat-farming feed without asking for permission.
“There are a varied number of text structures which can be stolen (iteratively if necessary) with this trick,” Evans warned.
Firefox was once vulnerable to similar attacks but maintainers of the open-source browser fixed the flaw in December 2008. That was the same month Microsoft was informed of the vulnerability, but it has been allowed to remain.
A Microsoft spokeswoman on Monday issued the following statement, which she attributed to Jerry Bryant, a spokesman for Microsoft response:
"Microsoft is aware of the public posting of a low severity information disclosure issue in Internet Explorer. A successful attack requires a victim website to be configured in a specific way which is non-standard for most sites. We are not aware of any attacks seeking to exploit this issue and will update customers if that changes." ®
