Legitimate browser functionality can be abused to trick users into believing that a trusted website has asked them to download a file, which is actually being served from a rogue server, Google security engineer Michal Zalewski demonstrated on Tuesday.
Zalewski’s proof-of-concept attack begins with a button on a page that, when clicked, opens the official Flash Player download website in a second tab and switches the browser’s focus to it. After a few seconds, the original page serves a file called flash11_updater.exe from Zalewski’s server, which causes the browser to display a download dialog.
However, because this happens while the active tab is the one with the official Flash Player website loaded into it and an adobe.com URL in the address bar, it appears as if the download was initiated by Adobe’s website.
“In a way this is a social engineer’s holy grail,” said Emmanuel Carabott, security research manager at security vendor GFI Software, via email. “What a social engineer is trying to do is getting you to trust what they are saying. The more authentic they can make it seem the more successful the attack will be.”
There have been many social engineering attacks in the past that tricked users into downloading malicious files by passing them as Flash Player updates. A lot of these attacks used spoofed pages that mimicked Adobe’s official Flash Player site.
Zalewski’s method removes the need for spoofed pages, therefore potentially increasing the success rate of such attacks by making the trick more credible.
Users who are presented with a Flash Player download request from the genuine Flash Player website will accept the download with little or no hesitation, said Bogdan Botezatu, a senior e-threat analyst at antivirus vendor BitDefender, via email. “The more legitimate an attack looks, the higher the rate of success.”
“All the top three browsers [Internet Explorer, Mozilla Firefox and Google Chrome] are currently vulnerable to this attack,” Zalewski said in a blog post.
Internet Explorer and Firefox do specify the hostname from where the file will be downloaded inside the download dialog. However, Zalewski feels that these indicators are inadequate and that as long as the download prompt is attached to the wrong browser window, many users will fail to notice them.
According to the researcher, the browser vendors have been notified about this issue since the beginning of April, but none of them is rushing to address it. Microsoft does not plan to fix it via a security patch for any of the current versions of Internet Explorer, Mozilla didn’t commit to a fix either and, while the Chrome developers agreed that it needs to be addressed, they haven’t set a specific date for it yet, Zalewski said.
“I think these responses are fine, given the sorry state of browser UI security in general; although in good conscience, I can’t dismiss the problem as completely insignificant,” Zalewski said.
Carabott and Botezatu both agreed that this issue should be fixed because attackers are likely to start exploiting it if it is left unaddressed. However, the fact that this attack method leverages legitimate browser functionality might make it harder to mitigate, Carabott said.
“We’re aware of a spoofing issue that can be mitigated by enabling Internet Explorer’s Smart Screen Filter,” said Yunsun Wee, director of Trustworthy Computing at Microsoft. “There have not been any known attempts to exploit this issue, and we continue to encourage customers to only visit trusted websites.”
Mozilla did not immediately return a request for comment.
“The best practice is to refuse any download that initiates automatically,” Botezatu said. “If this is not possible, users should scan the downloaded file with either an antivirus or online with a multi-engine antivirus service.”