A pair of researchers from the Indiana University uncovered a vulnerability in Facebook that allowed attackers to get their hands on user data.
Students Rui Wang and Zhou Li found a flaw in the Facebook platform code that enables a malicious site to impersonate other Websites and obtain the same access permissions those sites receive.
"Bing.com by default has the permission to access any Facebook users' basic information such as name, gender, etc., so our malicious website is able to deanonymize the users by impersonating Bing.com," Wang told eWEEK in an e-mail. "In addition, due to business needs, there are many websites requesting more permissions, including access to a user's private data, and publishing content on Facebook on her behalf. Therefore, by impersonating those websites (e.g., NYTimes, ESPN, YouTube, and FarmVille, etc.), our website can obtain the same permissions to steal the private data or post bogus messages on Facebook on the user's behalf."
Facebook patched the flaw shortly after it was reported to it, and said it is not aware of the issue having been exploited.
"Security is a top priority for us, and we devote significant resources to protecting people's accounts and information," a company spokesperson said. "We maintain a strong relationship with security experts around the world and work closely with them in the rare instances in which they find vulnerabilities on Facebook."
A YouTube video of the exploit in action can be viewed here.
"When I first experimented last week on a test site created for me by Zhou and Rui I couldn't precisely mimic what you see in the video,"blogged Graham Cluley, senior technology consultant at Sophos. "The demo website wasn't able to extract the name of my test Facebook account, and it displayed a 'failed' dialog box when it tried to post to my Facebook wall.
"Now it's possible that it didn't work because I had applied some pretty rigid privacy settings to my test account, and sure enough when I tried again (having installed the ESPN Facebook app onto my test account) it was then successful, and able to extract my name, email address, and post an 'evil' link seemingly via the app," he wrote.