Facebook has fixed a flaw that let hackers delete Facebook friends without permission.
The flaw was reported Wednesday by Steven Abbagnaro, a student at Marist College in Poughkeepsie, New York. It was patched Friday afternoon, Pacific time, after the IDG News Service notified Facebook of the issue.
The bug was a variation of an earlier vulnerability that Facebook learned about last week, which affected a range of features on the Web site. Hackers could have leveraged Abbagnaro’s bug to delete all of a victim’s contacts, one by one, but it does not appear that anyone ever exploited it in a malicious way.
For Abbagnaro’s attack to work, however, a user would have to have been tricked into clicking on a malicious Web link while still logged into Facebook.
Facebook has struggled this week to fix these bugs, which are called cross-site request forgery flaws. They exist because of relatively simple Web programming mistakes in the Web site’s code, and security researchers have criticized Facebook for not fixing them more quickly.
“We’re in the process of doing a full audit and are building additional protections for this type of potential attack across the code base,” said Simon Axten, a Facebook spokesman, in a Friday e-mail interview. “We began working on this one as soon as we learned about it and pushed a fix early this afternoon.”
Hackers Can Delete Facebook Friends, Thanks to Flaw
A bug in Facebook’s Web site lets hackers delete Facebook friends without permission.
The flaw was reported Wednesday by Steven Abbagnaro, a student at Marist College in Poughkeepsie, New York. But as of Friday morning, Pacific time, it had still not been patched, based on tests conducted by the IDG News Service on a reporter’s Facebook friends list.
A malicious hacker could combine an exploit for this bug with spam or even a self-copying worm code to wreak havoc on the social network, Abbagnaro said in an interview.
He’s written proof-of-concept code that scrapes publicly available data from users’ Facebook pages and then, one by one, deletes all of their friends. For the attack to work, however, the victim would first have to be tricked into clicking on a malicious link while logged into Facebook. “The next thing you know, you have no friends,” Abbagnaro said.
The security researcher is not going to release the code used in his attack until after Facebook fixes the flaw, but he says that technically competent hackers could figure out how to pull off the attack.
That’s because Abbagnaro’s code exploits the same underlying flaw that was first reported by M.J. Keith, a senior security analyst with Alert Logic.
Last week, Keith discovered that Facebook’s Web site was not properly checking code sent by users’ browsers to ensure that they were authorized to make changes on the site.
Called a cross-site request forgery bug, the flaw is a common Web programming error, but Facebook has had a hard time eradicating it from the site. After Keith first reported the issue, Facebook thought it had fixed the problem, only to discover that it could still be exploited to make users “like” Facebook pages without their consent.
Similarly, Facebook appears to have missed Abbagnaro’s delete-friend vector as well.
“I am just blown away that this keeps happening,” Keith said in an e-mail interview.
Facebook representatives couldn’t immediately be reached for comment.
Facebook’s security team has been under siege lately, with worm attacks and site flaws popping up on a regular basis. These security issues come as the social network has been hit with intense criticism for not adequately protecting users’ privacy, and inappropriately sharing user data with advertisers.
Users have been quitting the social network and a campaign proclaiming May 31 as Quit Facebook Day has gained some traction.
Despite all of its other problems, Facebook should have fixed this latest flaw by now, Abbagnaro said. “I’m not sure why they haven’t fixed it yet because it is pretty serious.”