Using a self-propagating worm that exploits a scripting vulnerability common to most dynamic websites, a Los Angeles teenager made himself the most popular member of community website MySpace.com earlier this month.
While the attack caused little damage, the technique could be used to destroy website data or steal private information — even from enterprise users behind protected networks, according to an internet security firm.
The unknown 19-year-old, who used the name "Samy," put a small bit of code in his user profile on MySpace, a 32-million member site, most of whom are under age 30. Whenever Samy's profile was viewed, the code was executed in the background, adding Samy to the viewer's list of friends and writing at the bottom of their profile, "... and Samy is my hero."
"This is an attack on the users of the website, using the website itself," says Jeremiah Grossman, chief technical officer at WhiteHat Security.
The worm spread by copying itself into each user's profile. Because of MySpace's popularity — it had 9.5 billion page views in September, making it the fourth most-popular site on the web, according to comScore Media Metrix — the worm spread quickly. On his website, Samy wrote that he released the worm just after midnight on October 4. Thirteen hours later, he had added more than 2,500 "friends" and received another 6,400 automated requests to become friends from other users.
"It didn't take a rocket or computer scientist to figure out that it would be exponential, I just had no idea it would proliferate so quickly," Samy says in an email interview posted on a blog. "When I saw 200 friend requests after the first 8 hours, I was surprised. After 2,000 a few hours later, I was worried. Once it hit 200,000 in another few hours, I wasn't sure what to do but to enjoy whatever freedom I had left, so I went to Chipotle and ordered myself a burrito. I went home and it had hit one million."
Samy also received hundreds of messages from angry MySpace users. He wasn't contacted by officials from Los Angeles-based MySpace, though his account was deleted. MySpace was purchased in July by Rupert Murdoch's News Corporation for US$580 million. MySpace didn't return requests to comment.
The attack depended on a long-known but little-protected vulnerability called cross-site scripting (XSS). XSS arises because many websites — apart from static sites that use only simple HTML code — are dynamic, allowing users to manipulate website source code.
Websites and browsers such as Internet Explorer and Firefox try to block such XSS holes, says Grossman. But the vulnerabilities continue to exist, for which he blames both the browser creators and the site operators.
Standard enterprise network security tools such as firewalls, antivirus and Secure Sockets Layer don't thwart XSS and other web application attacks because the affected user is already behind his firewall, says Grossman, whose 14-person firm consults businesses on how to prevent such attacks.
"The network is pretty locked down. But all of the new attacks are targeting where nobody is looking — the web application layer," he says.
Other web application-layer break-ins include a case earlier this year where more than a hundred applicants to Harvard Business School got an early peek into their admission files by simply modifying the URL typed into their browser address box. In a more serious phishing attack last year, someone injected code into SunTrust Bank's Web site designed to send emails from SunTrust's website asking account holders for account details.
An early version of an XSS-related vulnerability was discovered in Hotmail in 2001. That flaw allowed an attacker to send an email with malformed HTML code to a Hotmail user, whose browser would interpret the broken commands as legitimate script that would tell the site to steal the user's private information.
Grossman says most such cases go unreported.
While both Firefox and Internet Explorer promise security enhancements in upcoming versions, Grossman said he doubts they will entirely fix the XSS problems.