Wayne Huang, CTO and a researcher with Armorize Technologies, says he and his team spotted the drive-by download attack on the internationally popular www.goal.com soccer website between April 27 and April 28, which allowed the attackers to control the content on parts of the site. And this was no mass SQL injection attack, he says, but rather a very targeted one. Goals.com is ranked No. 379 on Alexa.
"We saw an attacker injecting code into the website and causing it to serve up malware [during that time frame]," Huang says. "We think that was just a test initially."
The attackers employed a lesser known malware kit called "g01pack" and appear to have installed a backdoor for wresting control of the site. "It went to malicious domains that we have not seen anywhere else yet. None of these domains were flagged by any AV product or by Google SafeBrowsing," he says. The team was not familiar with the malware kit, he says.
What's unclear is just how the attackers infected goals.com with the malicious iFrames, or what the attackers are after. "We don't know at this point ... the backdoor allows them to do a lot of things," Huang says.
Huang and his team also found a fake admin page for the toolkit that acts as a honeynet for researchers -- this lets the attackers study who is looking at their malicious domains. "This allowed the attacker to know someone linked back [to the domains'," he says.
"After we accessed the fake control panel, the injection was deleted from goal.com. We're still actively monitoring it, and it hasn't appeared again," Huang says.
The researchers alerted goals.com of their findings, but have not yet received a response.
Though the malware was injected on the site for a little more than 24 hours, all you had to do was visit the site during the time frame to get infected. "There definitely should be a lot of infections considering that they get around 23,000 visitors a day," Huang says.
What sets the goal.com hack apart is that it's a large website that was hit with a rare exploit pack that has a low detection rate -- only four of 41 AV products detected anything awry with the code, and it had multiple malicious URLs associated it with it, none of which had been blacklisted. "This means when they decide to activate it, the infection rate is going to be pretty high," Huang says.
Among the browser exploits used during the "test run" of the attack were CVE-2010-1423 (Java), CVE-2010-1885 (MS help center HCP), CVE-2009-0927 (PDF), and CVE-2006-0003 (MS MDAC), according to Armorize.
Armorize has posted a blog here with details about the malicious code as well as the infection logs.
Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message.