Attackers improve on JavaScript trickery

CanSecWest As JavaScript becomes an increasingly key component of online attacks, attackers are investing more energy in obfuscation and other techniques to make defenders' attempts at reverse engineering more difficult, a security researcher told attendees at the annual CanSecWest conference on Wednesday.

Attackers have adopted the same techniques used to hide the purpose of other types of malicious code, such as splitting up the code into many components and the use of custom encoders, to obfuscate JavaScript, said Jose Nazario, senior security engineer at network-protection firm Arbor Networks. Other advances include the addition of functions aimed at detecting any attempts at debugging or running the program in a virtual machine, he said.

"There is a lot of defensive JavaScript coming around," Nazario told attendees. "Attackers now will kill alerts and kill all sorts of inspection routines. They also will frequently only let a single IP (Internet protocol) address download the JavaScript."

A year ago, researchers warned about future worms that could spread through users' online profiles and data using JavaScript and interactive Web technology, similar to the Samy worm that infected MySpace in 2005. The increasing use of asynchronous JavaScript and XML (AJAX) for sharing data and adding interactivity to wesites has compounded the danger.

In the last year, malicious software written in JavaScript and AJAX has moved from an interesting research topic to a significant threat on the Internet, increasingly used by attackers to compromise users' computers. In February, researchers at security firm Websense discovered that the website for the Dolphin Stadium had been compromised in an attack that did not deface the site, but rather had infected the home page with malicious JavaScript code that attempted to force visitors to download a Trojan horse from one of three sites in China. Further research by incident responders found at least three dozen other sites that hosted similar code.

Nearly identical attacks, likely perpetrated by the same group, used the recent Microsoft animated-cursor flaw to compromise computers as well.

"I would not say that this is the end of their attacks," said Dan Hubbard, vice president of security research for Websense, said in February.

The advances in sophistication of malicious JavaScript make it more important for researchers to use proper debuggers and keep copies of any obfuscated code, Arbor's Nazario said. Many of the latest techniques are not aimed at fooling the victim, but the malware analyst, he added.

"The object is to make the attack vector that much more opaque, not just to your system, but to the analyst as well," Nazario said.

Other security experts have agreed that malicious software written in JavaScript will become more powerful.

Last month, security researcher Billy Hoffman showed off a JavaScript vulnerability scanner that could turn the computer of any visitor to a malicious Web site into an unwitting accomplice in an attack. While the proof-of-concept program, known as Jikto, had only rudimentary functionality, further development could create software that essentially turns Web site visitors into temporary zombies, said Hoffman, lead researcher for Web security firm SPI Dynamics.

"This is only going to make things worse," Hoffman said at the time. "It is like you (the victim) are in a bot net but without all the traditional malware traces that bot software usually leaves behind."

In the week following the presentation, someone leaked the source code - which Hoffman had intended to keep private - to the internet. This article originally appeared in Security Focus.

Copyright © 2007, SecurityFocus