Scammers have found a way around new token-based authentication systems that have been adopted by some banks.

Over the past few weeks, approximately 35 phishing websites have been set up that use the new attack. They attempt to trick users into divulging the temporary passwords created by the security token devices used by banks such as Citigroup, said Rich Miller, an analyst with Internet research company Netcraft.

Phishers have only recently begun looking for ways around token authentication, using what is known as a "man-in-the-middle" attack, Miller said. "These attacks are worrisome because they took advantage, fairly early on, of a system that’s seen as enhancing security for banking customers," he said.

Token devices are used to create a temporary second password for online banking customers. These passwords are valid for a very short period of time and can be used only once, making it impossible for attackers to steal them for later use. U.S. banks have been offering the tokens to users in an effort to comply with federal guidelines that call for stronger, two-factor authentication for online transactions by year’s end.

Security experts had predicted that phishers would eventually use a man-in-the-middle attack to circumvent token-based authentication, but these recent attacks mark the first time they have actually done so, Miller said.

Under an ongoing attack against Citibank customers, phishers have set up a fake website where victims are tricked into entering their passwords. The fake site instantly forwards the password information to Citibank’s real website, allowing the criminals to sign on before the victim.

With a total of 35 such phishing sites now spotted, it seems that the attack is becoming widespread, Miller said. "This is getting organized," he said. "It is not just an isolated incident of somebody coming up with a proof of concept or an exploit that’s unique to them."

Many of the 35 phishing sites found by Netcraft have now been shut down, although some are still operative, Miller said.

Although these new phishing techniques show that no technique is impervious to attack, token-based two-factor authentication remains a useful tool against malicious software such as Trojan horse programs, said Johannes Ullrich, chief research officer at the SANS Institute.

Ullrich also noted that these attacks rely on victims who will enter sensitive information into an untrusted website, a type of victim that is becoming harder to find as users clue into the phishing phenomenon.

"The real problem is not the phishing sites; it’s the Trojans and keyloggers," he said, adding that "they’ll have a harder time working around the two-factor authentication."