(pdf) gives the bad guys a solution to the biggest problem facing phishers these days: how to reach new victims. In a traditional phishing attack, the scammers send out millions of phoney e-mail messages disguised to look like they come from legitimate companies, such as banks or online payment companies.

Those messages are often blocked by spam-filtering software, but with in-session phishing, the e-mail message is taken out of the equation, replaced by a pop-up browser window.

Here's how an attack would work: The bad guys would hack a legitimate Web site and plant HTML code that looks like a pop-up security alert window. The pop-up would then ask the victim to enter password and login information, and possibly answer other security questions used by the banks to verify the identity of their customers.

For attackers, the hard part would be convincing victims that this pop-up notice is legitimate. But thanks to a bug found in the JavaScript engines of all the most widely-used browsers, there is a way to make this type of attack seem more believable, said Amit Klein, Trusteer's chief technology officer.

By studying the way browsers use JavaScript, Klein said he has found a way to identify whether or not someone is logged into a Web site, provided they use a certain JavaScript function. Klein wouldn't name the function because it would give criminals a way to launch the attack, but he has notified browser makers and expects the bug will eventually get patched.