In the attack, JavaScript is used to display a pop-up window in front of a trusted Web site. The pop-up appears to be part of the legitimate site, but actually is linked to a different, malicious site. A user might be fooled into sending personal information to the scammers.
Although the pop-ups could be used by attackers, overlaying multiple windows in a Web browser is a feature, not a vulnerability, according to an advisory posted Tuesday on Microsoft's TechNet Web site.
"This is an example of how current standard Web browser functionality could be used in phishing attempts," Microsoft said in the advisory.
Phishing is a prevalent type of online fraud that attempts to steal sensitive information such as usernames, passwords and credit card numbers. The schemes typically combine spam e-mail and fraudulent Web pages that look like legitimate sites.
Earlier this week, security monitoring company Secunia warned of the browser problem and rated it "less critical." The issue affects most major browsers, Secunia said.
The problem is that JavaScript dialog boxes do not display or include their origin. For an attack to occur, a user would have to visit a malicious Web site or click on a link before going to a trusted site, such as that of a bank. The attacker could then overlay part of the trusted site with a window asking for data such as a user name and password. Information entered would go to the attacker, instead of the bank.
Firefox developers at the Mozilla Foundation have been making moves to combat this kind of attack. In April, a patch was developed that allows people to block Java and Flash-based pop-ups unless they came from trusted sites.
Opera has said that its latest browser, 8.01, would display the pop-up's origin, letting a user inspect its URL to see if it came from a trusted site.
Graeme Wearden of ZDNet UK contributed to this report.
