A design error in the Firefox browser can allow phishers to conceal the true origins of a web page from the user. This could be used to place extremely deceptively genuine looking web pages from organisations such as banks, eBay, PayPal and other providers on the web (spoofing). Browser security specialist Michal Zalewski has provided a demonstration web page to enable interested users to understand the problem. The demo works with Firefox 1.5 and 2.0.
According to Zalewski, the problem lies in the way Firefox deals with the URL about:blank, which opens a blank page. The browser does not show either a URL in the address bar or information in the window's title bar. However, JavaScript can also open such a web page, and various JavaScript functions can be used to insert additional content into the web page. This is not normally possible for windows originating from different domains, but because about:blank is not assigned to any domain and document.location is not defined, it nevertheless works anyway. And, according to Zalewski, older spoofing bugs can also again be exploited in this way in Firefox.
The only remedy at present is to disable JavaScript or to use the NoScript plugin for Firefox, which only allows scripting on known, trusted websites.
Last November, Robert Chapin found a new security hole in Firefox, dubbing it ’Reverse Cross-Site Request (RCSR). It exposed saved passwords and could affect anyone visiting a weblog or forum website that allows user-contributed HTML codes to be added, said Chapin.
Then in January this year, the risks meant Firefox Password Manager should immediately be disabled, he warned.
Now Chapin, who runs Chapin Information Services, tells p2pnet an analysis he’s just finished concludes the February 23 release of Firefox version 2.0.0.2 handles less than 25% of the problems related to Bug #360493, as it’s been labelled, although the fixes will, "satisfy the current security needs of the average Internet user".
Chapin’s charts below are divided into lists of risk factors, "that may contribute to the overall success or failure of Password Management," he says, going on:
"Each risk is assigned a weight from zero to five such that zero means no risk and five means relatively high risk.
"A discussion of each risk follows the analysis tables. Where this document mentions risks that are not yet disclosed to the public, only status information will be given."