Mozilla confirms own URL handling bug
'Turns out it's a problem with Firefox as well'
The Mozilla Foundation acknowledged over the weekend that its own Firefox browser allows links that can send malicious code to external programs, a security issue that the group had previously argued should be fixed by the browser maker.
In early July, three researchers found a way to execute code in Firefox - and potentially other Windows programs - by passing it a malicious uniform resource identifier (URI) from Internet Explorer.
The discovery lit off a firestorm of finger pointing: The Mozilla Foundation argued that IE should validate the URI before passing it along to another program, while Microsoft stated that input validation is the responsibility of the receiving program.
Over the weekend, another researcher discovered that Mozilla Firefox has the same security issue. The Mozilla Foundation acknowledged the problem on Monday.
"We thought this was just a problem with IE," Mozilla's chief security officer Window Snyder said in a blog post. "It turns out, it is a problem with Firefox as well."
In the latest versions of their products, Microsoft and the Mozilla Foundation have focused on security. In Internet Explorer 7, Microsoft added anti-phishing features, the ability to run in protected mode on its latest operating system, Windows Vista, and severely culled problematic ActiveX controls. In Firefox 2.0, the Mozilla Foundation also added anti-phishing features and the ability to clear private data.
Mozilla is now looking into the issue to determine its response to the problem.
This article originally appeared in Security Focus.
Copyright © 2007, SecurityFocus
"the only way IE could validate parameters passed to a 3rd part app is if the app provided an interface for it to validate against. it couldn't just validate everything against an IE defined set of rules. what might be invalid for your app could be valid for mine."
But IE shouldn't validate the parameters passed to the 3rd part app. The 3rd part app should validate the parameters.
As Dillon Pyron pointed out: "You should NEVER EVER trust another program to protect you from the bad guys."
Your suggestion only covers 'good programming'. If I wrote a program taking advantage of the exploit, and wanted to pass data to IE (or another program) in an attempt to do something on your system that I really shouldn't, do you rely on my code to follow the rules, and validate the data before passing them?
the only way IE could validate parameters passed to a 3rd part app is if the app provided an interface for it to validate against. it couldn't just validate everything against an IE defined set of rules. what might be invalid for your app could be valid for mine
I love Noscript. I'm stuck with stable at work, so that's my cure. I haven't installed 18.104.22.168 RC at home because, well, because I'm too lazy. But I've got Noscript. I also resist using IE unless I absolutely have to.
I agree, IE should validate before sending to FF and FF should validate before sending to IE. BUT, FF should validate after receiving from IE and IE should validate after receiving from FF. You should NEVER EVER trust another program to protect you from the bad guys. To do so is either crap design, crap coding or pure laziness. Or more likely, a combination.