We verified the vulnerability in our lab, acquired it from the researcher, then promptly reported the vulnerability to the Mozilla security team shortly after. Successful exploitation of the vulnerability could allow an attacker to execute arbitrary code. Not unlike most browser-based vulnerabilities that we see these days, user interaction is required, such as clicking on a link in e-mail or visiting a malicious web page.
Here's how the flaw was described to me in an e-mail sent by TippingPoint's PR firm, publicizing their find, closely timed with the blog post, and soliciting interviews about their ZDI program:
Successful exploitation of this vulnerability could allow an attacker to execute arbitrary code, permitting the attacker to completely take over the vulnerable process, potentially allowing the machine running the process to be completely controlled by the attacker. TippingPoint researchers continue to see these types of "user-interaction required" browser-based vulnerabilities -- such as clicking on a link in email or inadvertently visiting a malicious web page.
I'm not about to disconnect from the Internet, or uninstall my copy of Firefox 3.0, just yet. First, in order to exploit the flaw, users would have to be enticed to click on a link, or end up at a bad guy's Web site. So the flaw isn't "worm-able," and would primarily only be a problem to those users who individually fell victim. (Granted, there's too many who don't watch their URLs, but the flaw requiring user interaction mitigates widespread risk.) Bottom line is that there's no Internet-virus pandemic on the horizon, and no real need to sound the alarm.
Second, the timing of the announcement is highly, very highly, suspect. From the ZDI blog:
What we can confirm is that about five hours after the official release of Firefox 3.0 on June 17th, our Zero Day Initiative program received a critical vulnerability affecting Firefox 3.0 as well as prior versions of Firefox 2.0.x.
So, within five hours of Firefox 3.0's release someone found and confirmed the flaw? Why wasn't the flaw found in any of the earlier release candidates? The flaw also resided in previous 2.0.x versions of Firefox. Sure, it's possible that no one found the flaw in the earlier versions, and researchers missed this flaw in the release candidates that have been available for months and months.
Common sense begs the more likely scenario:
Chances are that the flaw was discovered at an earlier date, and the researcher sandbagged the release to TippingPoint to coincide with 3.0's big download day. Can't say I blame the researcher for that, could be a nice way to ensure successful sale of the vulnerability. But TippingPoint's announcing the zero-day flaw before the patch is released is nothing more than this vendor's typical publicity grab.
Much better for us all if security vendors acted responsibly and quietly informed the software maker of the flaw, and only publish the fact that these vulnerabilities exist if the software vendor has been dragging their feet.
And there's been no sign of that in this case, just TippingPoint seeking some cheap PR.
Mozilla's update is available here.