A security researcher has discovered a sneaky social engineering trick that might be used to disguise the go-ahead to run hostile code on Windows 8 machines.
The so-called keyjacking technique, uncovered by Italian security researcher Rosario Valotta, is similar to clickjacking. However, instead of fooling marks into generating fake Facebook likes, the keyjacking involves disguising a "run executable" dialogue box within a CAPTCHA challenge.
Miscreants cover up the dialogue box with a window that looks like a CAPTCHA, with R as the first character prospective marks are invited to type. This R input authorises the computer to Run a downloadable file on a potential attack page.
Valotta has created a proof-of-concept demo that shows how a supposed sign-up to a movie-streaming site can be loaded with a fake CAPTCHA challenge that executes potentially hostile code, providing users are tricked into pressing "R".
"The attack technique allows for remote code execution on Internet Explorer and Google Chrome with a minimum user interaction. I'm actually talking of typing one key [on IE] or making one click [on Chrome]," Valotta told El Reg.
"It basically means that visiting a malicious website and pressing one key or one click is enough to download and install malicious executables without any notification for the user."
Valotta released a paper on his research last week after presenting his findings at security industry events Hack in the Box 2013, PhPdays 2013 and Nuit Du Hack 2013.
The attack works on IE9 and IE10 (Windows 7) and on Chrome for Windows 8, according to Valotta, who added that the approach doesn't work on IE8 because the browser features pop-up warnings.
The basic ruse behind the attack is not new and, in earlier incarnations, was clarified as a variant of clickjacking. Valotta's research, however, shows that the approach works on Windows 8 machines and not just in older executable warnings on Win 7 and earlier versions of Windows.
There are a couple of limitations to the technique, even on Win 8 machines with improved clickjacking defences. First, the malign application needs to make it past Microsoft's Smartscreen Reputation check. This cloud-based security technology is not 100 per cent reliable, so this might not be enough to thwart all attacks.
Even after clearing that hurdle, the prospective VXer would need to defeat Microsoft's User Access Control, which enforces a warning whenever an application requires administrative privileges. This warning can't be sidestepped simply using keyjacking but, as Valotta points out, it might be possible to do all manner of mischief with malicious code - even without administrative privileges. ®