A newly disclosed vulnerability in Skype for Android could be exploited by miscreants to bypass an Android phone's passcode screen to view photos, contacts, and even launch browser windows.

Bug-hunter Florian Kunushevci today told The Register the security flaw, which has been reported to Microsoft, allows the person in possession of someone's phone to receive a Skype call, answer it without unlocking the handset, and then view photos, look up contacts, send a message, and open the browser by tapping links in a sent message, all without ever unlocking the phone. This is handy for thieves, pranksters, prying partners, and so on. Here's a video demonstrating the bypass...

Youtube Video

Kunushevci, a 19-year-old bug researcher from Kosovo, said he was an everyday user of the Skype for Android app when he noticed that something appeared to be amiss with the way the VoIP app accessed files on the handset. Curious, he decided to put his white hat on, and take a closer look.

"One day I got a feeling while using the app that there should be a need to check a part which seems to give me other options than it should," he explained. "Then I had to change the way of thinking as a regular user into something that I can use for exploitation."

What he eventually found was that, once a Skype call has been received and opened, the application functions as normal, allowing features like photo-sharing and contact look-ups regardless of whether the rest of the phone was unlocked.

Much like the various iOS flaws spotted over the years, the bug is really down to a security oversight. In this case, the Skype app allows users to access the photo and contact features without first checking if the person using the device was authenticated.

"For the specific bug that I have found on Skype, it is more of a bad design and also a bug in coding," Kunushevci told El Reg. "I think to put it all together, humans make mistakes."

Prior to going public, Kunushevci alerted Microsoft to the hole in October and waited for a patch to land. The vulnerability is fixed in the latest versions of Skype, issued December 23, so users can protect themselves by making sure they have the latest build of the app installed.

The vulnerability affects Skype on all versions of Android, according to the bug hunter. We note that the Skype app version differs depending on which version of Android you have installed, though essentially we're told new builds of the application installed or updated after Christmas with a version number over 8.15.0.416 should be safe.

Though still a teenager, Kunushevci says he already has several years of experience in security research. Starting at the age of 12, he became interested in the reasons his own computer was crashing ,and began looking up the various causes of common security and stability flaws. Within a few years he was claiming bug bounties of his own.

"I started working in Bug Hunting when I was 15 years old trying to find web vulnerabilities for Microsoft, Apple, Dell, Intel, Adobe, Eset, Github and other companies, which I used to gain Hall of Fame status and T-Shirts in order to promote my self and learn new things," he said.

"After some years of development I started working on CTFs (Boot2Root) which taught me the most important thing, which is realizing that what you have learned till now is nothing of what should be learned."

A spokesperson for Microsoft was not available for immediate comment. ®

University of Maryland researchers have given Google a "welcome to 2019" gift by breaking its latest reCaptcha audio challenge.

The work is a follow-up to an attack published in April 2017 by the university's Kevin Bock, Daven Patel, George Hughey and Dave Levin, again attacking the audio challenges. Since then, Google has updated the code, and the boffins have updated their attack.

The audio challenge was created to solve reCaptcha's accessibility problem – someone using a screen reader can't see where to "tick the box" to prove they aren't a robot.

The 2017 attack, documented here, downloaded and segmented the audio captcha, sent the segments on to multiple online speech-to-text services, checked the responses for homophones, applied a weighted vote to those responses, and uploaded the answer to reCaptcha.

They claimed better than 85 per cent accuracy for that attack, and when Google fixed reCaptcha's audio challenge, the group set to work attacking the replacement. They demonstrated that the fixes made reCaptcha less secure, told Google in June 2018 (with a six-month disclosure deadline), and on Monday published unCaptcha2.

The group said: "Thanks to the changes to the audio challenge, [parsing] ReCaptcha is easier than ever before. The code now only needs to make a single request to a free, publicly available speech to text API to achieve around 90 per cent accuracy over all captchas."

The GitHub post notes that unCaptcha2 no longer needs to use multiple speech-to-text engines, and the fragmentation approach used in the first version has also been abandoned.

The boffins added that Google cleared them to release the code. "The Recaptcha team is aware of this attack vector, and have confirmed they are OK with us releasing this code, despite its current success rate."

They added: "While unCaptcha2 is tuned for Google's Demo site, it can be changed to work for any such site – the logic for defeating ReCaptcha will be the same."

Researchers wanting to check out unCaptcha2 for themselves will need their own API keys from the relevant services (speech-to-text engines from wit.ai, Bing, IBM and Google).

Since Google has had six months' notice, the boffins noted that unCaptcha2 could stop working at any time. ®

ESET eggheads have shed more light on the Unified Extensible Firmware Interface (UEFI) rootkit being used by the Kremlin's Fancy Bear hacking crew.

Dubbed Lojax, the software nasty embeds itself within the motherboard firmware of infected Windows PCs, allowing it to run as soon as the machine is powered up or reset, allowing it to ideally spy on the user and evade detection by the operating system or any antivirus tools. The firmware executes at the lowest levels, underneath OS kernels and apps, with full system access.

While ESET blew the whistle on Lojax with a disclosure back in September, the nitty-gritty of how the malware operates was withheld until the annual Computer Chaos Club conference at the end of December, where researcher Frederic Vachon described in a presentation how the UEFI-based rootkit is able to hide in modern firmware. ESET was able to get its hands on a copy of Lojax when one of its customers' computers picked up the cyber-nasty.

In short, Lojax – a modified version of the legitimate Lojack anti-theft software – starts out as a poisoned application delivered via spear phishing emails that, when run by hoodwinked victims, unpacks and runs code that hijacks a vulnerable driver, which is loaded by the UEFI firmware during startup, to install the rootkit in flash memory.

UEFI has a mechanism called DXE that locates and runs device drivers needed to get a machine going, before the OS is booted. Lojax tampers with one of the loaded drivers so that it flips the switch on the firmware's write protection, disabling the safeguard and allowing the chips' contents to be overwritten. At that point, the malware can inject its rootkit into the firmware so that it is run following a subsequent power or reset cycle.

So, the driver is fiddled with so that it unlocks the firmware during next boot, then Lojax alters the flash chip contents to inject its rootkit, and then on next boot, the spyware is up and running. This is because the firmware is expected to maintain its own security.

"The chipset exposes write protection mechanisms that need to be properly protected by the firmware," he explained. "Because there is no such thing as BIOS write protection by default, it is the job of the firmware to do that."

In order to turn off the firmware's write protection, Lojax exploits a known race-condition in Intel's flash memory controllers by repeatedly enabling BIOS updates and attempting to write to flash memory before the motherboard software steps in and automatically disables updates.

Once that is done, Lojax is able to write the full rootkit on to the UEFI firmware, which ensures the malware is installed and run during the operating system startup, making it nearly impossible to remove by anything short of a complete reflash of the board's SPI memory.

Fortunately, Vachon notes, vendors themselves can protect against the attack by fixing and patching the driver vulnerabilities, and/or enabling Secure Boot. This will cause the Lojax malware to automatically abort installation attempts. Secure Boot cryptographically ensures the firmware contents isn't altered. Also, not opening emailed applications as a system administrator helps a lot: the rootkit installer requires sysadmin access. However, it could exploit privilege escalation holes to gain admin control, so the real solution is enabling Secure Boot in the BIOS settings and set a password or other lock to prevent any changes.

"The tool only works if the platform is misconfigured," he explained. "If firmware vendors would have done their job correctly here, this tool would have failed at flashing the firmware, it's a great example of how [important] firmware security is." ®