Major Companies 'Fail' Social Engineering Test
Report details results of major social-engineering 'capture the flag' contest that targeted 14 companies in retail, airlines, food service, technology, and mobile services
October 31, 2011
Turns out employees in a retail setting are less likely to get duped by a social engineer than people who work in a call center or customer support site.
That's just one of the main takeaways from the postmortem report published today about the second annual live Social Engineering Capture The Flag contest that was held in August at DefCon. The contest targeted multiple companies in five different industries -- retail, airlines, food service, technology, and mobile services -- to determine how susceptible they are to a social engineer schmoozing potentially sensitive information out them.
This year's contestants tried to squeeze specific information, or "flags," out of Apple, AT&T, Conagra Foods, Dell, Delta Airlines, IBM, McDonalds, Oracle, Symantec, Sysco Foods, Target, United Airlines, Verizon, and Walmart during the two-day contest.
Prior to the live contest, contestants had two weeks to gather intel on the companies via passive information-gathering methods, such as Google searches, social networks, and Web research. That information was compiled in a dossier they turned in prior to the conference and was part of their overall scores.
It wasn't until DefCon that contestants got to make direct contact with their target companies; they dialed up the targets from a soundproof booth at DefCon, with an audience and organizers Hadnagy and James O’Gorman watching and listening. They had just a 25-minute time slot to capture as many flags as possible. There were more than 60 flags, and they ranged from the names of the food service providers in the company cafeterias to their antivirus programs and browser versions. All of the companies (unknowingly) surrendered flags, and only three employees resisted giving up information to the caller.
AT&T fared better than Verizon in the contest, but not because AT&T was necessarily more secure. "When they called AT&T, it was the retail store, not corporate. In that environment, those people aren't going to take any crap. They are really very hard to get through," says Chris Hadnagy, developer and community member of Social-Engineer.org, which created the contest. Another contestant targeting Verizon spoke with an employee in the customer service call center, and was able to more easily charm information out of that person.
Delta Airlines and United Airlines also scored low in the contest, which Hadnagy says was somewhat surprising since they regularly deal with consumers. "But the argument is that they were call centers, and security training is generally not high in call centers because there's such high turnover, so they don't want to spend the time and effort," Hadnagy says.
One of the lowest-scoring companies in the contest -- meaning that the contestant was able to score the most "flags" from them -- was Oracle. But that doesn't mean Oracle is the least secure company overall, Hadnagy notes: "There are so many factors. This was one caller calling one person in one Oracle location," he says. "It could be that the combination of the caller's voice, his pretext, and the person who got on the phone were enough to make that call so successful that it scored the lowest. Or another caller on another day calling another Oracle employee might not have that success."
All of the targeted companies' employees fell for one of the more potentially dangerous flags: visiting a URL offered by the caller under some pretext. Hadnagy says this flag was captured after the caller had struck a rapport with the employee. Several posed as fellow employees asking for help. "And even those [targeted] employees who knew going to a link was against security policy still tried to help," he says. One woman told the caller she wasn't allowed to go to the link, but still tried it out for him and was subsequently blocked.
Those companies whose employees were the toughest to crack were those who demonstrated what Hadnagy described as "critical thinking." One retail employee contacted during the contest questioned one of the contestants posing as a fellow employee, and eventually hung up on him. "She said [to him], 'There's no reason you should be asking me these questions. You should be asking the manager in your store.' That was a great example of critical thinking. He couldn't get anywhere with her," Hadnagy says. "That needs to be company-wide."
So the contestant called back the same number, and a different retail employee answered the phone. And this time, he was able to social-engineer information out of him.
For next year's contest at DefCon, the organizers hope to target other industries, and are looking for companies that will come forward as volunteer targets. They also hope to field more women contestants: There were only two this year, both of whom dropped out just prior to the live event at DefCon.
The full report is available for download (PDF).
Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message.
About the Author
You May Also Like
Cybersecurity Day: How to Automate Security Analytics with AI and ML
Dec 17, 2024The Dirt on ROT Data
Dec 18, 2024