Linkedin spurns bug bounty hunter
Will debug for food - but who will bite?
Earlier this month, employees for LinkedIn, a social network site that caters to business people, received an unusual proposition from a security researcher who had just uncovered a vulnerability that put many of its users at serious risk.
"If you are interested in the bug, we would like to give you first right of refusal to purchase it," security researcher Jared DeMott wrote in a July 10 email addressed to a LinkedIn executive. "If you wouldn't like to buy it then we are happy to re-sell or release as a full disclosure to help prevent security issues arising on end users servers."
As part of the offer, DeMott would provide working attack code so LinkedIn could verify the bug, which initially was believed only to allow an attacker to crash a LinkedIn user's machine. The price tag: $5,000. A week later, DeMott increased the asking price to $10,000 after writing exploit code that proved the vulnerability in the LinkedIn toolbar for Internet Explorer allowed attackers to completely hijack a user's PC if it visited a booby-trapped website.
Adhering to a policy of not paying for security vulnerabilities, LinkedIn executives didn't respond to the emails. On July 23, almost two weeks after first contacting LinkedIn, DeMott made good on his pledge, posting proof-of-concept code on his website that demonstrated the severity of the flaw he had discovered.
LinkedIn patched the critical security hole about 24 hours later.
The incident underscores the ethical minefield confronting both companies and researchers when serious vulnerabilities are discovered in a widely used program or service. There are no reports that the vulnerability was maliciously exploited. But LinkedIn's decision not to pay meant its users, many of whom are high-net worth individuals, were put at a higher risk of being victimized than would be the case had LinkedIn bought the exploit and fixed it quietly.
The decision also failed to compensate a researcher whose expertise can command fees as high as $500 per hour for expending considerable effort uncovering information that was crucial to LinkedIn.
"If you put yourself in Jared's shoes ... he's got kids he's got to support, so what other option is there?" says Charlie Miller, principal security analyst for Independent Security Evaluators, who submitted a paper on his experience selling a security vulnerability to a US governmental agency. "LinkedIn is the company that has the most to gain from this."
And yet even Miller, an acquaintance of DeMott, admits to feeling uncomfortable with the offer. "Your gut reflex is that they're basically blackmailing the company," he says.
DeMott says blackmail was the last thing on his mind when he sent the email to LinkedIn. He says he recently sold a vulnerability to a company very similar to LinkedIn for $5,000. The flaw, in an Internet Explorer plugin, also could have allowed the remote exploit of users.
"They were tickled and promised to hire us for consulting," DeMott says. Had I sold [the LinkedIn exploit] to the Russian mob, I could have gotten 100 grand. I don't know - but I could have gotten a lot more than $10,000."
DeMott also points out that the LinkedIn flaw was discovered as a result of pointing a fuzzer at the toolbar code. A spokeswoman for LinkedIn said she was unaware if the toolbar had ever been put to such a test, which is considered common security practice in some companies.
"How much money do they spend on other stuff?" DeMott asks of LinkedIn's engineering executives.
But plenty of security researchers say offers such as the one DeMott extended to LinkedIn have no place in legitimate security circles.
"I certainly don't feel comfortable with a model of pay us or we'll drop the proof-of-concept code," said Terri Forslof, manager of security response for TippingPoint, a division of 3Com that provides network security products. "From a vendor's perspective, there's no other way to look at it than as extortion, especially if the vendor doesn't have a defined model for handling this kind of request."
Forslof helps oversee Tipping Point's Zero Day Initiative, a program that pays bounties to researchers for responsibly disclosing vulnerabilities. Tipping Point uses the information to help keep its subscribers secure and also offers assistance to the affected company in plugging the security hole.
Echoing a familiar criticism of full-disclosure - in which researchers provide enough information for technically savvy individuals to reproduce the exploit - Forslof says going public with a vulnerability benefits no one.
"You've put an indefinite number of people at risk, given the recipe to the black market, and you didn't get paid for it because you just dropped it publicly."
DeMott says he approached representatives of the Vulnerability Contributor Program, a competing bounty program set up by VeriSign's iDefense division, to see if they might be interested in buying the LinkedIn exploit. He said talks got bogged down on issues he didn't want to detail and he never bothered to contact people at the Zero Day Initiative.
That's too bad, says Forslof, because the fees he was seeking "were not out of line". In her mind, the incident was a missed opportunity.
"If you're willing to take the time and the due diligence to try to get LinkedIn to hear your case, why not take the 5 or ten minutes to understand what those [third-party bounty] programs do with the information?" she says. "He would have gotten what he wanted, the vendor would have taken care of the problem, and users would not have been put at risk." ®