Anti-hacking laws 'can hobble net security'
Good Samaritans discouraged by threat of prosecution
Jeremiah Grossman has long stopped looking for vulnerabilities in specific websites, and even if he suspects a site to have a critical flaw that could be compromised by an attacker, he's decided to keep quiet.
The silence weighs heavily on the web security researcher. While ideally he would like to find flaws, and help companies eliminate them, the act of discovering a vulnerability in any site on the internet almost always entails gaining unauthorised access to someone else's server - a crime that prosecutors have been all too willing to pursue.
"I have long since curtailed my research," said Grossman, who serves as the chief technology officer for website security firm WhiteHat Security. "Any web security researcher that has been around long enough will notice vulnerabilities without doing anything. When that happens, I don't tell anyone, rather than risk reputational damage to myself and my company."
Grossman's fears underscore the fact that security researchers who find flaws in websites are crossing a line and trespassing on systems that do not belong to them. However, applying the law to good Samaritans interested in eliminating possible online risks only undermines the security of the Internet, a working group of researchers, digital-rights advocates and federal law enforcement officials concluded this week.
"I think that if you look at the software security world, there has been many, many cases of someone knowing about a vulnerability before you do and be using it out in the wild," said Sara Peters, editor for the Computer Security Institute. "There is no way to say that these same things are not happening in the web world. Assuming that nothing is going wrong, because you haven't heard about it is a very myopic and callow way of looking at it."
Dubbed the Working Group on Web Security Research Law, the panel of experts has started to study whether researchers have any ability to play the good Samaritan and find security flaws in websites without risking prosecution. The group met at the Computer Security Institute's NetSec on Monday and released an initial report that raises more questions about the status of web vulnerability research than provides answers to concerned bug hunters.
While security researchers have been able to test computer software and disclose details about any flaws found, the working group concluded that there is no way to test a web server without prior authorisation and not run the risk of being prosecuted. Software security researchers are free to disclose flaws fully or take part in a process that allows the vendor to plug the holes, while web researchers that disclose vulnerabilities in a way that angers the website owner could easily be reported to law enforcement.
"The way it is right now, if you find a vulnerability and the site owner finds about it, you can be held culpable for anything that happens after that," Peters said. "Perhaps, that is a bit of hyperbole, but not much. There is no culpability for the website owner."
The working group's report, available from the Computer Security Institute (registration required), includes four case studies including that of Eric McCarty.
In June 2005, McCarty, a prospective student at the University of Southern California, found a flaw in the school's online application system and notified SecurityFocus of the issue.
SecurityFocus contacted the school at the request of McCarty and relayed the information to USC, which initially denied the seriousness of the issue but eventually acknowledged the vulnerability after McCarty produced four records that he had copied from the database. In April 2006, federal prosecutors leveled a single charge of computer intrusion against McCarty, who accepted the charge last September.
As part of its policy, SecurityFocus did not publish an article on the issue until USC had secured its database.
While CSI's Peters believes that good Samaritans should be given some leeway, a few of the comments found on McCarty's computer by the FBI - and repeated in court documents - suggested that vengeance was a motive. For that reason, Peters suggests that security researchers who decide to look for vulnerabilities in websites use discretion in dealing with site owners.
"You can't let anyone run wild and hack into websites indiscriminately," Peters said. "If you publicly disclose a vulnerability in a website you are pointing a big red arrow at a single site, so there needs to be some discretion."
"AJAX is not necessarily adding more vulnerabilities to the landscape, it is making it more difficult for the scanner vendors to find the vulnerabilities," said WhiteHat Security's Grossman, who is also a member of the working group. "The sites still have vulnerabilities, but they are harder to find."
Independent researchers finding vulnerabilities in websites could put pressure on site owners to secure their part of the internet. However, the working group could not agree on whether the law should be changed to allow for good Samaritans.
That likely leaves liability as the best stick, said Grossman, who website owners should be held liable to some extent for any consumer data lost due to a vulnerability in their site.
"I think the motivation has to monetary," he said. "Right now, the website owners are the ones that have to pay for the security, but the consumer is the one bearing all the costs of failure."
Such an equation, he said, is unlikely to add up to better security.
This article originally appeared in Security Focus.
Copyright © 2007, SecurityFocus