The bug, which could have allowed attackers to take over user accounts, was patched in 48 hours
Microsoft has taken only two days to fix a security bug that could have enabled attackers to take over the accounts of users logging into Internet-based services such as Outlook.com or Windows Live.
British programmer Jack Whitton reported the problem to Microsoft earlier this year, on Sunday, 24 January, and a fix was implemented on the following Tuesday. Whitton first made the bug public in a blog post over the weekend.
When a user already logged into a Microsoft service running on one domain, such as outlook.com, accesses another service running on another domain, such as live.com, the user is issued a token proving their identity, so that they don’t have to login a second time. T
he token is used because the services run on completely separate domains, meaning cookies wouldn’t function.
The security bug affected the way the tokens are authenticated, and meant a malicious user who captured a user’s token via a fake website could have used the token to gain “complete access to the user’s account”, in an attack type known as a cross-site request forgery (CSRF), Whitton wrote.
“Despite CSRF bugs not having the same credibility as other bugs, when discovered in authentication systems their impact can be pretty large,” he wrote.
An attacker would need different tokens to access services hosted on different domains, but Whitton speculated that multiple hidden iframes could be used to harvest the tokens of various services.
While Whitton’s bug affected Microsoft’s online services, Windows administrators are currently awaiting a fix for a mysterious bug called “Badlock” that also affects an open-source tool called Samba.
Microsoft and Samba engineers warned that the bug is serious and to expect a fix on 12 April, Microsoft’s next scheduled patching date, but didn’t give any further information in order to prevent attackers from developing an exploit before the patch becomes available.
Are you a security pro? Try our quiz!