One day after the Great Twitter Meltdown, details of how pranksters gleefully redirected users and spewed out retweets full of gibberish continue to emerge. What’s surprising is that instead of criminals and attackers intent on taking over computers or stealing information, most of the perpetrators were just curious users trying to see what they could do with the security hole.
According to various reports, the programming XSS flaw on Twitter’s home page was initially discovered by Masato Kinugawa, a Japanese developer, who used the code to create rainbow-colored tweets. Another was a Norwegian Ruby programmer named Magnus Holm, who has the dubious distinction of having The New York Times claim he created the first worm to use the XSS exploit. Holm reportedly saw Kinugawa’s code and played with the idea. Holm extended the code to retweet itself by tweeting, “meh, this worm doesn’t really scale. the users can just delete the tweet,” and including a link with an “onmouseover” command.
The XSS bug caused the tweet to be retweeted the instant someone’s mouse cursor touched the link.
Holm said he just wanted to “experiment” with the flaw and was surprised his worm spread to at least 200,000 users.
Another curious user was Pearce Delphin, a 17-year-old boy from Australia, who modified Kinugawa’s code and created his own tweet “uh oh” with the onMouseOver JavaScript command embedded.
“I did it merely to see if it could be done … that JavaScript really could be executed within a tweet,” Delphin told AFP.
Others quickly picked up and modified the code to create their own versions, performing auto-retweets and redirecting to pornographic sites. Some were funny. Most were not.
Despite the havoc and huge amounts of spam, most of the created worms turned out to be harmless pranks intended to sow confusion. Kaspersky Lab estimated up to 500,000 people may have been affected.
“We are not aware of any issues related to it that would cause harm to computers or their accounts,” Twitter said.
Visitors on the Twitter page belonging to Sarah Brown, the wife of the U.K. prime minister, were redirected to a Japanese porn site. Another worm sent gibberish spam messages to White House Press Secretary Robert Gibb’s 100,000 followers. One worm even “rickrolled” users, a widespread Web prank in which unwitting users are instantly sent to a Website playing 1980s singer Rick Astley’s “Never Gonna Give You Up” music video.
Despite the company’s claims that no harm was done, The New York Times claimed one instance of a worm downloading malicious code from a Russian server onto the user’s computer.
Kinugawa discovered the bug and notified Twitter on Aug. 14. Twitter said in the blog post that developers had fixed the security hole last month, but a site update, unrelated to the new site redesign unveiled earlier this month, reintroduced the bug on the old design. Users who already switched to the new design or are using third-party apps to read their tweets were unaffected. Users with JavaScript blockers such as the NoScript add-on for Mozilla Firefox browsers were also spared.
Twitter acted quickly to close the hole. Reports indicate the company first became aware of the problem at 5:24 a.m. EDT and had the flaw patched by 10:23 a.m. EDT. According to the company, there was no need to change passwords because user account information was not compromised through this exploit.
The XSS bug that brought Twitter down to its knees is a fairly elementary mistake. In fact, cross-site scripting flaws exist in seven out of 10 of all Websites, according to WhiteHat Security. Considered relatively benign by Website operators and developers, what happened to Twitter indicates that it can be a real problem. The bug was easy to exploit and spread amazingly fast, and it was done by a handful of hobbyists, not malicious hackers.