Sunday, June 01, 2008

May 2008: Web site compromises record month!

Here are the highlights of the notable Web site compromises I have seen in the past month:

May 2 - One Year Later, Italian Job Still Working Overtime

It’s been a year since the infamous Italian Job attack of 2007. And in an apparent observance of its anniversary, a similar attack was seen compromising about 90 varied Italian Web sites, all hosted in Italy by a single hosting provider—the same one that hosted the thousands in last year’s large-scale.

May 7 - A Very Convoluted Chinese Gaming-Info-Stealing Campaign

Web sites numbering approximately 9,000 were compromised via SQL injection with embedded malicious JavaScript redirecting users to two major malicious URLs. Among these Web sites were legitimate medical, educational, government, and entertainment sites from around the world.

A survey of the site locations includes India, UK, Canada, France, and China. This observation suggests the attack as the work of an automated Chinese hacktool programmed to search through Web sites for vulnerabilities, creating the same .HTML file that has been used to launch various exploits.

May 10 - More of The Same: Another Half Million Web Sites Compromised

Meanwhile, a malicious script was injected into half a million Web sites believed to be either using poorly implemented or older exploitable versions of phpBB. This event was involved a ZLOB Trojan among others that changes an affected system’s local DNS and Internet browser settings.

May 19 - Chinese Weekend Compromise

Chinese-language Web sites were targeted in an attack that was meant specifically against China, Taiwan, Singapore, and Hong Kong. Google search results at the time of the attack showed 327,000 pages containing the malicious script tag.

May 19 - More Weekend Compromises Reach Other Shores

Another string of Web site compromises was discovered the following week, involving at least four (4) Web sites of various affiliations and different countries. These were injected with a malicious JavaScript that redirects to two sites. Both eventually lead to their own series of redirections, and finally the download and execution of malware: a backdoor and Trojan, respectively.

May 21 - It’s Not Over: Asian Sites Injected with Nasty Code

Two days later, hundreds of thousands of Web sites were again found compromised and inserted with malicious JavaScript code, some of which are sites from the APAC region. Hackers have apparently conducted another massive SQL injection attack. A Google search for the malicious URL turned up 197,000 results.

May 22 - Malicious Domains Found in Compromised Japanese Sites

The next day, several Web sites in Japan — including a popular music download site and a music company site — have been found injected with malicious code.

These are the hard facts, and these developments tell us that there could indeed be a trend that cyber criminals seem to favor this type of attack over other methods. For what it’s worth, our engineers also think that mass compromises are common (or at least not as uncommon as we think), it’s just that they are either found soon enough, or they remain unnoticed and consequently unreported.

And I'm not even talking about thousand other websites which were defaced in nearly every country of the world even in Belgium (eg. VTM Broadcast site) ... it's definitely not a good sign and trend.

A lot of XSS methods seems to be used as will in those or a lot of other compromises.

XSS has been around for a long time, it has neither become less of an attractive attack method, nor has a fool-proof solution against it has been properly formulated.
XSS vulnerabilities can cause a variety of problems for the casual web surfer. These problems range in severity from mere annoyance to complete credential compromise. Some XSS attacks incorporate disclosure of the user’s session cookies, allowing an attack perpetrator to have complete control over the victim’s session and to (in effect) take over the account & hijack the HTTP session.
XSS attacks may also include redirecting the user to some other page or website, and modifying the content of a HTTP session. Other damaging risks include the exposure of the victim’s files, and subsequently the installation of Trojans and other damaging malware — and to what purpose? One can only guess because once the compromise is successful, the criminal’s next actions are open to unlimited possibility.
An XSS attacker utilizes varying methods to encode the malicious script in order to be less conspicuous to users and administrators alike. There are an unaccounted number of variations for these types of attacks, and XSS attacks can come in the form of embedded JavaScript — one of the more common implementations. But be forewarned — any embedded active content is also a potential source of danger, including: ActiveX (OLE), VBscript, Flash, and more.
XSS issues can and do exist as well in the underlying Web and application servers too. Most Web and application servers use error mechanisms to display content access error pages, such as “404 page not found “and “500 internal server error”. If these pages reflect back any information from the user’s request, such as the URL they were trying to access, there are even greater chances that they are vulnerable to an XSS attack.
The possibility that a website contains XSS vulnerabilities is extremely high. There are countless ways to mislead Web applications into relaying maliciously injected scripts. Developers and website administrators seem to have a knack for missing these vulnerable application areas in their web implementations, but finding these configuration errors seems to be a walk in the park for attackers, since all they need is a browser and time (time which most of us don’t have).