Core Security Technologies issued an advisory disclosing a vulnerability that could affect millions of individuals and businesses using Microsoft’s Internet Explorer web browsing software.
A vulnerability researcher working in CoreLabs, the research arm of Core Security Technologies, discovered that in some cases when affected versions of Internet Explorer are used to access an external website, the browser does not apply the appropriate security permissions, thus allowing unknown sites or applications to be treated as trusted URLs. This could potentially lead to malicious or infected URLs remotely executing scripts on systems running the affected versions of IE, via either drive-by or downloaded attacks, without the end user’s knowledge or permission to do so.
CoreLabs initially discovered the vulnerability in Internet Explorer as part of its ongoing research efforts. The flaw specifically affects IE versions 5, 6 and 7 under Windows 2000/2003/XP and Vista. Although it is present, the vulnerability cannot be exploited when a vulnerable version of IE is used in a security-enhanced mode called “Protected Mode.” Protected Mode is enabled by default in IE 7 for Vista. At the time of the original report, Internet Explorer 8, then in the pre-release Beta phase, was also found to be vulnerable. However, the problem was fixed in the commercially released version of IE 8 and this version is therefore no longer vulnerable.
Internet Explorer utilizes a feature known as “URL Security Zones,” which defines a set of privileges for websites and applications depending on their apparent level of trustworthiness. IE zone settings include Internet Zone, Local Intranet Zone, Trusted Sites Zone (for URLs that are considered to be more reputable or trustworthy), Restricted Sites Zone (for websites that contain content that can cause or have previously caused problems) and Local Machine Zone (an unrestricted special zone that is used only by internal components of the operating system).
Internet Explorer users can assign specific websites or domains to any of the available zones except for the Local Machine Zone. The ability for a given website to perform security-sensitive operations on the web browser is determined by the Security Level of the zone to which the site was assigned. Each zone can be set to one Security Level out of three available pre-sets (Medium, Medium-High or High) or one customized by the user or system administrator.
By default, all websites that are determined not to be on the Intranet Zone and are not explicitly listed in the Restricted Sites or Trusted Sites Zones are assigned to the Internet Zone, which has a default Security Level setting of Medium-High. Thus, for most IE users the security-sensitive actions that a browser can perform while connected to a site on the Internet are those allowed by the security policy settings of the Internet Zone at the Medium-High Security Level.
A vulnerability that allows a website to perform security-sensitive actions that are disallowed by the Security Level of a given Security Zone is known as a Security Zone bypass or Security Zone elevation vulnerability and may lead to breach of other security policies, for example circumventing same domain policy restrictions.
Based on CoreLabs research, in some cases a malicious website may leverage a vulnerability and a combination of security weaknesses in the affected versions of Internet Explorer to bypass Security Zone restrictions by first serving HTML content that IE will cache in known locations in the user’s computer, and then redirecting the browser to load it from the local file system and render it as HTML. In this manner, arbitrary content provided by a potentially malicious site would able to run scripting code or ActiveX controls on vulnerable browsers and gain read access to any file stored in the user’s computer.
Exploitation of the vulnerability allows an attacker to retrieve security and privacy-sensitive data such as authentication credentials, HTTP cookies and other details of HTTP session state, as well as the contents of any local file. To successfully execute an attack, the attacker must either obtain or guess the username of the user visiting the website that delivers the exploit in order to predict the exact pathname to the cached content. In that context, lack of egress filtering for SMB connections, username leakage flaws or simple brute forcing of known usernames can facilitate attacks.
Active exploitation can be prevented by using several workarounds, which include setting the Security Level for the Internet Zone (and if necessary the Intranet Zone) to High; disabling ActiveX controls and scripting for the Internet Zone; using IE’s Protocol Lockdown feature control to restrict use of the “file:” protocol; or running IE in Protected Mode when available. These workarounds may limit or disrupt functionality of certain web applications that rely on security-sensitive actions to run properly.
The vulnerability is a variation of a similar issue catalogued by Microsoft as an Outlook Express/Windows Mail problem that involved serving HTML in the contents of HTTP cookies and remotely forcing IE to render it, which Core Security Technologies reported to the vendor in January 2008 and published, in coordination with the release of the corresponding patch by Microsoft, in August 2008 as CORE-2008-0103 Security Advisory.
Microsoft has received and acknowledged the report of this vulnerability in October 2008, fixed it for the release (RC and RTM) versions of Internet Explorer 8, and issued a security patch for the other vulnerable versions of IE.
Information on this patch and a related security bulletin published by Microsoft can be found here.