[VulnWatch] Script Injection to Custom HTTP Errors in Local Zone (GM#014-IE)
From: GreyMagic Software (security_at_greymagic.com)
To: "email@example.com" <firstname.lastname@example.org> Date: Tue, 17 Jun 2003 10:14:35 "GMT"
GreyMagic Security Advisory GM#014-IE
By GreyMagic Software, Israel.
17 Jun 2003.
Available in HTML format at http://security.greymagic.com/adv/gm014-ie/.
Topic: Script Injection to Custom HTTP Errors in Local Zone.
Discovery date: 18 Feb 2003.
Microsoft Internet Explorer 5.01, 5.5 and 6.0.
Note that any other application that uses Internet Explorer's engine
(WebBrowser control) is affected as well (AOL Browser, MSN Explorer, etc.).
Internet Explorer ships with various internal HTML resource files. The
majority of these files are meant to handle custom HTTP errors in web sites
(also called "Friendly HTTP error messages"). They all use the same basic
pieces of code, with minor changes to the actual content of each resource.
One of the main functions included in the resources is a method to extract
the real URL from the resource URL hash. For example, if "site.com"
generated a 404 HTTP error, the following URL will be internally requested
by IE: res://shdoclc.dll/404_HTTP.htm#http://site.com/file.html.
The function takes the part after the # sign and attempts to extract the
domain of the site, in order to embed it in the content of the custom
We found that the above-mentioned parsing procedure has a flaw in it that
may cause arbitrary script commands to be executed in the Local Zone.
Leading to potential arbitrary commands execution, local file reading and
other severe consequences.
However, Exploiting this procedure requires user-interaction. The user must
click the URL presented to it by the resource for the malicious code to
Here is the vulnerable function, precisely as it appears in the resources:
// in real bits, urls get returned to our script like this:
//For testing use DocURL =
DocURL = document.location.href;
//this is where the http or https will be, as found by searching for ://
but skipping the res://
//this finds the ending slash for the domain server
serverIndex=DocURL.indexOf("/",protocolIndex + 3);
//for the href, we need a valid URL to the domain. We search for the #
symbol to find the begining
//of the true URL, and add 1 to skip it - this is the BeginURL value. We
use serverIndex as the end marker.
//urlresult=DocURL.substring(protocolIndex - 4,serverIndex);
BeginURL=DocURL.indexOf("#",1) + 1;
if (protocolIndex - BeginURL > 7)
//for display, we need to skip after http://, and go to the next slash
displayresult=DocURL.substring(protocolIndex + 3 ,serverIndex);
// Security precaution: must filter out "urlResult" and "displayresult"
forbiddenChars = new RegExp("[<>\'\"]", "g"); // Global search/replace
urlresult = urlresult.replace(forbiddenChars, "");
displayresult = displayresult.replace(forbiddenChars, "");
document.write('<A target=_top HREF="' + urlresult + '">' +
displayresult + "</a>");
The comments in this function teach us that Microsoft had indeed attempted
to protect this resource from being exploited in this way, but unfortunately
failed to do so. A specially crafted value appended after the # sign can
Exploit and Demonstration:
document, which will execute when the user clicks on it:
Copy and paste the above URL in your browser, then click the red link in
order to test it.
Microsoft was notified on 20-Feb-2003. They were able to reproduce this on
IE6 Gold and all versions below it. We managed to reproduce it on all
versions, including IE6 SP1, with no exceptions.
They plan to fix this flaw in a future service pack.
The information in this advisory and any of its demonstrations is provided
"as is" without warranty of any kind.
GreyMagic Software is not liable for any direct or indirect damages caused
as a result of using the information or demonstrations provided in any part
of this advisory.
Please mail any questions or comments to email@example.com.
- Copyright © 2003 GreyMagic Software.