Srikar Nadipally
XSS Vulnerabilities
Outline
Finding and Exploiting XSS VulnerabilitiesStandard
Reflected XSS
Stored XSS
DOM based XSS
Prevention of XSS attackReflect
Stored
DOM
Standard
Use standard proof of concept
“><script>alert(document.cookie)</script>
Detecting Reflected XSS
Systematic approach of finding reflected attackFind all the entry points of the user input
Submit a benign alphabetical string in each entry point
Identify all the locations where the string is reflected in the applications response
For each reflection identify the syntactic context in which string appears
Submit modified data tailored to the reflection syntactic context, attempting to introduce the arbitrary script in to response.
If reflected data is blocked, try to understand and circumvent the application defensive filter
Example 1
Tag attribute valueReturn page contains the code
<input type=“text” name=“address1” value=“myxssscript”>
Ways to craft XSS exploit “><Script>alert(1)</script>“ onfoucus=“alert(1)
Example 2
Java Script String
If return page code is
<script> var a=‘myxsstest’; var b =123; </script>
Ways to craft XSS exploit
‘; alert(1); var foo=‘
Example 3
Atrribute Containing URL
Returned page code
<a href=“myxssscript”>Click here…</a>
Ways to craft XSS exploito javascript:alert(1)o #”onclick=“javascript:alert(1)
Stored XSS
Stored XSS vulnerability identification is quite similar to that of reflected XSS – submitting a unique string in every entry point within the application.
Once you identify every instance in which user controllable data is stored by the application and later displayed back to the browser, same process is followed as that of reflected XSS – determining what input to be submitted to embed valid JavaScript within the surrounding HTML and then besiege the filters that intervene with your attack payload process.
Techniques to test for stored XSS
Testing for XSS in Web Mail applicationsSend all kinds of unusual HTML content within emails to test for bypasses in input filters.
Restricting to standard email client will not give you enough control over the raw message content or the client may itself sanitize or clean up your malformed syntax.
Using UNIX sendmail command a raw email can be created in a text editor and send it.
Sendmail –t [email protected] < email.txt
Example
Raw email fileMIME-Version: 1.0From: [email protected]: text/html; charset=us-asciiContent-Transfer-Encoding: 7bitSubject: XSS test<html><body><img src=``onerror=alert(1)></body></html>
Techniques to Detect stored XSS
Testing for XSS in Uploaded Files
Hybrid File attacks – “hybrid files” - two different formats
Example: GIFAR (GIF + JAR)
Uploaded file attack using GIFAR
The attack using GIFAR files can be prevented in current versions of Java browser plug-in by validating whether JAR files being loaded actually content hybrid content.
Detecting DOM-based XSS
Manually walk through the application with your browser and modify each URL parameter to contain a standard test string.
“<script>alert(1)</script>
“;alert(1)//
‘-alert(1)-’
Displaying each returned page in the browser causes all client side scripts to execute referencing the modified URL parameter.
If you see a dialog box containing cookies, you will have found a vulnerability.
Detecting DOM-based XSS
Effective approach:
Review all client-side JavaScript for any use of DOM properties that may lead to a vulnerability.
DOM-tracer is a tool that helps you to automate this process
Preventing XSS attacks
Due to the different root causes different defense mechanisms needs to be applied for reflected and stored XSS on one hand and DOM-based on the other.
Reflected and Stored XSSIdentify every instance within the application where user-controllable data is being copied into responses including data that is copied from immediate request and also any stored data that is originated from any user at any prior time, including via out-of-band channels.
After identification, follow a threefold approach to prevent any actual vulnerabilities from arising.
…Continued
Threefold approach-Validate input
Validate output (Encode the output)
Eliminate dangerous insertion points
Validate input: The application should perform context-dependent validation of data when application receives user-supplied data that may copy into one of its responses at any future point.
Potential features to validate – data is not too long, contains only a certain permitted character set, matches a particular regular expression.
Different Validation rules should also be applied – names, email id’s, account numbers etc.
…Continued
Validate OutputData should be HTML-encoded to sanitize potentially malicious characters.
HTML encoding involves replacing literal characters with their corresponding HTML entities.
HTML encodings of the primary problematic characters are as follows –
“ — "
‘ — '
& — &
< — <
> — >
…Continued
Eliminate dangerous insertion pointsInserting user-controllable data directly into existing script code should be avoided wherever possible. This applies to the code within <script> tags, and also code within event handlers.
Allowing limited HTML
…Continued
Preventing DOM-based XSSApplication should avoid using client-side scripts to process DOM data and insert it into the page.
DOM-based XSS flaws can be prevented through two types of defenses-
Validate input
Validate output
Validate input:Attack can be prevented by validating the data about to be inserted into the document containing only alphanumeric characters and white space.
…Continued
Example:<script>
var a = document.URL;
a = a.substring(a.indexOf(“message=”) + 8, a.length);
a = unescape(a);
var regex=/^([A-Za-z0-9+\s])*$/;
if (regex.test(a))
document .write(a);
</script>
…Continued
Server-side validation can also be employed to detect requests that may contain malicious exploits by verifying the following:
The query string contains a single parameter.
The parameter’s name is message (case-sensitive check).
The parameter’s value contains only alphanumeric content
…Continued
Validate output:Applications can perform HTML encoding of user-controllable DOM data before it is inserted into the document.
HTML encoding can be implemented in client-side Javascript with a function like the following:
function sanitize(str)
{
var d = document.createElement(‘div’);
d.appendChild(document.createTextNode(str));
return d.innerHTML;
}
Request forgery
Also known as “Session riding” is related to session hijacking attacks, in which an attacker captures a user’s session token and thereby uses that application as that user.
Request Forgery vulnerabilities comes in two flavors:
On-site Request Forgery (OSRF)
Cross-Site Request Forgery (CSRF)
OSRF
Attack for exploiting stored XSS vulnerabilities.
OSRF vulnerabilities can exist even in situations where XSS is not possible.
Consider a message board application that lets users submit items that are viewed by other users. Messages are submitted using a request like the following:
POST /submit.phpHost: wahh-app.comContent-Length: 34type=question&name=daf&message=foo
…Continued
This request results in the following being added to the messages page:
<tr> <td><img
src=”/images/question.gif”></td> <td>daf</td> <td>foo</td>
</tr>
CSRFAttacker creates the innocuous-looking website causes the user’s browser to submit a request directly to the vulnerable application to perform some unintended action that is beneficial to the attacker.
CSRF attacks are “one-way” only.
Consider an application in which administrators can create new user accounts using requests like the following:
POST /auth/390/NewUserStep2.ashx HTTP/1.1Host: mdsec.netCookie: SessionId=8299BE6B260193DA076383A2385B07B9Content-Type: application/x-www-form-urlencodedContent-Length: 83
realname=daf&username=daf&userrole=admin&password=letmein1&confirmpassword=letmein1
…Continued
This request has three key features that make it vulnerable to CSRF attacks:
The request performs a privileged action. In the example shown, the request creates a new user with administrative privileges.
The application relies solely on HTTP cookies for tracking sessions. No session-related tokens are transmitted elsewhere within the request.
The attacker can determine all the parameters required to perform the action. Aside from the session token in the cookie, no unpredictable values need to be included in the request
…ContinuedAttacker can construct a web page that makes a cross-domain request to the vulnerable application containing everything needed to perform the privileged action.
Example of such attack:<html><body><form action=”https://mdsec.net/auth/390/NewUserStep2.ashx” method=”POST”><input type=”hidden” name=”realname” value=”daf”><input type=”hidden” name=”username” value=”daf”><input type=”hidden” name=”userrole” value=”admin”><input type=”hidden” name=”password” value=”letmein1”><input type=”hidden” name=”confirmpassword” value=”letmein1”></form><script>document.forms[0].submit();</script></body></html>
Questions