Lesson Plan Title: How to Perform Http Splitting

 

Concept / Topic To Teach:

This lesson teaches how to perform HTTP Splitting attacks.

 

How the attacks works:

The attacker passes malicious code to the web server together with normal input. A victim application will not be checking for CR (carriage return, also given by %0d or \r) and LF (line feed, also given by %0a or \n)characters. These characters not only give attackers control of the remaining headers and body of the response the application intends to send, but also allows them to create additional responses entirely under their control.
The effect of an HTTP Splitting attack is maximized when accompanied with a Cache Poisoning. The goal of Cache Poisoning attack is to poison the cache of the victim by fooling the cache to believe that the page hijacked using the HTTP splitting is a good one and it is indeed the server's copy.
The attack happens using the HTTP Splitting attack plus adding the Last-Modified: header and setting it to a future date. This will force the browser to send If-Modified-Since request header, which gives the attacker the chance to intercept the server's reply and replace it with a '304 Not Modified' reply. A sample of a 304 response is:
HTTP/1.1 304 Not Modified
Date: Fri, 30 Dec 2005 17:32:47 GMT

 

General Goal(s):

This lesson has two stages. Stage 1 teaches you how to do HTTP Splitting attacks while stage 2 builds on that to teach you how to elevate HTTP Splitting to Cache Poisoning.
Enter a language for the system to search by. You will notice that the application is redirecting your request to another resource on the server. You should be able to use the CR (%0d) and LF (%0a) to exploit the attack. Your exercise should be to force the server to send a 200 OK. If the screen changed as an effect to your attack, just go back to the homepage and after stage 2 is exploited successfully you will find the green check in the left menu.

 

 

Solution:

Please note that this solution is written for Windows. If you use Linux you have to alter it. Windows uses a CR and LF for new Line. Linux uses only LF. So all the %0d%0a have to be replaced by %0a if you are using Linux.

Because the input is not validated you can inject any HTTP syntax, carriage returns and line-feed you want.

 

Enter a language to examine what’s going on. You do have WebScarab intercepting HTTP requests and responses?

 

Figure 1 Language en

Figure 2 HTTP Request

 

Figure 3 First HTTP Response

 

Figure 4 Second HTTP Request

 

Now inject for the language en%0d%0a%0d%0a%0d%0a

 

Figure 5 First HTTP Request

 

Figure 6 First HTTP Response

 

The Content-Length: 0 will tell the server that the first request is over.

A 200 OK message looks like this: HTTP/1.1 200 OK

 

Lets see what you can do with: foobar%0d%0aContent-Length:%200%0d%0a%0d%0aHTTP/1.1%20200%20OK%0d%0aContent-Type:%20text/html%0d%0aContent-Length:%2047%0d%0a%0d%0a<html>Hacked J</html>

 

Figure 7 HTTP Splitting attack

 

Figure 8 HTTP Response

 

Figure 9 Second HTTP Request

 

Figure 10 Second HTTP Response

 

Figure 11 Hacked!

 

Hit the “Back” button of your browser.

 

Figure 12 Stage 1 completed

 

Now you know how to do HTTP Splitting. You can abuse this technique to do a cache poisoning attack.

 

Cache poisoning requires manipulating the Last-Modified header. This must be changed to a date in the future.

Inject: foobar%0d%0aContent-Length:%200%0d%0a%0d%0aHTTP/1.1%20200%20OK%0d%0aContent-Type:%20text/html%0d%0aLast-Modified:%20Mon,%2027%20Oct%202003%2014:50:18%20GMT%0d%0aContent-Length:%2047%0d%0a%0d%0a<html>Hacked J</html>

 

Figure 13 Inject cache poisoning

 

Solution by Erwin Geirnaert ZION SECURITY