Cookies Continued...part 3

Now since we have a general idea of Cookies, it's time to reverse engineer them. But before we do this, we must first understand how to provide "a secure method of session management". In other words, how do we keep our cookie safe? Well, the first step is to make the data "hard-to guess" or unpredictable. In doing this, we are making it hard to forge the cookie which makes it harder to break into another user's session. To make it unpredictable, we have to create random values or cryptography.  Second step is to make it tamper resistance. Having a cookie with a parameter isAdmin and value "no" is not smart. It is extremely easy to temper with and change the "no" to  a "yes".  A better way of doing it is encrypting the  parameter and value.  Third step would to make sure the cookie does expire at a certain point. Make the cookie only exists in the time period that it should. Lastly, for important cookies, make sure to use the secure flag so transmission of the cookie can only occur in an encrypted channel.

If an application follows the above guidelines when generating cookies, it will be extremely difficult for hackers to temper with the cookies. Unfortunately, some cookies are not this safe and that is why reverse engineering can be done. The basic premise is that we get a handful of cookies from a website and look for patterns.  When collecting cookies, make sure the website is in the same state for all cookies collected. For example, if the first cookie collected was right after authentication, then the rest of the cookies of the sample as to be collected at the same time.  It is also important to note down the time when the cookie was collected. As stated in the last post, the time the cookie was created can be part of the cookie value.

When looking to reverse engineer a cookie, it is good to know the value of type or types. Many times, a cookie can hold numerous values of different types.  Below is an example of a cookie.

ID=5a0acfc7ffeb919:CR=1:TM=1120514521:LM=1120514521:S=j3am5KzC4v01ba3q

In the next post, we'll break down what each part of the cookie means.

Cookies Continued...pt2

So our first step is cookie collection. In here we're trying to figure out to answer some basic questions such as :

• "how many cookies does the application create?"
• The way I did this is by first clearing all the cookies in my browser, and then visiting the application. Then, I check the cookie list.  Doing this tells me about all the cookies created when accessing the application.

• "which parts of the application generates a cookie?"
• Go to different parts of the web application and check the browser to see if any cookies are created.

• "which parts of the application causes a modification of the cookie?".
• Same as above - when we're messing around with the application, see if the status of any of the cookies has changed.

Our next step is to analyze the cookie. For example, look at the session ID of multiple cookies. Are they just linearly made? Are they random? Does duplicates exists?  We can also examine the value of the cookie. Often times, developers make the mistake of storing actual sensitive data in the cookie. What their actually suppose to do is store a generic value and reference the real data using the generic value. Some developers store real data in cookies and just encode various methods like Hex, Base64, or MD5. What they fail to understand is that these strings can easily be decoded and the real data can be seen by unwanted eyes.

If  tokens are in fact encoded, they could contain multiple sets of data in one long string. For example, if the token is 32 bit string, the first 16 bits could represent something static like a username or IP address and the next 16 bits could by dynamic. If this second 16 bit string changes incrementally, this could represent a date or time for each access.  One way to verify this is access the site at the same time using different browsers and using the same username/IP address. In that case, the cookies from each browser should be almost the same.