<div>There is nothing that you have provided that indicates that this is a new class of vulnerability; this is a classic state management issue. The application is designed to have a password reset function, and in order for that function to behave properly application state must be enforced.
</div>
<div> </div>
<div>Essentially the issue here is that the application is not maintaining its state properly, and the result is that an attacker can force the application to engage in some unplanned functionality. Any web application must leverage some form of state management, and this is pretty easy for most developers to learn. What most developers have difficulty with is the idea that process flow should be enforced; if you have a 'Registration' module with 3 states, and a 'Password Reset' module with 4 states, then any state transition from the Registration module into the Password Reset module should result in the use being forced to start at Password Reset modules initial state.
</div>
<div> </div>
<div>Without a strong means of enforcing the application logic it is trivial to move an application from a known to an unknown state, and this is never a good thing in an application, let alone an application that must make decisions about security based on application state.
</div>
<div> </div>
<div>The reuse of the 'login' session variable in your example is still a case of poor software development, but if application state was properly maintained, then it wouldn't be an issue.<br> </div>
<div><span class="gmail_quote">On 1/13/06, <b class="gmail_sendername">Alla Bezroutchko</b> <<a href="mailto:alla@scanit.be">alla@scanit.be</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Frank Knobbe wrote:<br>> The proposed fix is -- besides being only specific to this example --<br>> equally flawed. The underlying issue is that you trust user supplied
<br>> data. When a user supplies a user name for login purposes, you should<br>> only use that input to perform a search in your database. If a match has<br>> been found, take a *trusted* value from your database, for example an
<br>> index ID, and carry that in the session object to identify the user.<br><br>Suppose in my example resetpw3.php instead of doing<br><br>$_SESSION['login'] = $_POST['login'];<br><br>did<br><br>$_SESSION['login'] = $db->getOne("SELECT login FROM users WHERE login=?
<br>AND secret_answer=?", array($_POST['login'], $_POST['secret_answer']));<br><br>As you suggest it takes a trusted value from the database. It is still<br>does not prevent using register2.php to initialize $_SESSION['login']
<br>and then jump to resetpw4.php and have it use the value to change password.<br><br>> In other words, don't accept any user input (even after proper input<br>> validation) and carry it as trusted data in your session object. Don't
<br>> base further decisions on this data. Since it is user supplied it can<br>> not be trusted.<br><br>Are you saying that user supplied data should never ever be stored in<br>the session? Where do you suggest to keep the data that needs to be
<br>passed between different pages in the application?<br><br>Say you have a form that is filled in in several steps, each step being<br>a separate page. Where would you store the data that the user entered to<br>have it all available on the last page? I can only think of carrying it
<br>around in hidden form fields, and this is not always possible.<br><br>> Your example is further flawed in that it allows the change of a<br>> password without being properly authenticated. Just having a valid<br>
> 'login' session object present doesn't indicate that the user is<br>> authenticated.<br><br>I think it does, if 'login' session object is only ever set by the<br>authentication procedure after verifying user's credentials. I believe
<br>the usual approach to authentication in web applications is to check<br>user's credentials and then store something in the session that<br>indicates that the credentials were correct. Usually it is something<br>like a user ID or an object or JavaBean storing user information. Then
<br>every pafge that needs to check auth just checks if the session object<br>is present and valid.<br><br>How do applications that you encounter check if the user is authenticated?<br><br>> I really, really hope this was just an example you made
<br>> up, and not something you actually saw being used. If so, go back to<br>> that web site with a clue-by-four and give it a few whacks.<br><br>I work for a company that does penetration testing, vulnerability<br>
assessments, and stuff like that. The contracts we usually sign do not<br>authorise us to apply blunt instruments to customers' web sites. :)<br><br>This particular example is made up. Giving real ones would violate<br>non-diclosure agrements. It is similar to the real ones in the way that
<br>it has the same problem - one storage location is used to keep both<br>trusted and untrusted data. I used the password reset example because<br>the security implications are clear and one don't need application<br>specific context to understand it. The bug does not have to do anything
<br>with authentication as such.<br><br>And, in case you are wondering, the real ones were just as bad. :)<br><br>Alla.<br>_______________________________________________<br>Full-Disclosure - We believe in it.<br>Charter:
<a href="http://lists.grok.org.uk/full-disclosure-charter.html">http://lists.grok.org.uk/full-disclosure-charter.html</a><br>Hosted and sponsored by Secunia - <a href="http://secunia.com/">http://secunia.com/</a><br></blockquote>
</div><br><br clear="all"><br>-- <br>____<br>ygjb<br>Computer Science is no more about computers than astronomy is about telescopes. E. W. Dijkstra