Web Application Security

Why You Should Overload WebSite Errors

I often have to explain to people that they need to overload error messages on their websites, and I occasionally get weird looks. It’s important to know that there are three scenarios that regularly occur that can be used to an attacker’s advantage.

The first scenario is where an error occurs in the application logic and creates a 500 error (Internal Server Error). These errors can often include stack traces, which provide incredibly useful information for crafting SQL injection attacks, among others. There’s no reason a user should see these.

The second is 400 errors (Bad Request). If an attacker can send overly large cookies, the server will respond with a bad request and will include the cookie payload in the response. This response, which includes the over-long cookie, can be read by an XMLHTTPRequest, even if the HTTPOnly flag is set. This allows an attacker who can set cookies (via XSS for instance) to break the protection that HTTPOnly provides, so it’s critical that these errors are trapped.

The third scenario is that the difference between a 404 (File Not Found) and a 403 (Forbidden) is noticeable to a scanner and can help tools that do directory enumeration figure out which directories exist verses aren’t there, at which point it is possible to brute force the files that are contained within. Ideally all directories that have Directory Indexing turned off as well as files that aren’t there should respond with 200 OK to confuse scanners. Disabling Directory Indexing is important and useful, but so is making it impossible to tell which directories are there or not there. Overloading a 404 (File Not Found) error is also useful in that it can be used to your advantage for instance, to tell users in a graceful way that they can go back to another page, offer a search box to find what they need, or whatever makes sense for your users.

There may be many other use cases as well, but this should give you some ideas as to why it’s important to trap errors. Don’t worry that you will lose visibility into the problems your site may have. Just because the user can’t see it doesn’t mean you can’t. You can disable this type of error trapping in QA, and you can review your logs and identify pages that are problematic. This allows you to fix your issues without introducing additional security issues.

  • http://anantshri.info Anant Shrivastava

    Another tip that i have used at few places. overload pages with your home page, so that the overloaded pages load your home screen which will again be full of links. Makes the spider go wonky. plus to avoid making that 404 page you should keep adding some extra garbage just to make sure that two response of errors are never equal.