Difference between revisions of "Reviewing The Secure Code Environment"
Revision as of 04:51, 11 January 2007OWASP Code Review Guide Table of Contents
Contact author: Eoin Keary
Secure Code Environment
Another important thing to be aware of is when you receive the code make sure it is identical in deployment layout to what would go to production. Having well-written code is a great start, but deploying that great code in unprotected folders on the application server is not a great idea. Attackers do code reviews also and what better than to code review the potential target application. For example: try in “Google”: http://www.google.com/search?q=%0D%0Aintitle%3Aindex.of+WEB-INF
This lists exposed “Web-Inf” directories on WebSphere®, Tomcat and other app servers.
The WEB-INF directory tree contains web application classes, pre-compiled JSP files, server side libraries, session information and files such as web.xml and webapp.properties.
So be sure the code base is identical to production. Ensuring that we have a “secure code environment” is also an important part of an application secure code inspection.
The code may be “bullet proof” but if it is accessible to a user this may cause other problems. Remember the developer is not the only one to perform code reviews, attackers also do this. The only visible surface that a user should see are the “suggestions” rendered by the browser upon receiving the HTML from the backend server. Any request to the backend server outside the strict context of the application should be refused and not be visible. Generally think of “That which is not explicitly granted is denied”.
Example of the Tomcat web.xml to prevent directory indexing:
<servlet> <servlet-name>default</servlet-name> <servlet-class>org.apache.catalina.servlets.DefaultServlet</servlet-class> <init-param> <param-name>debug</param-name> <param-value>0</param-value> </init-param> <init-param> <param-name>listings</param-name> <param-value>false</param-value> </init-param> <init-param> <param-name>readonly</param-name> <param-value>true</param-value> </init-param> <load-on-startup>1</load-on-startup> </servlet>
So to deny access to all directories we put:
<Directory /> Order Deny,Allow Deny from All </Directory>
And then override this for the directories we require access to:
Also in Apache HTTP server to ensure directories like WEB-INF and META-INF are protected the following should be added to the httpd.conf, the main configuration file for the Apache web server
<Directory /usr/users/*/public_html> Order Deny,Allow Allow from all </Directory> <Directory /usr/local/httpd> Order Deny,Allow Allow from all </Directory>
On Apache servers, if we wish to specify permissions for a directory and subdirectories we add a .htaccess file.
To protect the .htaccess file itself we palce:
<Files .htaccess> order allow,deny deny from all </Files>
To stop directory indexing we place the following directive into the .htaccess file: IndexIgnore * The * is a wildcard to prevent all files from being indexed.
Protecting JSP pages
If using the Struts framework we do not want users access any JSP page directly. Accessing the JSP directly without going through the request processor can enable the attacker to view any server-side code in the JSP. Lets say initial page can is a HTML document. So the HTTP GET from the browser retrieves this page. Any subsequent page must go through the framework. Add the following lines to the web.xml file to prevent users from accessing any JSP page directly:
<web-app> ... <security-constraint> <web-resource-collection> <web-resource-name>no_access</web-resource-name> <url-pattern>*.jsp</url-pattern> </web-resource-collection> <auth-constraint/> </security-constraint> ... </web-app>
With this directive in web.xml a HTTP request for a JSP page directly will fail.
A clean environment
When reviewing the environment we must see if the directories contain any artefacts from development. These files may not be referenced in any way and hence the application server gives no protection to them. Files such as .bak, .old, .tmp etc should be removed as they contain source code.
Source code should not go into production directories. The complied class files are all that is required in most cases. All source code should be removed and only the “executables” should remain.
No development tools should be present on a production environment. For example a java application should only need a JRE (Java Runtime Environment) and not a JDK (Java Development Kit) to function.
Test and debug code should be removed from all source code and configuration files. Even commented out code should be removed as a precaution. Test code can contain backdoors that circumvent the workflow in the application and at worst contain valid authentication credentials or account details.
Comments on code and Meta tags pertaining to the IDE used or technology used to develop the application should be removed. Some comments can divulge important information regarding bugs in code or pointers to functionality. This is particularly important with client side code such as JSP’s and ASP files.
A copyright and confidentiality statement should be at the top of every file. This mitigates any confusion regarding who owns the code. This may seem trivial but it is important to state who owns the code.
To sum up, code review includes looking at the configuration of the application server and not just the code. Knowledge of the server in question is important and information is easily available on the web.