The Web Standards Group recommends that web developers stay up-to-date regarding vulnerabilities and threats which may affect the systems that they work with. The best practices and examples below are provided as a starting point for web security and should not be considered a “complete” or “current” resource.
Review the OWASP The Top 10 Application Security Risks list which is maintained by The Open Web Application Security Project to become familiar with the most current threats.
A vulnerability disclosure database is maintained and published by the National Cyber Security Division of the U.S. Department of Homeland Security. This database can be used to research vulnerabilities common to a specific development platform.
Reporting Security Incidences
The Security Operations Center (SOC) responds to reports of intrusions, denial of service and other information security incidents that take place on UCSB computers and Networks. To report an incident, send email to: email@example.com.
Please review Security Incident reporting information.
Secure Coding Practices
- Initialize all variables before use
- Validate all user input before use
- Restrict administrative permissions on servers and databases
- Handle errors and don’t display system error messages to end users
- Provide accounts with the least amount of permissions and privileges required
- Don’t store secrets (e.g. passwords, keys) in your code
- Use tested, reliable libraries or modules for common functions (e.g. authentication, encryption, session tracking)
Common Web Application Vulnerabilities and Mitigation
An attacker may attempt to authenticate into your application in a number of ways, including determining a correct username and password through attempting several random combinations (“brute force”), gathering the information from one of your users via a vulnerability, or simply bypassing the authentication process altogether through some security oversight.
A few practices for avoiding these security risks include:
- Secure login pages and pages protected by authentication with HTTPS
- Provide the same response for all authentication errors
- Check that the user has been authenticated on each secure page
Servers should only have the features and capabilities that are absolutely required. Installing unnecessary software or enabling extraneous options may create vulnerabilities within your server. Furthermore, servers should be configured to provide as little information as possible to external users and visitors.
Keep the following best practices in mind when administering a web server:
- Ensure that server components (OS, software/apps) are up-to-date
- Avoid installing unnecessary applications on production servers
- Remove unused and backup pages from the web server
- Uninstall programming language SDKs and runtimes that are not in use
- If possible, make code libraries and configuration files inaccessible from the web
- Disable directory browsing
- Avoid making operating system calls based on user input
There are a number of mechanisms for maintaining a user’s “state” while authenticated within your application, including encrypted variables within the URL, hidden form elements, and cookies. None of these methods are secure from an attack: URL variables can be easily modified and both hidden form elements and cookies can be manipulated with client-side browser development tools and proxies.
In general, the best practice is to make use of the session tracking mechanism built into your development framework.
There are a variety of attacks that stem from malicious input being fed into a system, including operating system command injection, buffer overflow attacks, SQL injection attacks, and cross-site scripting attacks. In most cases, these problems are caused by a central root cause: feeding user input into a system without properly sanitizing it.
To help protect from these issues, keep the following best practices in mind:
- Avoid basing system calls or database calls on user input
- Sanitize user input using a mechanism built in to your development framework
- Ensure user input is an expected and valid length
- Ensure that any errors reported to the user do not provide insight to web internals
- Log detailed error information to a monitored log file
- If outputting user input on a web page, sanitize the output for display
Provide secure HTTPS connections, enabled via Transport Layer Security (TLS), using valid, up-to-date certificates, across your entire website, not simply areas of the website that handle secure information.
This recommendation is echoed by the White House Office of Management and Budget (OMB)’s HTTPS-Only Standard directive.