Severity: Low to Medium
Unchecked redirects occur when an application redirects to a destination controlled by attackers. This often occurs in functionality returning users to a previous page, e.g. after authenticating.
http://example.com/?return_to=http://example.com/home
An attacker can control the value of the return_to parameter and make it redirect to a malicious endpoint.
http://example.com/?return_to=http://evil.com
Here is a proof of concept to demonstrate how an open redirect occurs. Please note that this particular example is not a vulnerability and just here for demonstration purposes.
https://httpbin.org/redirect-to?url=https://www.hacker101.com
The URL looks like it should go to httpbin.org, but you are redirected to www.hacker101.com.
Attackers may be able to use this to execute believable phishing attacks, bypass authentication, or (in rare circumstances) violate CSRF mitigations.
When possible, do not allow user input to directly control redirect destinations; rather, generate them on the server side (e.g. via ID -> URL mapping). When this is not an option, a strict whitelist is highly recommended. Finally, a last-ditch mitigation can be performed by removing protocol specifiers from user input prior to redirection. This last method will not fix intra-site redirect exploits, but can prevent redirects to an attacker-controlled website.