In this article, We’ll talk about the topic of cross-site request forgery (CSRF) vulnerabilities and I’ll share my personal experience of successfully executing a one-click account takeover on an e-commerce website. This serves as a reminder of the critical role that proper security measures play in keeping our sensitive information safe.
Cross-Site Request Forgery (CSRF) is a type of security vulnerability that affects web applications. It happens when a malicious website is able to trick a user’s browser into sending a request to another website, without the user’s knowledge or consent. This can be used to perform unauthorized actions, such as modifying the user’s profile, posting unwanted content, or making unauthorized purchases.
How it works?
Imagine that a user, let’s call him victim A, has logged into his account on an online bank website that is vulnerable to CSRF. Then he send a request to transfer money to his nephew, Little Timmy. The normal request would look something like this:
After that, let’s say victim A visits a malicious website created by attacker B. Since the online bank website has a CSRF vulnerability, the malicious website could send a forged prefilled request to the bank website to transfer money to attacker B, without victim A even noticing. This is possible because victim A’s browser automatically sends a valid session token with each request, which allows the attacker B to take actions on behalf of the victim A.
How to protect?
To protect against CSRF attacks, websites need to implement proper security measures. One common method is to include a unique token, known as a CSRF token, with each form submission. This token is checked on the server to make sure that the request came from the same website that it was intended for. This way, even if a malicious website sends a request, the server will know it’s not legitimate and reject it.
In summary, CSRF is a serious security vulnerability that can have devastating consequences for both users and websites. It’s important for web developers to be aware of this threat and take the necessary steps to protect their users.
The target is an E-commerce website from United States. Like most E-Commerce websites, this website provided a feature edit profile on customer’s account. I can’t disclose the name of the website, so let’s call it as redacted.com 🙂
First, I created a dummy victim account and filled in the details.
During my testing of updating a user profile using Burp Suite, I noticed that the server uses a POST request to send the updated information and there is no protection against CSRF in place. The profile update request appears as follows:
POST /secure/profile/profile_form_handler.php HTTP/2
User-Agent: Mozilla/5.0 (Windows VT 10.0; Win64; x64, rv:102.0) Gecko/20100101 Firefox/102.0
Accept: text/html, */*; q=0.01
Accept-Encoding: grip, deflate
Content-Type: application/x-www-form-urlencoded, charset=UTF-8
Now, I copied this CSRF POC from https://hackerone.com/reports/834366 , modify the data, and saved it as an html page. Here’s my POC:
<form action="https://redacted.com/secure/profile/profile_form_handler.php" method="POST">
<input type="hidden" name="nextpage" value="%2Fsecure%2Fprofile%2Fupdated_profile"/>
<input type="hidden" name="prevpage" value="%2Fsecure%2Fprofile%2Fprofile"/>
<input type="hidden" name="selectedRadialValue" value="790"/>
<input type="hidden" name="txtEmail" value="email@example.com"/>
<input type="hidden" name="txtFirstName" value="hacked"/>
<input type="hidden" name="txtLastName" value="hacked"/>
<input type="hidden" name="txtDayPhone" value="999999999"/>
<input type="hidden" name="hdnSmsPhone" value="999999999"/>
Then I opened the POC in a victim’s browser, and it displayed a “success” response. Upon re-visiting the redacted.com profile page, I observed that the attacker had successfully updated the victim’s profile.
Jackpot! Just one click from the victim and the victim profile is entirely updated. At this point, the integrity aspect of the user has been compromised, but can we do further?
As victim’s email has been replaced, how to escalate this into account takeover? Yep, forgot password!
Next, I opened a incognito browser window to assume the role of the attacker. I visited the redacted.com login page and clicked on the “forgot password” link. Knowing that the victim’s email was changed to “firstname.lastname@example.org” that the attacker have access to it, I requested for the password reset link to be sent to that email.
A password reset link was then sent to the attacker’s email. I opened the link, entered a new password, and that was it. I was able to take over the victim’s account without ever having access to their original email or password 🙂
I immediately reported this vulnerability to the concerned enterprise through its Public Vulnerability Disclosure Program on Hackerone.
Thank you for reading this article 🙂
- Verify the presence of a CSRF token in all sensitive API requests using Burp Suite.
30 Jun 2022 : Bug was discovered.
1 Jul 2022 : Bug was reported to Hackerone public program.
1 Jul 2022 : Bug Status Triaged as High Severity.
27 Oct 2022 : Bug fixed (ineligible for bounty).