Authorization token leak from verify email endpoint
While testing a website I found that the verify email endpoint was leaking the authorization tokens of any verified users by just passing their email as an input.
As I can not disclose the website name let's say the website is example.com
As the program had one domain in its scope. I opened the burp suite in the background to collect all the requests and responses so I could review them later. First I used the website like an average user to understand its functionality and flow of the website.
Then I proceeded to register a new user for the application. The email verification was done using a POST request with “email” and “code” as a parameter.
POST /api/user/verifymail HTTP/1.1
In response, email verification was successful and the token was displayed along with other information.
After noticing the email parameter in the request, I thought of registering two new users using different emails and test if the code assigned for the first user’s email can be used to verify the second user. But the test failed. Then I tested by removing the code parameter from the request and just forwarded the request with the email only. The application did not verify the email. So with no luck, I just went on to test the other functionalities of the application.
After some testing, I was closing the tabs in the repeater on the burp suite. Then I saw /api/user/verifymail endpoint with just email as a parameter. I clicked on send and to my surprise, the response was 200 OK with the email verified successful message along with the token and other information. As I had already verified this email before, it showed the email verified successful again without displaying any error message like email already verified or verification failed.
I got curious and started testing the endpoint further. I found that if an already verified email is passed as a value in the email parameter, the application responded with the token. Using the token obtained, any action can be performed on behalf of the victim.
Then I tried with the admin account of the application which I found during the testing process. The admin token was leaked in the response.
Then I created a POC and reported the vulnerability.
Using this vulnerability an attacker could obtain tokens of any verified users by just using the email address of the victim.