In ScratchVerifier before commit a603769, an attacker can hijack the verification process to log into someone else's account on any site that uses ScratchVerifier for logins. A possible exploitation would follow these steps: 1. User starts login process. 2. Attacker attempts login for user, and is given the same verification code. 3. User comments code as part of their normal login. 4. Before user can, attacker completes the login process now that the code is commented. 5. User gets a failed login and attacker now has control of the account. Since commit a603769 starting a login twice will generate different verification codes, causing both user and attacker login to fail. For clients that rely on a clone of ScratchVerifier not hosted by the developers, their users may attempt to finish the login process as soon as possible after commenting the code. There is no reliable way for the attacker to know before the user can finish the process that the user has commented the code, so this vulnerability only really affects those who comment the code and then take several seconds before finishing the login.
When an actor claims to have a given identity, the product does not prove or insufficiently proves that the claim is correct.
| Link | Tags |
|---|---|
| https://github.com/ScratchVerifier/ScratchVerifier/security/advisories/GHSA-99cr-hvf7-85g9 | third party advisory |
| https://github.com/ScratchVerifier/ScratchVerifier/commit/a603769010abf8c1bede91af46e4945314e4aa4a | third party advisory patch |