open-formulieren vulnerabilities and threats

Open Forms allows users create and publish smart forms. Prior to 3.3.13 and 3.4.5, to be able to cosign, the cosigner receives an e-mail with instructions or a deep-link to start the cosign flow. The submission reference is communicated so that the user can retrieve the submission to be cosigned. Attackers can guess a code or modify the received code to look up arbitrary submissions, after logging in (with DigiD/eHerkenning/... depending on form configuration). This vulnerability is fixed in 3.3.13 and 3.4.5.

Open Forms allows users create and publish smart forms. Prior to versions 3.2.7 and 3.3.3, forms where the prefill data fields are dynamically set to readonly/disabled can be modified by malicious users deliberately trying to modify data they're not supposed to. For regular users, the form fields are marked as readonly and cannot be modified through the user interface. This issue has been patched in versions 3.2.7 and 3.3.3.

Open Forms allows users create and publish smart forms. Versions prior to 2.2.9, 2.3.7, 2.4.5, and 2.5.2 contain a non-exploitable multi-factor authentication weakness. Superusers who have their credentials (username + password) compromised could potentially have the second-factor authentication bypassed if an attacker somehow managed to authenticate to Open Forms. The maintainers of Open Forms do not believe it is or has been possible to perform this login. However, if this were possible, the victim's account may be abused to view (potentially sensitive) submission data or have been used to impersonate other staff accounts to view and/or modify data. Three mitigating factors to help prevent exploitation include: the usual login page (at `/admin/login/`) does not fully log in the user until the second factor was succesfully provided; the additional non-MFA protected login page at `/api/v2/api-authlogin/` was misconfigured and could not be used to log in; and there are no additional ways to log in. This also requires credentials of a superuser to be compromised to be exploitable. Versions 2.2.9, 2.3.7, 2.4.5, and 2.5.2 contain the following patches to address these weaknesses: Move and only enable the API auth endpoints (`/api/v2/api-auth/login/`) with `settings.DEBUG = True`. `settings.DEBUG = True` is insecure and should never be applied in production settings. Additionally, apply a custom permission check to the hijack flow to only allow second-factor-verified superusers to perform user hijacking.