heisec@social.heise.de - BSI warnt vor KeePassXC-Schwachstellen
Das BSI warnt vor Schwachstellen im Passwort-Manager KeePassXC. Angreifer können Dateien oder das Master-Passwort ohne Authentifzierungsrückfrage manipulieren.
[The BSI warns of vulnerabilities in the password manager KeePassXC. Attackers can manipulate files or the master password without authentication confirmation.]
KeePassXC is not affected by this vulnerability.
CVE-2023-32784
That’s not the issue mentioned in the article. Which is
CVE-2023-35866.This is also the vulnerability that made many people delete Keepass 2 for XC many months ago so it is very strange that they make an article that sounds like it’s a new vulnerability.
Wrong vulnerability. The discovered one is CVE-2023-35866, which is still pending verification* (analysis).
This affects KeePassXC. https://nvd.nist.gov/vuln/detail/CVE-2023-35866
Thanks for the correction. In that case going to be interesting how this issue progress.
Here is KeePassXC’s response: https://keepassxc.org/blog/2023-06-20-cve-202335866
Basically some random guy with weird misconceptions about security decided this was an issue, it’s obviously not. Honestly concerning that he was able to easily get a CVE for this and even get articles about it published on some websites.
Lock the pc, if you leave and lock the db, if pc is locked, lid is closed and this is absolute a non-issue.
German BSI is sometimes a little bit over motivated ;-)
You don’t even need to lock the pc, locking the db is sufficient. The issue allows changing the settings on unlocked databases without needing to re-confirm (at least according to the article).
Can’t read German. What is required to perform this attack?
Ok I checked it up (CVE-2023-35866). It basically says an attacker may export everything if they have access to your unlocked database. Which seems… obvious? The project contributors says it’s not a vulnerability which I incline to agree.
You mean to say that if I leave my door unlocked, somebody might come in? This is shocking news!
Do we know mechanism of access. You have to be on same users account or you have to be on same machine only as any user? If same users account, what do they expect? Anyone running as you has total access to your stuff anyway. Is there anyway around that?
Thoughts?
deleted by creator
It’s a denial of service vulnerability. Requiring the existing master password to change the master password will stop a drive by miscreant denying you access to your db. And password change system I’ve ever used has required the existing password to he entered first.
Likewise a full db export feel like a big enough deal to require authorization.
If you’re careful and lock your machine when you leave it then you should be pretty safe. I’m surprised these aren’t already features.
No, requiring the existing master password won’t help. A drive by miscreant with access to an unlocked computer with an unlocked DB can delete all the DB entries. If the DB is locked they can just delete the DB file. KeePassXC can’t defend against this, that takes properly functioning versioned backups.
yeah, maybe denial of service isn’t it. I replied to a comment above why I think it should still be protected functionality to help prevent data leak.
They could just delete the file to deny you access to your db?
Yeah, that’s fair. But a full db export that they could then email themselves. It’d be nice to have some more protection against that. Or Change the master password and email the encrypted file to themselves.
Why is there no link to the article?
There is if you click on the image.
On Jerboa(List View) the link is on the thumbail, maybe it’s the same in the browser version. Keep in mind that the article on heise is in german.
deleted by creator
Attackers can manipulate files or the master password without authentication confirmation.
This is in my view incorrect. It should be: If an attacker gains access to a computer where a user has unlocked the KeePass DB, the attacker has access to all data and can switch off 2FA and change password, the software does not ask to enter the password again.
The key is that the user must unlock the data first. It is a vulnerability, but the way it is classified by BSI is questionable.
