On Nov. 3, 2026, current and former mSecure customers opened an email that told them to protect everything in their vaults and almost nothing about why. The sender was mSeven Software, the company behind the password manager. The subject line was an important security notice. The body used the phrase “suspicious activity,” named no systems, gave no timeline and offered a menu of resets: passwords, two-factor methods, passkeys. For an IT team triaging a password manager incident, that is a difficult starting point. The advice is maximal. The facts are minimal. That gap is the real story of the mSecure password manager security incident, and it is why operators should separate what customer-shared replies attribute to mSeven from what remains unverified.
What the mSecure Password Manager Security Incident Notice Actually Said
Customers pasted the Nov. 3 email into mSecure’s own support forum the same day. In a thread titled Important Security Notice Nov 3, 2026, the quoted text reads, in part: “mSeven Software (“mSeven”) recently learned of suspicious activity associated with the mSecure password management application.” It says mSeven launched an investigation, took mitigation steps and hired outside cybersecurity help. Recipients were told they were current or former customers. Then came the reset menu: “This may include changing passwords, resetting or re-enrolling two-factor authentication methods, or removing existing passkeys.”
That is the entire public substance of the first notice, as customers reproduced it. No count of affected accounts. No list of systems. No statement that vault plaintext was read. No statement that it was not. The company also did not, in that email, say whether Dropbox sync, iCloud, Wi-Fi sync or local-only use mattered. Forum posters asked those questions immediately. For a while, the only answers were more customer posts.
What Support Later Told Customers
A clearer paragraph showed up the next day, not as a homepage advisory, but as a paste from support tickets. In the forum thread Update from mSecure Team, a customer named Caitlin published the reply she said she received. According to that reply:
- mSeven “learned that an unauthorized actor acquired an encrypted copy of a database containing information belonging to current and former users” of mSecure.
- “mSeven has no evidence that the unauthorized actor has accessed the confidential information stored within the database in the mSecure Cloud.”
- The company still recommends changing the master password, changing “vital stored passwords,” resetting two-factor methods and replacing passkeys “as a standard best practice.”
- Deleting the app “will have no impact on this situation.”
- “This issue only pertains to users that were synching with the mSecure Cloud.”
- mSeven says it has “no evidence that mSecure is not currently safe for use,” and that it has taken steps “to bolster security.”
Read that carefully. The customer-shared reply attributes the encrypted-copy claim to mSeven; it is not an independently authenticated public advisory. Decrypted vault contents are not. “No evidence of access” is not the same sentence as “we proved the ciphertext cannot be read.” The company also still has not, in that reply, explained how the copy was taken, what fields sat in the database, whether any keys or verifiers left with it, or how many people got the email.
What mSecure Says Its Architecture Is Supposed to Do
mSeven’s published security model is the reason the first email felt so jarring. In mSecure’s Security Model – Secure by design, the company says data is encrypted before it leaves the device, that it uses AES-256, that a locally generated 46-character Account Key “is never stored on our servers,” and that “even we, the creators, have no way to access your information.” It calls that a zero-knowledge model.
If that model held as described, stealing an encrypted cloud copy should not equal reading every login. That is the architecture’s pitch. It is also why customers asked, reasonably, why the first notice told every current and former user to reset the world. The later Cloud-only caveat narrows the blast radius on paper. It does not answer whether anyone who briefly tried Cloud sync and later switched still has a copy sitting in that database. mSeven’s reply does not say.
How I Would Triage This, as an Operator
The following is my opinion, not a vendor playbook.
- Treat Cloud sync as in-scope until mSeven publishes otherwise. If you never used mSecure Cloud, the support reply says you are outside this issue. If you did, act as if the encrypted copy exists.
- Do not wait for a prettier advisory to rotate the crown jewels. Email accounts, identity providers, banking, domain registrars and anything that can reset everything else come first. Labels like “vital stored passwords” in mSeven’s reply are doing that triage for you without saying the quiet part.
- Reset authenticator secrets and passkeys that lived in the vault. The first notice called those out specifically. If mSecure held the TOTP seeds, rotating the site-side 2FA is the point, not just changing the mSecure login.
- Change the mSecure master password from a device you trust, then decide whether to stay. Deleting the app, according to the customer-shared reply, does not resolve the incident. Changing the master password also does not retroactively protect an encrypted copy already acquired by an attacker.
- Watch for follow-on phishing. Hypothetical: a week from now you get a polished “mSecure breach help desk” message with a password-reset link. That is a standard second act after vague vendor mail. Type the URL yourself.
I am not saying vault plaintext has been proven stolen. The public record, built from the Nov. 3 email and the later support paste, does not support that claim. I am saying a password manager that emails “reset everything” before it can say “encrypted cloud database copy, Cloud sync only, no evidence of confidential access” has already spent a lot of trust. Operators should spend the next change window on the accounts that matter, document which sync method each vault used, and keep the distinction between ciphertext taken and secrets read in every status update they write for their own users.
