Security Model of Unlocked 1Password Desktop App in Offline Mode (Passkey-Specific Attack)

Forum|Forum|7 days ago
July 8, 2026
1 reply
8 views

I would like to better understand the architectural security boundaries of 1Password under the following threat model, which focuses exclusively on passkey operations (not on reading or stealing standard passwords):

OS State: A Windows PC is powered on and unlocked. There is no physical surveillance or bystander monitoring of the device (the owner has stepped away).
Attacker Capabilities: An attacker has gained full remote control over this PC. However, they do not know the Windows Hello PIN or biometrics.
1Password State: The 1Password Desktop app is already in an unlocked state.
Account State: The owner had already changed the master password of the account from another device in advance. As a result, the 1Password Desktop app has lost synchronization with the 1Password.com servers and is operating strictly offline.
Access Channel Context: Normally, the official browser extension allows using passkeys via an already unlocked Desktop app without prompting for Windows Hello or a PIN. However, in this model, the extension (and any other client sessions) was never previously authorized on this PC. Since the Desktop app is now strictly offline, authorizing any new sessions through it is technically impossible.
System Integration: Attempting to use passkeys via native Windows WebAuthn will not work either, as this path strictly requires entering the Windows Hello PIN, which is unknown to the attacker.

Consequently, the only active component left on the machine is the unlocked 1Password Desktop app itself.

Question: Does the 1Password security model anticipate and defend against this specific scenario?

To put it plainly: can a "god-mode" hacker — utilizing reverse engineering, debugging, or process hooking directly on this machine — bypass standard channels and force the unlocked 1Password Desktop app to sign an incoming passkey challenge via an unconventional path?

Or is the 1Password Desktop app designed in such a way that, in the absolute absence of legitimate request channels (like an authorized extension), it is fundamentally isolated from such signing attempts, allowing us to rest easy?

Pinned Reply By 1P_Dave

Hello @auriOrient ! 👋

Thank you for the thoughtful questions! This gets at a real tension that every password manager faces. The short answer is that if an attacker has gained full control over a PC where 1Password is unlocked, 1Password is limited in what it can do to protect your data.

Our Security Design White Paper is direct about this. Under "Beware of the Leopard," we write:

Malware that can inspect the memory of your computer or device when the 1Password client has unlocked your data will be able to extract secrets from the 1Password process. Malware that can install a malicious version of 1Password and have you execute it will also be able to extract your secrets. After malware running on a system has gained sufficient power, there’s no way in principle to protect other processes running on that system.

That said, passkeys do have meaningful protections that set them apart from regular passwords. Unlike passwords, they aren't surfaced or exported through standard workflows. The private key isn't meant to be something a user sees or handles directly. Beyond that, in your specific scenario where the browser extension isn’t being used, sign-in attempts using a passkey on Windows prompt for Windows Hello, which requires biometrics or your Windows Hello PIN.

Your data is most protected against local attackers when 1Password is locked, where vault contents remain encrypted even against privileged attackers. You can read more about what 1Password can and can’t protect you against here: How 1Password protects information on your devices (and when it can’t)

-Dave

1P_Dave

1Password Employee