New Pass-ta-key attack reveals all the things we didn't know about passkeys

The Pass-ta-key research made for a dramatic headline, but the reality is far less alarming. The attack, as outlined by Arie Olshtein, requires malware already running on a Windows machine. At that point, the game is over regardless of passkeys. The attacker can capture keystrokes, read memory, and hijack sessions. Extracting passkeys from Google Password Manager is just another move in an already lost position.

The confusion stems from a misunderstanding of where passkeys live on Windows. Many assume they are sealed inside the TPM, the hardware enclave designed to protect cryptographic material. That is true for some passkey implementations, but not for Google Password Manager. GPM stores passkeys in its own vault, protected by the operating system's security mechanisms. The TPM is not a universal vault for every passkey on the system. It is a component that can be used, but its use is not automatic.

This distinction matters because it shapes the threat model. Passkeys are designed to resist remote phishing attacks, not to defend against a compromised device. If malware is already executing with user privileges, it can do anything the user can do, including accessing their password manager. No authentication system can fully protect against a fully compromised endpoint. The security boundary is the device itself.

The broader lesson is not that passkeys are broken, but that endpoint security remains the foundation. Passkeys eliminate a class of attacks that rely on stolen passwords and credential reuse. They do not eliminate the need for robust malware protection, patch management, and user awareness. The Pass-ta-key attack is a reminder that no single technology is a silver bullet.

For enterprises evaluating passkey adoption, the risk profile is unchanged. The attack does not introduce a new vulnerability in the protocol or in the TPM. It simply demonstrates that a compromised device is compromised. The decision to move to passkeys should be based on their proven ability to reduce phishing and credential theft, not on a misreading of this research. The noise around Pass-ta-key will fade, but the underlying security principles remain constant.