Security researchers have disclosed three new attack techniques that could allow malware to access passkey-protected accounts on Windows PCs without requiring a fingerprint, PIN, or any user interaction. The attacks specifically target Google Password Manager’s cloud authenticator in Chrome and show how malware running on an already compromised device can abuse the passkey implementation.
The research comes from Unit 42, the threat intelligence team at Palo Alto Networks. The researchers have named the three attack techniques Pass-ta-key, Silver Pass-ta-key, and Golden Pass-ta-key.
It is important to note that these attacks do not break passkey encryption or the underlying cryptography. Instead, they exploit the way Chrome manages and protects synced passkeys on Windows.
The attacks only work if malware is already running on the victim’s Windows computer. They are post-compromise techniques, meaning they do not provide an initial way to infect a system.
The research is also limited to Google Password Manager in Chrome on Windows systems that use a Trusted Platform Module (TPM). There is currently no evidence that these techniques are being actively exploited in the wild, and no CVE identifiers have been assigned.
According to the researchers, each technique targets a different part of Chrome’s passkey implementation.
The first attack, called Pass-ta-key, abuses Chrome’s TPM-protected device identity key. Malware can trick the TPM into signing authentication requests without asking the user for biometric verification or a PIN. The authentication response does not contain the User Verified (UV) flag, which normally indicates that the user has completed biometric or PIN verification.
This means websites that properly validate the UV flag remain protected. During testing, GitHub rejected such authentication attempts, while eBay reportedly accepted them before fixing the issue after responsible disclosure.
The second technique, Silver Pass-ta-key, targets Chrome’s device re-enrollment process. Researchers found that malware can force Chrome to register a new user verification key controlled by the attacker before Chrome creates its own secure key.
If successful, future authentication requests can appear fully verified and include the required UV flag, allowing attackers to sign in from another device without needing the victim’s computer.
The third attack, Golden Pass-ta-key, is considered the most powerful. It targets the Security Domain Secret (SDS), a 32-byte master secret that protects synced passkey private keys. Researchers claim malware can temporarily read this secret from Chrome’s memory during the re-enrollment process and later use it to decrypt synced passkeys.
Unlike the first technique, the last two attacks could allow attackers to retain access even after they no longer control the victim’s computer.
The researchers noted that Google has already addressed one earlier issue involving the Security Domain Secret appearing in Chrome’s FIDO logs. They also confirmed that eBay now properly validates the User Verified flag.
However, it remains unclear whether all three attack paths have been fully fixed. Public Chrome documentation does not currently explain whether changing a Google Password Manager PIN or deleting stored Password Manager data invalidates a previously stolen Security Domain Secret.
As of August 3, 2026, security databases also did not list any CVEs related to the three newly disclosed techniques.
Although the findings sound alarming, they do not mean passkeys are broken. Every attack begins with malware already running on the victim’s computer. If an attacker has reached that stage, they may already have access to sensitive files, browser sessions, cookies, and other credentials. The newly disclosed techniques simply show that passkeys are not immune to post-compromise attacks if malware gains control of the device.
For users, the best protection remains preventing malware infections by keeping Windows and Chrome updated, avoiding untrusted downloads, using reliable security software, and enabling additional account security measures wherever possible.
This research shows an important point about passkeys that is often misunderstood. Passkeys are designed to replace passwords and defend against phishing attacks, credential theft, and database breaches. They were never meant to protect accounts after a device has already been fully compromised by malware.
The study also shifts some responsibility to websites. Properly enforcing the User Verified (UV) check prevented the first attack on services like GitHub, while weaker implementations left other platforms exposed until fixes were applied.
Another takeaway is that secure authentication depends on the entire ecosystem, not just cryptography. Even if encryption remains mathematically secure, weaknesses in key management, device recovery, memory handling, or validation logic can still create opportunities for attackers.
At this stage, the research should be viewed as an important warning for browser vendors and relying parties to strengthen passkey implementations rather than evidence that passkeys themselves have failed.
Related Posts
