0

This question is about the use of a TPM by a password manager used by an end-user on a PC to store and use passwords to log in to web sites and other things that are protected only by a password.

Windows 11 requires at least TPM 2.0. Support for Windows 10 is scheduled to end in 2025. This suggests that sometime soon password managers might be able to assume the existence of TPMs on Windows workstations.

The technical details of the use of the TPM by password managers seems to be rather vague. The 1password whitepaper contains this gem, for example:

enter image description here

On the other hand, the 1password Windows desktop application does seem to make use of the TPM when it's available:

If you use the Trusted Platform Module with Windows Hello:

  • 1Password delegates the responsibility of authentication to Windows Hello.
  • The encrypted secret is stored in the Trusted Platform Module instead of your computer’s memory.
  • Windows Hello can immediately unlock 1Password after you quit the app or restart your PC.

Unfortunately the above are the clearest statements I have found about any password manager's use of the TPM. My intuition is that a properly-designed Windows password manager application (or browser, browser extension, etc) could mitigate some risks using a TPM that would not otherwise be possible.

  1. What threats could, in principle, be mitigated by a password manager's use of a TPM that could not otherwise be mitigated?
  2. Are there any password managers that includes in their official security model documentation how they use the TPM?

1 Answer 1

1

Brute Force Attack

TPM encrypts the Key Encryption Key (KEK) of the password manager that encrypts its own master key. Password managers ties TPM's decryption of KEK with the system authentication backed by TPM. The benefit of enforcing system authentication for key use is TPM can enforce an exponential cooldown timer after some incorrect attempts. It also allows use of biometric authentication for user convenience. This prevents offline brute force attack to pass authentication. The password manager can also clear encrypted master key and vault altogether after being notified of incorrect attempts by the operating system. This will force the attacker to login again into the account which may not be feasible to crack. This also prevents the attacker to use methods like chip-off to take out the storage and brute force the key on its own machine.

After First Unlock State Vulnerabilities

Operating system can encrypt user and app data from first boot which is backed by TPM. But over the years, spyware agencies have worked around storage encryption to extract data by using exploits for the operating system. TPM protects KEK from system vulnerabilities. Even if the attacker compromises the OS or the password manager, TPM will refuse decryption of KEK unless system authentication is successful while assuming that it is infeasible for the attacker to exploit TPM.

To increase user trust, password manager can provide their own offline password based authentication that encrypts the KEK together with the encryption of KEK by TPM. This will ensure that even if the attacker manages to compromise TPM, the KEK will not be fully decrypted. But opting custom authentication like this comes with the unavailability of biometric authentication.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .