Original title: Beyond the Trust Boundary: A Critical Reassessment of the FIDO2 Threat Model
Aditya Mitra, Kolluru Sai Abhiram, Sibi Chakkaravarthy Sethuraman, Anitha S · September 2026
Many systems use FIDO2 and physical security keys to stop phishing. People often assume this setup is completely safe as long as the math behind the digital keys works. The problem is that logins do not happen in isolation. The researchers re-examined what can actually go wrong in real-world use. They tested eight ways an attacker could strike outside the core math, including infected browsers, malicious software, and rogue USB hardware. Even if the cryptographic keys stay secure, bad actors can break the system if they compromise the web browser, the computer's operating system, or connected hardware devices. Attackers can also chain separate attacks across these layers to bypass intended defenses, or track users with device timing and identity codes. Ultimately, the researchers found that the biggest security danger is not the cryptography itself, but the surrounding software and hardware that users trust.
Imagine you have a safe with an unbreakable lock, but you keep it in a room with a broken door and no alarm system. A thief can simply walk in and take the safe without picking the lock. In this paper, FIDO2 is like the safe with strong cryptography, but attackers target the browser, operating system, or hardware—which are the broken door and missing alarm—to steal authentication without breaking the cryptographic code.
People using FIDO2 for secure logins, like with security keys, might be at risk because attackers can target the device or software instead of the keys themselves. This means that even strong authentication can fail in real-world setups.
The abstract does not specify limitations, but it notes that attacks vary in privilege, skill, and resource requirements, and focuses on eight specific vectors without claiming comprehensive coverage.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Beyond the Trust Boundary: A Critical Reassessment of the FIDO2 Threat Model — Aditya Mitra, Kolluru Sai Abhiram, Sibi Chakkaravarthy Sethuraman et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.