Andrii Bidochko

• Updated: February 22, 2026
• 6 min read

Study Reveals Critical Flaws in Popular Cloud‑Based Password Managers

The recent ETH Zurich study proves that three of the most popular cloud‑based password managers—Bitwarden, LastPass, and Dashlane—contain critical security flaws that let attackers read and even modify stored passwords, challenging the industry’s “zero‑knowledge” promises.

Why the ETH Zurich Study Matters

The study, published by the Applied Cryptography Group at ETH Zurich, is the first comprehensive, peer‑reviewed analysis that demonstrates real‑world attacks against the “zero‑knowledge” encryption model claimed by leading password‑manager providers. With over 60 million users worldwide, the findings have immediate implications for security‑conscious professionals, IT managers, and anyone who relies on a password manager for secure password storage.

Read the original research article for full technical details: ETH Zurich password manager security study.

Vulnerabilities Uncovered in Bitwarden, LastPass, and Dashlane

The researchers executed a total of 25 distinct attacks:

• 12 attacks on Bitwarden
• 7 attacks on LastPass
• 6 attacks on Dashlane

All attacks succeeded in one of two ways:

1. Integrity violations: targeted user vaults were altered without the user’s knowledge.
2. Full vault compromise: the attacker obtained every stored password for an organization.

Crucially, the attackers needed only to impersonate a legitimate server—a scenario that can arise after a server breach or a DNS hijack. No exotic hardware or massive compute clusters were required.

Bitwarden: 12 Attack Vectors

Bitwarden’s client‑side encryption was found to be bypassable when the server returned malformed synchronization payloads. The researchers demonstrated that a malicious server could:

• Inject crafted ciphertext that, when decrypted, revealed the master key.
• Force the client to re‑encrypt data with a key known to the attacker.

LastPass: 7 Attack Vectors

LastPass relied on a legacy key‑derivation function that allowed a malicious server to perform a downgrade attack, effectively weakening the encryption strength. The team also exploited a race condition in the vault‑sync process to overwrite passwords.

Dashlane: 6 Attack Vectors

Dashlane’s “zero‑knowledge” claim was undermined by an insecure implementation of the password‑recovery feature. By sending a specially crafted recovery request, the attacker could retrieve the master password hash and perform offline cracking.

Methodology: From Malicious Server to Real‑World Exploit

The researchers adopted a malicious server threat model. They set up rogue servers that mimicked the official endpoints of each password manager. When a user’s browser or mobile app connected, the rogue server responded with altered data, exploiting client‑side assumptions about server honesty.

“All we needed were interactions that users already perform—login, vault opening, and sync. The attacks required no zero‑day exploits, only clever protocol manipulation.” – Prof. Kenneth Paterson, ETH Zurich

This approach mirrors realistic attack scenarios such as:

• Compromised DNS records directing traffic to attacker‑controlled servers.
• Insider threats where a rogue employee gains server access.
• Supply‑chain attacks that replace legitimate binaries with malicious versions.

Code Complexity: The Hidden Attack Surface

Matteo Scarlata, a PhD student on the project, highlighted that the rapid addition of “user‑friendly” features—password recovery, family sharing, and cross‑device sync—has inflated the codebase of each manager. This complexity creates obscure code paths that are difficult to audit, increasing the likelihood of subtle bugs.

Key observations:

• Feature creep leads to duplicated cryptographic logic.
• Inconsistent error handling opens side‑channel opportunities.
• Legacy cryptographic primitives from the 1990s remain in production code.

Researchers’ Recommendations

To restore confidence in cloud password managers, the ETH Zurich team proposes a three‑pronged remediation plan:

1. Adopt modern cryptographic standards: Replace outdated key‑derivation functions with Argon2id or scrypt, and enforce TLS 1.3 with perfect forward secrecy.
2. Implement transparent, third‑party audits: Publish audit reports and enable independent verification of the zero‑knowledge claim.
3. Offer opt‑in migration paths: Allow existing users to move to a hardened architecture without losing access to their vaults.

Practical Advice for Users

While providers work on fixes, users can take immediate steps to protect their credentials:

• Prefer open‑source clients: Verify the code yourself or rely on community‑audited builds.
• Enable two‑factor authentication (2FA): Even if a vault is compromised, 2FA adds a barrier to unauthorized access.
• Use a master password with high entropy: Combine length (≥16 characters) with a mix of symbols, numbers, and unrelated words.
• Consider offline vaults for high‑value accounts: Store critical passwords in a locally encrypted file that never syncs to the cloud.
• Regularly review security updates: Subscribe to vendor security bulletins and apply patches promptly.

How UBOS Helps You Stay Secure

At UBOS homepage, we understand that security is non‑negotiable. Our UBOS platform overview offers a zero‑knowledge architecture built on modern cryptography, ensuring that even our own engineers cannot read your data.

For teams looking to integrate secure password management into custom workflows, the Workflow automation studio lets you create automated vault‑access policies without exposing master keys.

Start quickly with pre‑built solutions from our UBOS templates for quick start. For example, the AI SEO Analyzer template demonstrates how to embed secure API keys using our encrypted storage.

SMBs can benefit from the UBOS solutions for SMBs, which include role‑based access controls and audit logs that satisfy compliance requirements.

Startups looking for a scalable security foundation can explore UBOS for startups, which offers a free tier and easy migration paths.

Enterprise customers can leverage the Enterprise AI platform by UBOS to combine AI‑driven threat detection with encrypted credential storage.

Additional Resources from UBOS

Deepen your knowledge with these related articles and tools:

• AI marketing agents – Learn how AI can automate secure content generation.
• UBOS partner program – Partner with us to deliver secure solutions.
• UBOS pricing plans – Find a plan that fits your security budget.
• UBOS portfolio examples – See real‑world deployments of our secure platform.
• Web app editor on UBOS – Build custom secure web apps without writing backend code.
• AI Article Copywriter – Generate SEO‑friendly content while keeping API keys safe.
• AI Email Marketing – Securely manage email campaign credentials.

Conclusion: Act Now, Secure Your Digital Identity

The ETH Zurich study is a wake‑up call: even industry‑leading password managers can fall short of their zero‑knowledge promises. By understanding the password manager security landscape, adopting the researchers’ recommendations, and choosing a platform that prioritizes transparent, modern cryptography—such as UBOS—you can safeguard your credentials against emerging threats.

Don’t wait for the next breach. Review your current password‑manager settings today, enable multi‑factor authentication, and consider migrating to a solution that offers verifiable zero‑knowledge encryption.

Ready to protect your passwords with a truly secure platform? Explore UBOS now and take the first step toward uncompromised digital security.

Andrii Bidochko

CTO UBOS

Andrii Bidochko is an AI entrepreneur and researcher focused on AI agents, reinforcement learning, and autonomous systems. He writes about the technologies shaping the future of machine intelligence, from frontier models and agent architectures to real-world AI applications.