Claude can now enter all your passwords for you – if you give him permission

1Password Unveils Claude Partnership, Enabling Anthropic’s AI to Authenticate on Users’ Behalf via Zero Exposure Architecture
Users approve every login with biometrics
New agent mode in browser extension locks interface when AI agents take over

Leading password management company 1Password has launched a new tool that allows artificial intelligence assistant Claude to authenticate on behalf of its user and thus accomplish tasks that were previously impossible without major security compromises.

1Password for Claude is built on a “zero exposure architecture” – so in practice this means that Claude can essentially ask 1Password to complete the login process, but he will never see the credentials and they will never be loaded into his memory.

In turn, 1Password will notify the user and request biometric approval before continuing. Once granted, it will automatically fill in the credentials and check whether they have been exposed on the page or not. If the submission fails, the populated values will be cleared and the report will be resent.

Agent mode

“We need a new security model built specifically for agents, not just humans,” said Nancy Wang, CTO of 1Password. “The answer isn’t to hand over your secrets to agents. It’s about allowing a user to give an agent permission to use an ID without the agent seeing it. Claude knows he used your ID; he doesn’t need the password or one-time code in context. This distinction is the starting point for trust in agents and the foundation we’re building with Anthropic.”

To further strengthen its security posture, 1Password also announced Agentic Mode, a new feature of the browser extension that gives users visibility and control over browser-based AI agents. When a compatible AI agent takes over, the 1Password extension automatically locks and hides the interface. The agent can only use connections and OTPs explicitly approved for the current task.

Even if the integration is not set up, and even if 1Password is not required for the current agent task, agent mode works, the company emphasized. Other agents, besides Claude, are also supported.

Currently, a major debate is underway over how many permissions AI agents should receive and under what rules. We’ve already seen horror stories where AI agents deleted people’s entire email inboxes or somehow ruined days of hard work. It remains to be seen whether this trend will continue or whether most people will remain skeptical about AI access to certain services.