HYPR launches AgentPass to link AI agents with human supervisors

HYPR launched AgentPass on November 18, saying the platform is available to connect AI agents with named human supervisors and enforce access policies before actions run, according to its announcement.
Supervisors sign in through their organization’s identity provider. Agents begin without access and receive permissions based on the supervisor’s role. HYPR says credentials remain in a central gateway, which issues short-lived tokens limited to individual tools.
For supported coding tools on enrolled macOS and Windows machines, AgentPass routes selected traffic through that gateway. Administrators can observe policies before enforcing them and require approval for sensitive operations.
High-risk actions and new agent connections can be sent to a supervisor’s separate mobile device for passkey approval. Disabling an agent blocks its next request, rather than reversing work it already completed.
These are vendor-described controls, not proof that all agent risks are eliminated. The launch addresses an operational gap as agents move from generating suggestions to using enterprise tools: identifying who authorized their work and checking permissions at execution time.
Sources and citations
Revision history
Published .