Keeper Security, a leading provider of zero-trust and zero-knowledge identity security and Privileged Access Management (PAM) software, announced the integration of agentic AI governance into its Keeper Endpoint Privilege Manager. The newly rolled-out capabilities bring automated policy enforcement to autonomous artificial intelligence systems, establishing a centralized security layer for both human and non-human identities directly at the device and endpoint level.
The software update addresses a rapidly evolving threat vector within enterprise network architecture. As coding assistants and autonomous AI applications proliferate across workforce workstations, they routinely execute tasks, make system calls, and seek administrative clearance outside the visibility of legacy privilege management software. Keeper Endpoint Privilege Manager mitigates this visibility gap by intercepting and governing agent activity in real time, regardless of whether the system interacts through a direct API, Model Context Protocol (MCP), or local automation scripts.
“AI agents are not assistants; they are principals,” said Darren Guccione, CEO and Co-founder of Keeper Security. “Every agent running on an endpoint has an identity, requests access and takes actions on behalf of your organization. If you are not governing them with the same rigor you apply to your human workforce, you have blind spots that adversaries will find before you do. Keeper closes that gap today.”
Also Read: SEALSQ Corp Advances Post-Quantum Cybersecurity Infrastructure
Enforcing Local Control via OS-Level Process Observation
Instead of routing security evaluations through external network proxies, Keeper handles policy management from the existing endpoint software agent tasked with regulating human privilege requests. Operating natively at the operating system level, Keeper monitors the exact machine environment where the artificial intelligence executes to manage five specific local vectors:
-
Child Process Spawning: Prevents unauthorized background executables or cascading sub-routines from initiating without administrative consent.
-
Direct Filesystem Adjustments: Regulates the writing, modification, or deletion of sensitive information blocks across restricted system volumes.
-
Local Shell Invocations: Blocks autonomous systems from opening terminal prompts or executing unverified, low-level operational commands.
-
Privilege Elevation Paths: Restricts AI applications from modifying local operating system permissions or obtaining persistent root access.
-
Sensitive File Ingestion: Safeguards corporate intellectual property by controlling which database paths or directories an agent can read.
Because every action undergoes evaluation against a unified policy engine, the platform produces a single, cohesive audit trail that matches the compliance rigor applied to human operators.
Combining Machine Learning Diagnostics with Human Review
To simplify administrative overhead, the platform calculates a real-time risk index ranging from 0 to 100 for every application active on a managed workstation. Any system crossing a preconfigured compliance threshold automatically triggers agentic AI protective policies at runtime, eliminating the need for manual file tagging or signature updates.
“This update stops the emerging threat of autonomous AI in its tracks, allowing enterprises to adopt cutting-edge AI agents without opening the floodgates to catastrophic risk.”
The agentic governance matrix introduces three dedicated policy models built on existing Endpoint Privilege Manager configurations, alongside a human-in-the-loop review interface:
-
Agentic AI Policy: Manages user permissions to determine which specific personnel are authorized to launch automated agents on local hardware.
-
Agentic Access Policy: Controls what data, executables, and background terminal commands a running agent can access on a user’s behalf.
-
Agentic Privilege Elevation Policy: Regulates the mechanism by which an active agent can request temporary or permanent administrative clearance from the OS.
The expanded toolkit provides the necessary controls to help enterprises satisfy the National Institute of Standards and Technology (NIST) AI Risk Management Framework directly at the workstation layer. The release also debuts a redesigned management console complete with automated agent version control, unique workload monitoring screens, and dedicated visibility dashboards.
The agentic security features are globally available today, offered as a standalone endpoint solution or fully integrated into the cloud-native KeeperPAM® platform. Corporate security directors, risk compliance managers, and enterprise cloud architects can examine deployment blueprints, review identity verification frameworks, and analyze cryptographic structures by visiting Keeper’s official digital platform.