Posted by Singapore: Cybersecurity firm Keeper Security has unveiled a new AI Agent integration capability that allows organisations to safely connect artificial intelligence agents to their secrets management systems. The update introduces the Model Context Protocol (MCP), a secure framework that enables AI-driven automation while maintaining strict control over access and data security.
AI Agent integration
Keeper’s latest enhancement to its Secrets Manager product is designed to address growing enterprise demand for AI integration in workflows. The new protocol allows businesses to authorise AI agents—whether locally hosted or cloud-based—to interact with sensitive systems under tightly controlled conditions, without compromising on the company’s core principles of zero-trust and zero-knowledge architecture.
“AI agents are becoming powerful tools for operational efficiency, but their access to sensitive data must be governed by strong controls,” said Craig Lurey, CTO and Co-Founder of Keeper Security. “With our Model Context Protocol integration, organisations can adopt AI responsibly while ensuring their digital assets remain protected.”
The Model Context Protocol acts as a secure and auditable interface between AI tools and Keeper’s encrypted vaults. It ensures that while AI agents may operate outside the bounds of zero-knowledge from the customer’s perspective, Keeper itself retains no visibility into the data being accessed.
Key Features of the MCP Integration:
Zero-Trust Enforcement: AI agents are granted access only to specific folders, supporting least-privilege access principles.
Human-in-the-Loop Oversight: Sensitive operations require real-time user approval, reducing the risk of unintended data exposure.
Improved Efficiency: Users can automate data retrieval without manual copying, minimising errors and memory leaks.
Comprehensive Auditing: All AI activity is logged and auditable to meet compliance and regulatory standards.
Cross-Platform Compatibility: Available on Linux, macOS, Windows, and Docker environments.
Admin-First Activation: The feature is off by default and must be manually enabled through the Keeper Admin Console.
With the MCP integration, organisations can securely enable AI agents to perform tasks such as generating secure passwords, managing secrets and attachments, and running system diagnostics—all within policy-driven constraints.
“Model Context Protocol provides the secure framework enterprises need to confidently deploy AI agents alongside their secrets management infrastructure,” said Jeremy London, Director of Engineering for AI and Threat Analytics at Keeper Security. “This integration allows us to support automation without compromising on security or oversight.”
The feature is part of Keeper’s broader KeeperPAM platform, which includes tools for password rotation, privilege access management, and secure credential storage. Keeper Secrets Manager is certified under SOC-2, SOC-3, ISO 27001/27017/27018, and is FIPS 140-3 validated, as well as FedRAMP and GovRAMP authorised.