Lumos Introduces MCP Governance for Enhanced AI Agent Security Management

Lumos Launches MCP Governance to Ensure Runtime Security for AI Agents

Lumos, the identity management platform at the forefront of the agentic era, has introduced a revolutionary approach to managing AI agent permissions with the launch of MCP Governance. This innovative tool specifically targets the growing need for enterprises to maintain control and visibility over the actions of AI agents, particularly when leveraging Claude Code and Codex.

The introduction of MCP Governance comes at a crucial time for businesses that are rapidly adopting AI technologies to enhance productivity. As organizations integrate AI coworkers into their workflows, the challenge of ensuring that these agents operate within designated permission parameters becomes increasingly significant. Unlike human actions, which can be monitored and regulated post-factum, AI agents operate at unprecedented speeds, with the capability of executing thousands of tasks within seconds—an issue that can lead to substantial consequences if left unchecked.

Andrej Safundzic, CEO and co-founder of Lumos, emphasized the urgency of tackling this challenge, stating, "We spent twenty years learning to govern humans and still have not finished. Now we have agents doing the same work ten times faster. Just here at Lumos, with fewer than 200 employees, we measured over 450,000 agent actions in a single week. That kind of scale is impossible to track with old methods."

The standard practice in cybersecurity has often included tracking agent inventories, but this merely informs security teams of the existence of agents without providing insights into their permissions or actions. In contrast, MCP Governance shifts the focus to real-time decision-making—asserting control right at the moment an action is proposed by an agent.

As Safundzic noted, organizations are not attempting to restrain AI but are striving to empower it safely. He shared an example of a Lumos client who withheld integration for their marketing team due to excessive access to underlying tools, resulting in lost opportunities. "Governance at the moment of action is how you turn that no into a yes," he explained. The ability to govern actions before they occur is pivotal in adapting to the fast-paced advancements in AI technologies.

This perspective on identity governance was further reinforced by Leo Mehr, co-founder of Lumos, who stated, "Identity has always governed what someone is permitted to do. However, it hasn't provided control over what they actually did, as human actions were manageable with post-event reviews. But with agents making decisions so rapidly, waiting for a review is no longer viable. The only opportunity left to govern is just before the action is executed."

MCP Governance is designed to complement Lumos's existing framework of managing identities across human, machine, and AI platforms. By mapping identities and permissions, the tool offers a comprehensive solution that ensures both visibility and control over how these identities operate.

Available now for teams utilizing Claude Code and Codex, MCP Governance exemplifies Lumos’s commitment to security and efficiency in an ever-evolving digital landscape. The company is poised to expand support to additional agents in the near future, making it easier for organizations to navigate the complexities of AI governance.

For those interested in understanding how MCP Governance can transform their operational security, Lumos invites teams to schedule a demo today.

About Lumos

Lumos is pioneering the first identity management platform tailored specifically for autonomous agents, focusing on real-time governance and security. Unlike traditional identity governance systems, which were designed for slower human workflows and frequent reviews, Lumos addresses the scale and complexity introduced by AI. As a result, companies like Mars, Netskope, Assurant, and GitLab leverage Lumos to accelerate their operations while minimizing risks and ensuring compliance, all under the watchful guidance of human oversight.