AI Agents are gaining stronger runtime protection as Delinea has introduced new runtime authorization capabilities that enforce security policies on actions before they execute. The company says it is the only platform that can enforce policy on what AI agents do inside a session, rather than simply controlling how they connect.

The new capabilities are part of the Delinea identity security platform, which governs what humans, machines, and AI agents can do after they gain access. The launch addresses growing security concerns as AI agents increasingly operate autonomously across production databases, SSH hosts, Kubernetes clusters, cloud consoles, and Model Context Protocol (MCP) servers.

According to Delinea, most existing identity security solutions either verify access when a session begins and audit activity afterward, or they monitor only traffic routed through their own servers. As a result, they cannot enforce security policies on actions taking place within a session before they are executed. Delinea’s runtime authorization is designed to close that gap.

“The security industry spent years solving credential theft, but AI agents have introduced a new problem: authorized access doing unauthorized things,” said Art Gilliland, CEO at Delinea.

“You can have perfect credential hygiene and still have an agent tear through your production environment in milliseconds. Recording what happened or revoking access after the fact doesn’t stop that. Enforcing policy on the action as it runs does. The perimeter has moved to inside the session, and no one has figured out how to address that until now.”

The runtime authorization capabilities provide consistent policy enforcement and a complete audit trail regardless of how AI agents connect. Whether agents access databases, SSH hosts, Kubernetes clusters, cloud consoles, or other environments directly, the Delinea Platform authorizes and records every connection through a single control point across multiple protocols. This gives security teams consistent visibility and enforcement across deployments.

Additionally, the platform removes standing credentials from the agent. Instead, credentials are injected just in time when access is required, limited to the assigned task, and automatically revoked when the task is completed. Since the credentials are never exposed to the agent, the risk of credential theft between sessions is significantly reduced.

Delinea also evaluates every individual tool call made by AI Agents during a session. Rather than relying on session-level authorization, the platform can allow, block, or require human approval for each action before it executes. This approach helps contain the impact of unexpected or unauthorized behavior.

Furthermore, the platform distinguishes between human-driven and agent-driven connections before granting access. It applies policies specifically designed for AI agents at the time of connection, ensuring the appropriate security controls are active before any action begins.

The platform also maintains a detailed audit record of every tool call, database query, and SSH command. Each action is linked to a named identity, providing organizations with complete accountability and supporting investigation and response when needed.

Delinea said the same enterprise-grade platform already used to govern privileged access across thousands of enterprise environments now extends those capabilities to AI agents. Organizations can also apply Zero Standing Privilege without deploying additional infrastructure, helping strengthen security where higher levels of protection are required.

Runtime authorization for AI Agents is available immediately as part of the Delinea Platform. The company said the new capabilities enable organizations to enforce policy before actions are executed, while improving visibility, accountability, and security across autonomous AI environments.

Related post

View all