Skip to main content
The HiddenLayer AI Runtime Security is managed using a policy. The policy configuration customizes the LLM proxy’s behavior in various scenarios. For instance, all detection methods offer Ignore, Alert, and Alert and Block policy configurations. Additional container behavior, such as routing to different LLM providers, can be configured via settings during container deployment; in fact, to use Runtime Security in its enriched, forward-proxy mode, the backend LLM must be configured in this way. This guide covers the various behaviors that can be configured using environment variables in the container.

​
Configuration Options

There are seven different policy groups, each with specific environment variables which can be tailored to the specifications and requirements of your organization. By default, the policy will be alert only for all detections.
Configuration SettingsMost configuration settings are true or false, with false being the default setting. For configurations with different settings, the settings are identified in the Description.