Configure Ai Agent Runtime Protection
In brief
The documentation now notes that Audit mode may incorrectly report a prompt injection block, and Claude Code may continue processing a tool response after a reported block in Block mode.
What Defender admins need to know
Administrators evaluating the preview should interpret these alerts and runtime behaviors with these known issues in mind.
Summaries are generated from the documentation change itself.
Documentation change
The comparison below shows only the changed extract. Use the full-page view for complete context.
Review and investigate detections
Important
During preview, a prompt injection alert might incorrectly report a block when runtime protection is configured in Audit mode. In Block mode, Claude Code might continue processing a tool response despite a reported post-tool-response block. See Known issues (Preview) for details.
After you enable runtime protection, review alerts to validate detection accuracy and tune your configuration before broadening enforcement. This step is critical during the audit phase because it helps you understand what agents are encountering and whether detections represent real threats.
When runtime protection detects prompt injection, Defender raises a Suspicious AI prompt injection alert and takes action based on the configured mode. The alert appears on the device timeline, and related alerts are correlated into incidents for security operations center (SOC) investigation. In block mode, the alert severity is Critical, High, Medium, or Low based on assessed risk. In audit mode, the alert is Informational, so your team can review what would have been blocked without triaging it as an active threat.
@@ -9,8 +9,8 @@ feedback_product_url: https://techcommunity.microsoft.com/t5/security-compliance uhfHeaderId: MSDocsHeader-MicrosoftDefender manager: bagol description: Learn how to configure Microsoft Defender for Endpoint AI agent runtime protection to detect, audit, and block prompt injection on Windows devices.-author: lwainstein-ms.author: lwainstein+author: j0shbregman+ms.author: joshbregman ms.service: defender-endpoint ms.topic: how-to ms.date: 2026-09-25T00:00:00.0000000Z@@ -206,6 +206,10 @@ Close any terminal windows used to run agents, and then open a new terminal wind
## Review and investigate detections
+Important
+
+During preview, a prompt injection alert might incorrectly report a block when runtime protection is configured in Audit mode. In Block mode, Claude Code might continue processing a tool response despite a reported post-tool-response block. See [Known issues (Preview)](ai-agent-runtime-protection-overview#known-issues-preview) for details.
+
After you enable runtime protection, review alerts to validate detection accuracy and tune your configuration before broadening enforcement. This step is critical during the audit phase because it helps you understand what agents are encountering and whether detections represent real threats.
When runtime protection detects prompt injection, Defender raises a **Suspicious AI prompt injection** alert and takes action based on the configured mode. The alert appears on the device timeline, and related alerts are correlated into incidents for security operations center (SOC) investigation. In block mode, the alert severity is **Critical**, **High**, **Medium**, or **Low** based on assessed risk. In audit mode, the alert is **Informational**, so your team can review what would have been blocked without triaging it as an active threat.
