Overview
Agent On-Queue for Defender XDR allows ContraForce Security Delivery Agents to automatically detect and respond to incidents originating directly from Microsoft Defender XDR with no Microsoft Sentinel forwarding required. Previously, agents could only auto-trigger on incidents ingested through Sentinel. Customers using Defender XDR directly had to manually trigger agent actions. With this update, ContraForce polls Defender XDR for new incidents approximately every 2 minutes, automatically triggering your agent to triage and respond in order to deliver the autonomous SOC experience with minimal time-to-respond.How It Works
ContraForce continuously polls the Microsoft Defender XDR security API for new incidents across all eligible customer accounts. When a new incident is detected:- The platform identifies the incident and checks it against previously processed incidents to prevent duplicates.
- A notification is queued for processing.
- ContraForce validates that the account has a deployed agent with the appropriate severity capability for the incident.
- The agent is automatically triggered on-queue to investigate and respond — just as it would for Sentinel-ingested incidents.
Prerequisites
Before Agent On-Queue for Defender XDR can activate for a workspace, all of the following must be true:Defender XDR Module Enabled
The Defender XDR security provider must be enabled for the workspace. See Defender XDR Module Deployment for setup instructions.
Partner Consent Granted
Microsoft Defender partner consent must be granted during onboarding. This is the same consent flow completed when deploying the XDR module.
Sentinel Forwarding Disabled
The workspace must not be configured to forward Defender XDR incidents to Sentinel. If
FetchToSentinel is enabled, incidents are ingested through the existing Sentinel pipeline instead.Agent Deployed with ProcessIncident Capability
A Security Delivery Agent must be deployed via Agent Center with the ProcessIncident capability enabled for the target incident severity levels (High, Medium, Low, Informational).
What Changes for You
If you use Defender XDR without Sentinel
This is the feature for you. Once the prerequisites are met, your agent will begin automatically processing Defender XDR incidents within approximately 2 minutes of their creation. No configuration changes are needed on your end — the platform handles everything.If you already forward to Sentinel
Nothing changes. Your incidents continue to flow through the Sentinel ingestion pipeline as before. The Defender XDR polling pipeline automatically excludes accounts with Sentinel forwarding enabled.If you use both
Accounts are evaluated individually. Workspaces with Sentinel forwarding enabled use the Sentinel pipeline. Workspaces without Sentinel forwarding use the new Defender XDR polling pipeline. There is no overlap or duplicate processing.Configuring Your Agent for Defender XDR Incidents
If you already have a Security Delivery Agent deployed and configured, no additional setup is required. The platform automatically detects eligible workspaces and begins polling. To deploy or configure an agent:- Navigate to Agent Center from the left navigation menu.
- Deploy your agent following the Agent Center Deployment guide.
- Configure the agent’s ProcessIncident capability and select which severity levels the agent should handle automatically.
- Set the agent mode to On Queue.
Verifying It’s Working
After setup, confirm that the pipeline is active:- Check Agent Center — Verify your agent status shows On Queue and the mode is set to your preferred level (Manual, Automatic, or Autonomous).
- Monitor the Command Dashboard — New Defender XDR incidents should appear on the Command Dashboard within approximately 2 minutes of creation in Defender.
- Review Gamebook Activity — When the agent processes an incident, you’ll see corresponding Gamebook activity in the Gamebook Activity widgets and the incident’s Workbench.
- Check Agent Execution History — Navigate to Agent Center to review the agent’s execution history and confirm incidents are being processed.
Incident Detection Timing
ContraForce polls Defender XDR approximately every 2 minutes. This means:- New incidents are typically detected within 2 minutes of appearing in Defender XDR.
- The agent is triggered immediately after detection and validation.
- End-to-end time from incident creation to agent response initiation is typically under 5 minutes.
Troubleshooting
| Issue | Likely Cause | Resolution |
|---|---|---|
| Agent not processing Defender XDR incidents | Agent not set to “On Queue” mode | Navigate to Agent Center and set the agent mode to On Queue |
| Incidents not appearing from Defender XDR | Partner consent not granted | Re-run the consent flow from Workspace Settings → Module Configuration with Global Admin credentials |
| Duplicate incidents appearing | Sentinel forwarding is also enabled | Verify the workspace is not forwarding Defender XDR incidents to Sentinel. Only one ingestion path should be active |
| Agent triggers for wrong severity levels | ProcessIncident capability misconfigured | Review the agent’s severity capability settings in Agent Center and adjust which severity levels are handled |
| No incidents detected despite active Defender XDR incidents | Account eligibility not met | Verify all prerequisites are satisfied — subscription status, module enablement, partner consent, and agent deployment |
| Agent was working but stopped processing | Subscription lapsed or consent revoked | Confirm subscription is still Active/Trial and re-authorize partner consent if needed |
- Workspace name
- Agent status screenshot from Agent Center
- Approximate timestamp of the incident that was not processed
- Any error messages visible in the portal