Skip to content

CHAIN-008 — Defender disabled open ports to blind execution

Summary

Severity: Critical · Likelihood: High · Logic: ANCHOR_PLUS_ONE · Anchor: zt_vis_003

Why this chain matters

Microsoft Defender for Cloud Servers plan is turned off, management ports are open to the internet, and activity logs are not shipped to a SIEM. An attacker brute-forces or exploits the exposed port, executes payloads on the VM, and neither the host-based Defender sensor nor the control-plane audit trail reports anything. The environment becomes a blind spot: compromise happens in full darkness.

Component rules

This chain fires when its trigger conditions are met by the following rules. Click any rule to see its detection logic and compliance mappings.

Rule ID Role
zt_vis_003 Anchor
zt_net_001 Trigger
zt_net_002 Trigger
zt_vis_002 Trigger

Attack walkthrough

Step 1 — Locate a VM exposing SSH/RDP/WinRM to the internet.

Actor: External attacker
MITRE ATT&CK: T1595.001
Enabled by: zt_net_001

Shodan / internet-wide TLS banner scan; NSG permits 0.0.0.0/0 on port 22 or 3389.

Attacker gain: Reachable compromise target.

Step 2 — Exploit or brute-force the exposed service to gain interactive access.

Actor: External attacker
MITRE ATT&CK: T1110.003
Enabled by: zt_net_002

Credential spray against local accounts; or exploitation of unpatched SSH/RDP CVEs.

Attacker gain: Shell on the VM as a local user.

Step 3 — Execute tooling with zero host-level detection.

Actor: Attacker on VM
MITRE ATT&CK: T1562.001
Enabled by: zt_vis_003

Defender for Servers plan is Free/off: no MDE sensor, no file behaviour monitoring, no EDR telemetry generated.

Attacker gain: Unobserved execution of discovery, credential dumping, and persistence tools.

Step 4 — Operate without control-plane telemetry either - Activity Log is not exported to a SIEM.

Actor: Attacker on VM
MITRE ATT&CK: T1562.008
Enabled by: zt_vis_002

No diagnosticSettings streaming to Log Analytics / Event Hub; on-box actions translated into ARM calls are not correlated anywhere.

Attacker gain: Complete blind spot across both host and cloud audit surfaces.

Blast radius

Initial access Internet-exposed management port on an unmonitored VM.
Lateral movement Whatever the compromised VM can reach - and nobody will see it happening.
Max privilege Local admin on the VM; potentially more via managed identity (see CHAIN-001).
Data at risk Everything on the VM and everything reachable from it
Services at risk Compute, Any service the VM can call
Estimated scope Unknown - the absence of telemetry is the finding

How the logic works

The chain fires when the anchor rule fires AND at least one of the other triggers fires. The anchor represents the initial foothold; the second rule amplifies it into a meaningful attack. Remediate the anchor to eliminate the entire chain.