Detections

Abuse of cloud messaging platforms to send mass spam or consume quota-based resources.

Detects adversarial archiving using libraries (zlib, zip APIs) invoked by scripts or binaries. Correlates process executions of Python, PowerShell, or custom .NET binaries with DLL/module loads linked to compression libraries, followed by archive file creation.

Detects adversarial archiving by scripts or binaries calling compression libraries (libzip, zlib, bzip2). Correlates execution of Python, Perl, or compiled binaries with dynamic linking to archiving libraries and creation of compressed files in /tmp or user directories.

Detects malicious archiving via system or third-party libraries (libz, libarchive) invoked by Python, Swift, or Objective-C binaries. Correlates unified logs of library loads with creation of compressed or encrypted archives (.zip, .gz, .bz2, .dmg).

Logon via RDP or WMI by a user account followed by uncommon command execution, file manipulation, or lateral network connections.

SSH session from new source IP followed by interactive shell or privilege escalation (e.g., sudo, su) and outbound lateral connection.

Remote login via ARD or SSH followed by screensharingd process activity or modification of TCC-protected files.

Use of cloud-based bastion or VM console session followed by commands that initiate outbound SSH or RDP sessions from the cloud instance to other environments.

vSphere API logins (vimService) or SSH to ESXi host followed by unauthorized shell commands or lateral remote logins from the ESXi host.

Adversary modifies Group Policy Objects (GPOs), domain trust, or directory service objects via GUI, CLI, or programmatic APIs. Behavior includes creation/modification of GPOs, delegation permissions, trust objects, or rogue domain controller registration.

Adversary modifies tenant policy through changes to federation configuration, trust settings, or identity provider additions in Microsoft 365/AzureAD via Portal, PowerShell, or Graph API. Includes setting authentication to federated or updating federated domains.

Detects anomalous process access to LSASS on domain controllers, suspicious module loads of authentication DLLs, and registry or file modifications indicative of Skeleton Key–style patching. Correlates LSASS access attempts with subsequent abnormal logon activity patterns.

Detects unauthorized modification of network device authentication by correlating OS image file changes, checksum mismatches, or memory verification failures with anomalous authentication events. Focus is on behaviors where patched images introduce hardcoded passwords or bypass native authentication.

Processes that normally do not initiate network connections establishing outbound encrypted TLS/SSL sessions, especially with asymmetric traffic volumes (client sending more than receiving) or non-standard certificate chains. Defender observations correlate process creation with unexpected network encryption libraries being loaded.

Processes like curl, wget, python, socat, or custom binaries initiating TLS/SSL sessions to non-standard destinations. Defender sees abnormal syscalls for connect(), loading of libssl libraries, and persistent outbound encrypted traffic from daemons not normally communicating externally.

Applications or launchd jobs initiating encrypted TLS traffic to rare external hosts. Defender observes unified logs showing ssl/TLS API calls by processes not baseline-approved, and payload entropy suggesting encrypted C2 sessions.

VMware management daemons or guest processes initiating encrypted connections outside expected vCenter, update servers, or internal comms. Defender identifies hostd or vpxa initiating outbound TLS flows with uncommon destinations.

Unusual TLS tunnels through ports not normally encrypted (e.g., TLS on port 8080, 53). Defender sees NetFlow/IPFIX or packet inspection indicating high-entropy traffic volumes and asymmetric client/server exchange ratios.

Correlation of registry key modification for Run/RunOnce with abnormal parent-child process relationships and outlier execution at user logon or system startup

Correlates creation/modification of systemd service files or /etc/init.d scripts with outlier process behavior during boot

Observes creation or modification of LaunchAgent/LaunchDaemon property list files combined with anomalous plist payload execution after user logon

An adversary leverages built-in tools such as certutil.exe, powershell.exe, or copy.exe to decode, reassemble, or extract hidden malicious content from obfuscated containers or encoded formats. The decoding utility often spawns shortly after file staging or download and may be chained with script interpreters or further payload execution.

The adversary uses native utilities like base64, gzip, tar, or openssl to decode, decompress, or decrypt files that were previously staged or downloaded. These tools may be chained with curl/wget and executed via bash/zsh, often to extract an embedded payload or reverse shell script.

The adversary invokes built-in scripting or decoding tools like base64, plutil, or AppleScript-based utilities to decode files embedded in staging artifacts. Decoding often occurs post-download or as part of post-exploitation payload deployment via zsh, python, or osascript.

Detection of rogue Domain Controller registration and Active Directory replication abuse by correlating: (1) creation/modification of nTDSDSA and server objects in the Configuration partition, (2) unexpected usage of Directory Replication Service SPNs (GC/ or E3514235-4B06-11D1-AB04-00C04FC2DCD2), (3) replication RPC calls (DrsAddEntry, DrsReplicaAdd, GetNCChanges) originating from non-DC hosts, and (4) Kerberos authentication by non-DC machines using DRS-related SPNs. These events in combination, especially from hosts outside the Domain Controllers OU, may indicate DCShadow or rogue DC activity.