This YARA rule targets specific patterns in PGMPACK v0.14, a known PE file packer frequently used by adversaries to compress executables and obscure their code from static analysis. Proactively hunting for this signature in Azure Sentinel helps identify potentially malicious or obfuscated binaries that may be leveraging this specific packing tool to evade initial detection mechanisms.
rule PGMPACKv014
{
meta:
author="malware-lu"
strings:
$a0 = { 1E 17 50 B4 30 CD 21 3C 02 73 ?? B4 4C CD 21 FC BE [2] BF [2] E8 [2] E8 [2] BB [2] BA [2] 8A C3 8B F3 }
condition:
$a0 at pe.entry_point
}
This YARA rule can be deployed in the following contexts:
This rule contains 1 string patterns in its detection logic.
Scenario: Execution of the PGMPACK utility by the IT Asset Management team during a quarterly software inventory audit. The tool is invoked via a scheduled task named AssetScan_Monthly on domain-joined workstations to compress and package application metadata for the CMDB.
C:\Program Files\AssetManager\Tools\ or C:\Scripts\Inventory\ where the parent process is svchost.exe (for scheduled tasks) or powershell.exe with arguments containing -Inventory or -Package.Scenario: DevOps engineers using PGMPACK as part of a custom CI/CD pipeline (e.g., Jenkins or Azure DevOps) to compress build artifacts before uploading them to an internal artifact repository. The binary is executed within a Docker container or a dedicated build agent VM.
jenkins or buildagent service accounts, or exclude execution paths within C:\Jenkins\workspace\ or C:\AzureDevOps\Agent\.Scenario: Database Administrators (DBAs) using PGMPACK to compress large log files or backup sets prior to transferring them to off-site storage. This is often done via a PowerShell script triggered by a SQL Server Agent job named NightlyBackupCompression.
-compress or -archive, and the working directory is under C:\Backup\ or D:\DBBackups\.Scenario: Endpoint Detection & Response (EDR) agents or antivirus software (e.g., CrowdStrike, SentinelOne) temporarily invoking PGMPACK during a deep scan or