AI-Generated Carbon Black Detection Rule for DarkCrystal RAT Campaign

How It Works

Uncoder AI processes threat reports like CERT-UA#14045 on DarkCrystal RAT and generates Carbon Black-compatible detection logic. This feature maps observed file hashes, execution patterns, and C2 infrastructure into a rule that’s ready to deploy within Carbon Black’s behavioral telemetry stack.

On the left, the threat report details the DarkCrystal campaign, including:

• Malicious file names (Fine Reader.exe, pdfDecode.exe)
  
• SHA256 hashes for droppers and RAT payloads
  
• Embedded infrastructure (e.g., 45.130.214.237, 62.60.235.190)
  
• Known URL paths with embedded download scripts and traffic exfiltration components
  

Uncoder AI transforms these disparate IOCs into a tightly scoped query using Carbon Black’s event model.

Detection Logic Breakdown

process_name:("Fine Reader.exe" OR "DecodPDF.exe" OR "PDF-Office2025version46.v.exe" OR "pdfDecode.exe")
AND (
  md5:("e8720e86ca8937...") OR
  md5:("f8e86a48a9b06a...") OR
  md5:("951e32343290b5...") OR
  md5:("2ae8e804a92249...")
)
AND (
  url:("http://45.130.214.237/cdn/Time/updateexternalprotonAsync/multiImagepythoncdn/SqIwp/..." OR
       "http://62.60.235.190/mariadpPhp/videogameApiServer.php" OR
       "http://87.249.50.64/Phploadhtm.php?id=qq8235_labdotarinoq.txt")
)
AND (
  dst_ip:("45.130.214.237" OR "62.60.235.190" OR "87.249.50.64" OR "217.25.21.61" OR "83.147.253.138")
)

Components Mapped by AI

• Process Names:
  Captures decoy executables mimicking document readers and PDF tools
  
• File Hashes (MD5):
  Targets known variants of droppers and RAT stagers
  
• URLs and Path Patterns:
  Includes JavaScript loaders, file exfiltration paths, and async PHP modules used for delivery or callback
  
• Destination IPs:
  Flags known attacker-controlled infrastructure tied to C2 operations
  

Why It’s Innovative

Writing detection rules for Carbon Black traditionally involves manually stitching together IOC elements, matching syntax, and testing field compatibility. Uncoder AI solves this by:

• Recognizing Carbon Black’s telemetry model (process, md5, url, dst_ip)
  
• Auto-aggregating threat intelligence into logical groupings
  
• Generating a production-ready rule from unstructured IOC reports
  

This eliminates the need for context-switching between PDFs, spreadsheets, and the Carbon Black console.

Operational Value

For Carbon Black defenders and incident response teams:

• Accelerated IOC-to-Rule Pipeline
  Go from CERT-UA threat reports to Carbon Black rules in minutes, not hours
  
• High-Fidelity Threat Matching
  Simultaneously match known malware hashes, behavior patterns, and infrastructure
  
• Improved C2 Discovery
  Flag outbound connections to domains and IPs used for DarkCrystal RAT staging and communication
  

Uncoder AI makes threat intel actionable in Carbon Black by automating the most technical and error-prone parts of detection rule authoring.

Explore Uncoder AI