Xillen Stealer v5 AI Evasive InfoStealer

Xillen Stealer v5 AI Evasive InfoStealer

Xillen Stealer is a Python-based cross-platform information stealer whose v4 and v5 updates expand capabilities to harvest credentials, browser data from 100+ browsers, cryptocurrency from 70+ wallets, container and cloud configurations, and biometric/TOTP data while adding persistence, polymorphism, and P2P C2 features. The malware is marketed via Telegram by a group calling itself Xillen Killers and has public GitHub activity and sample hashes published in the appendices. #XillenStealer #XillenKillers

Keypoints

  • Xillen Stealer v4/v5 is a Python-based infostealer targeting credentials, browser data, cryptocurrency wallets, system information, and biometric/TOTP artifacts across multiple platforms.
  • New functionality includes persistence, Kubernetes/Docker scanning, container credential collection, cloud API exfiltration via a CloudProxy, and P2P command-and-control mechanisms (including blockchain and IPFS options).
  • Anti-analysis and evasion features comprise an AIEvasionEngine (behavioral mimicry, timing/randomization, API call and memory obfuscation), polymorphic Rust engine, and multiple planned process-injection and bypass techniques.
  • Collectors target developer and cloud tooling (IDE configs, AWS/GCP/Azure credentials), password managers (OnePass, LastPass, BitWarden, Dashlane, NordPass, KeePass), SSH keys, SSO tokens, Kerberos tickets, and Windows biometric stores.
  • Exfiltration methods include steganography, NTFS ADS, whitespace encoding, cloud-service routing for blend-in traffic, Telegram reporting for stolen data, and support for DGA/.onion and anonymizing networks for C2.
  • The malware is commercially marketed on Telegram with a GUI for buyers; source code and commits have been posted to a public GitHub repository and multiple sample hashes are listed in the appendices.

MITRE Techniques

  • [T1059.006 ] Python – Xillen Stealer is implemented in Python and uses Python-based components to run stealing routines (‘Python-based information stealer “Xillen Stealer”’)
  • [T1555 ] Credentials from Password Stores – The stealer collects credentials from multiple stores and password managers (‘Ability to steal credentials from password managers’)
  • [T1555.003 ] Credentials from Web Browsers – The malware harvests browser data including history, cookies and saved passwords from many browsers (‘browser data (history, cookies and passwords) from over 100 browsers’)
  • [T1555.005 ] Credentials from Password Managers – Targets specific password managers to extract stored credentials (‘OnePass, LastPass, BitWarden, Dashlane, NordPass and KeePass’)
  • [T1649 ] Steal or Forge Authentication Certificates – The enterprise collector targets corporate certificates for credential or identity theft (‘corporate certificates’)
  • [T1558 ] Steal or Forge Kerberos Tickets – SSOCollector attempts to obtain Kerberos tickets via system commands for authentication theft (‘Kerberos tickets obtained through the klist command’)
  • [T1539 ] Steal Web Session Cookie – The stealer collects web session cookies as part of browser data harvesting (‘browser data (history, cookies and passwords) from over 100 browsers’)
  • [T1552.001 ] Unsecured Credentials: Credentials in Files – The tool scans files and config artifacts (e.g., .env) to harvest API keys and other credentials (‘API keys from .env files’)
  • [T1552.004 ] Unsecured Credentials: Private Keys – Xillen scans for key material such as SSH keys and private key files (‘SSH keys’)
  • [T1552.005 ] Unsecured Credentials: Cloud Instance Metadata API – The malware targets cloud credentials and may abuse cloud metadata/APIs for credential access and exfiltration (‘Cloud credentials and configurations’ / ‘exfiltrating data via Cloud APIs’)
  • [T1217 ] Browser Information Discovery – The stealer enumerates installed browsers and harvests stored browser artifacts (‘steal credentials … browser data (history, cookies and passwords) from over 100 browsers’)
  • [T1622 ] Debugger Evasion – Code includes modules and placeholders intended to evade debugging and analysis (API obfuscation and behavior mimicry) (‘AIEvasionEngine is a module designed to help malware evade AI-based or behavior-based detection systems, such as EDRs and sandboxes’)
  • [T1082 ] System Information Discovery – Xillen collects system and environment information to profile victims and prioritize targets (‘steal cryptocurrency, credentials, system information, and account information’)
  • [T1497.001 ] Virtualization/Sandbox Evasion: System Checks – The evasion engine includes system checks and timing/randomization to avoid sandbox detection (‘mimics legitimate user and system behavior…randomizes execution patterns’)
  • [T1115 ] Clipboard Data – The listed MITRE mappings include clipboard collection capability as part of the stealer’s data collection set (‘T1115 – Clipboard Data’)
  • [T1001.002 ] Data Obfuscation: Steganography – The stealer implements multiple steganography methods to hide data within images and other carriers before exfiltration (‘The SteganographyModule uses steganography (hiding data within an image)’)
  • [T1567 ] Exfiltration Over Web Service – CloudProxy routes stolen data through cloud service domains using HTTP POST to blend traffic with legitimate cloud services (‘exfiltrating data by routing it through cloud service domains’)
  • [T1657 ] Financial Theft – A central goal is stealing cryptocurrency and other financial assets from wallets and related applications (‘Ability to steal … cryptocurrency from over 70 wallets’)

Indicators of Compromise

  • [File Hash ] malware samples – 395350d9cfbf32cef74357fd9cb66134 (confid.py), F3ce485b669e7c18b66d09418e979468 (stealer_v5_ultimate.py), and 1 more hash
  • [File Name ] observed sample filenames – xillen_v5.exe, stealer_v5_ultimate.py (mapped to reported hashes)
  • [Repository URL ] developer/distribution links – https://github[.]com/BengaminButton/XillenStealer, https://github[.]com/BengaminButton/XillenStealer/commit/9d9f105df4a6b20613e3a7c55379dcbf4d1ef465
  • [File Paths ] targeted system artifacts – C:WindowsSystem32WinBioDatabase (Windows biometric DB), TokenBrokerCache (Azure AD token cache)
  • [Cloud Services ] exfiltration/cover domains – AWS, GCP, Azure cloud domains used as CloudProxy placeholders for data exfiltration attempts


Read more: https://www.darktrace.com/blog/xillen-stealer-updates-to-version-5-to-evade-ai-detection