Fast, Broad, and Elusive: How Vidar Stealer 2.0 Upgrades Infostealer Capabilities

Fast, Broad, and Elusive: How Vidar Stealer 2.0 Upgrades Infostealer Capabilities

Trend Research analyzes Vidar Stealer v2.0, a full rewrite in C that adds multithreaded data theft, advanced anti-analysis and AppBound bypass techniques, and an automatic polymorphic builder, enabling faster and more evasive credential and wallet exfiltration. The release aligns with declining Lumma Stealer activity and a spike in Vidar adoption, suggesting threat actors may migrate to Vidar and StealC; #Vidar #LummaStealer

Keypoints

  • Vidar 2.0 is a complete rewrite from C++ to C, claimed to improve stability and speed.
  • Multithreaded architecture enables parallelized data collection that scales with CPU cores and memory.
  • Advanced browser credential extraction includes memory injection to bypass Chrome AppBound protections and uses named pipes for inter-process communication.
  • An automatic polymorphic builder generates unique binaries and control-flow flattening obfuscation to hinder static detection and reverse engineering.
  • Targets a wide range of data: browser credentials, crypto wallets, cloud credentials, FTP/SSH files, gaming/social platform data, and screenshots.
  • Exfiltration uses HTTP multipart form submissions to a round-robin C2 infrastructure including Telegram bots and Steam profiles, followed by cleanup and self-deletion.
  • Release timing correlates with Lumma Stealer decline and a measurable spike in Vidar activity, indicating likely increased adoption by underground actors.

MITRE Techniques

  • [T1622 ] Debugger Evasion – Vidar performs debugger detection as part of extensive anti-analysis checks: ‘debugger detection’.
  • [T1497.001 ] Virtualization/Sandbox Evasion – The malware uses timing verification, system uptime validation, and hardware profiling to avoid execution in analysis environments: ‘timing verification, system uptime validation, and hardware profiling’.
  • [T1027 ] Obfuscated Files or Information – Uses control flow flattening and numeric state machines to obscure program logic and hinder analysis: ‘heavy use of control flow flattening, implementing complex switch-case structures with numeric state machines’.
  • [T1055.001 ] Process Injection: Dynamic-link Library Injection – Vidar injects malicious code into running browser processes using reflective DLL or shellcode injection to extract encryption keys from memory: ‘injects malicious code directly into running browser processes using either shellcode or reflective DLL injection’.
  • [T1055.002 ] Process Injection: Portable Executable Injection – The malware performs injections into active processes to steal keys from memory rather than from storage: ‘injects malicious code directly into running browser processes … The injected payload extracts encryption keys directly from browser memory’.
  • [T1082 ] System Information Discovery – Conducts system profiling and hardware enumeration to adapt thread count and behavior: ‘system profiling to collect victim information’ and ‘Thread count is dynamically calculated based on CPU Core count and available physical memory’.
  • [T1083 ] File and Directory Discovery – The file grabber searches user directories and removable drives for valuable files and wallets: ‘systematically searches for valuable files across user directories and removable drives’.
  • [T1087.001 ] Account Discovery: Local Account – Performs account and environment discovery as part of initialization and profiling (enumeration of local accounts implied in discovery steps): ‘system profiling to collect victim information’.
  • [T1518.001 ] Software Discovery: Security Software Discovery – Conducts security software checks during initialization to decide whether to proceed: ‘anti-analysis checks including … security checks’ (implied in evasion checks and software profiling).
  • [T1555.003 ] Credentials from Password Stores: Credentials from Web Browsers – Extracts browser credentials from storage and memory across Chrome, Firefox, Edge and other browsers: ‘comprehensive browser credential extraction capabilities targeting both traditional browser storage methods and Chrome’s latest security protections’.
  • [T1555 ] Credentials from Password Stores – Uses DPAPI decryption of Local State files and memory theft to obtain stored credentials: ‘attempts to extract encryption keys from Local State files using standard DPAPI decryption’.
  • [T1552.001 ] Unsecured Credentials: Credentials In Files – Searches for credential files and configuration files such as FileZilla recentservers.xml and WinSCP sessions: ‘AppDataRoamingFileZillarecentservers.xml’ and ‘SoftwareMartin PrikrylWinSCP 2Sessions’.
  • [T1528 ] Steal Application Access Token – Targets cloud and application tokens such as msal.cache and IdentityService tokens to harvest access tokens: ‘.IdentityService – Azure identity tokens’ and ‘msal.cache – Microsoft Authentication Library cache’.
  • [T1005 ] Data from Local System – Collects files and databases (browser storage, wallet files, documents) from the local system and removable media: ‘file grabber component systematically searches for valuable files across user directories and removable drives’.
  • [T1113 ] Screen Capture – Captures screenshots for additional intelligence prior to exfiltration: ‘screenshot capture for additional intelligence value’.
  • [T1071.001 ] Application Layer Protocol: Web Protocols – Exfiltrates data via HTTP multipart form submissions to C2 infrastructure: ‘data packaging and exfiltration through HTTP multipart form submissions’.
  • [T1102.001 ] Web Service: Dead Drop Resolver – Uses web services like Telegram bots and Steam profiles as communication channels for C2 and data transfer: ’round-robin command-and-control (C&C) infrastructure that includes Telegram bots and Steam profiles as communication channels’.
  • [T1573 ] Encrypted Channel – Employs encrypted channels and authenticated tokens for communication and victim tracking (C2 authentication tokens and build identifiers): ‘uses specific authentication tokens and build identifiers for tracking and victim management’.
  • [T1041 ] Exfiltration Over C2 Channel – Sends stolen data to command-and-control infrastructure over application-layer channels: ‘Exfiltration through HTTP multipart form submissions to a round-robin command-and-control infrastructure’.
  • [T1020 ] Automated Exfiltration – Uses automated routines to package and exfiltrate data, categorizing stolen data and using different operation modes: ‘The malware employs different operation modes to categorize stolen data’ and ‘streamlined exfiltration routines’.

Indicators of Compromise

  • [File names ] targeted data and artifacts – examples: Local State (browser key store), login Data (Chrome/Edge passwords), logins.json (Firefox passwords).
  • [File paths ] credentials and config locations – examples: %AppData%RoamingFileZillarecentservers.xml, %UserProfile%.aws (AWS CLI credentials).
  • [Artifacts ] browser and wallet stores – examples: key4.db (Firefox master key), 0.indexeddb.leveldb (wallet storage), and other wallet extension IndexedDB files.
  • [Communication channels ] C2 mechanisms – examples: Telegram bots, Steam profiles used as command-and-control channels.
  • [Obfuscation/behavioral ] binary characteristics – examples: unique polymorphic builds (automatic morpher) and heavy control-flow flattening observed in samples.
  • [Operational indicators ] runtime behaviors – examples: multithreaded worker creation based on CPU/core count and named pipes usage for IPC when exfiltrating keys from browser memory.


Read more: https://www.trendmicro.com/en_us/research/25/j/how-vidar-stealer-2-upgrades-infostealer-capabilities.html