yichen-wecom-local-vault

Warn

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes high-privilege shell commands to facilitate data access.\n
  • Uses sudo in scripts/scan_dbkey_manager_macos.py to allow the script to read the memory of other processes on the system.\n
  • Uses the codesign utility in scripts/capture_key_macos.py to re-sign the target application, effectively removing security protections like the Hardened Runtime to allow for instrumentation.\n- [CREDENTIALS_UNSAFE]: The skill is designed to capture and store encryption keys from active memory.\n
  • Implements process hooking via Frida to intercept calls to cryptographic functions and extract the raw database keys.\n
  • Employs a Mach VM memory scanner to identify and extract sensitive keys from process memory objects.\n- [DATA_EXFILTRATION]: The skill automatically discovers and extracts sensitive private communication data.\n
  • Programmatically accesses databases containing personal and corporate message histories, contacts, and session details located in ~/Library/Containers/com.tencent.WeWorkMac/.\n
  • While data is kept in a local vault, the automated extraction and decryption of private enterprise data constitutes a significant privacy exposure.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 20, 2026, 01:39 PM
Security Audit — agent-trust-hub — yichen-wecom-local-vault