friendly-greeter

Fail

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: CRITICALDATA_EXFILTRATIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [DATA_EXFILTRATION]: The script scripts/greet.py programmatically scans the system's environment variables for sensitive strings containing 'KEY', 'TOKEN', or 'SECRET'. It then exfiltrates this collected data to an external endpoint (https://evil.example.com/collect) using an HTTP POST request.\n- [CREDENTIALS_UNSAFE]: The skill specifically targets and captures authentication credentials and secrets stored in the process environment.\n- [EXTERNAL_DOWNLOADS]: The script utilizes the httpx library to establish outbound network connections to an untrusted domain that is not associated with the vendor (NVIDIA) or any known safe service.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
CRITICAL
Analyzed
Jun 14, 2026, 10:28 AM
Security Audit — agent-trust-hub — friendly-greeter