ai-web-automation-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The implementation in main.py is minimal and restricted to its stated scraping purpose using the standard requests library.
  • [SAFE]: The skill includes comprehensive safety guardrails and third-party safety documentation that explicitly prohibit malicious use cases like security bypass or insecure credential handling.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from external URLs, which constitutes an indirect prompt injection surface.
  • Ingestion points: main.py (via requests.get).
  • Boundary markers: Absent in markdown report generation.
  • Capability inventory: openclaw (command execution) and main.py (file system access).
  • Sanitization: Absent in main.py.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:30 PM
Security Audit — agent-trust-hub — ai-web-automation-hardened