family-menu

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill researches recipes and restaurant information from the web to populate menu data. This ingestion of untrusted external content represents a minor attack surface. Ingestion points: Web search results for recipes and local restaurants (SKILL.md). Boundary markers: None present to delimit search results from agent instructions. Capability inventory: Local file writing and PDF generation via scripts/generate_menu.py. Sanitization: No validation or sanitization of search-derived content is mentioned.
  • [Unverifiable Dependencies] (SAFE): The skill utilizes the 'reportlab' library for PDF generation, which is a standard and reputable Python package.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 09:56 PM