restaurant-pricing
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and logic are consistent with its stated purpose of competitive pricing analysis and do not contain any malicious patterns, obfuscation, or unauthorized access.
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because it processes untrusted data from external web searches (competitor menus).
- Ingestion points: WebSearch tool outputs in Step 1 and Step 2.
- Boundary markers: None; the instructions do not implement markers like XML tags to distinguish search data from system instructions.
- Capability inventory: The skill uses web search tools and performs file-write operations to save the pricing report.
- Sanitization: No explicit validation or filtering of external search content is performed before data interpolation.
Audit Metadata