shopping-assistant
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process large amounts of untrusted data from the open web to provide shopping recommendations.
- Ingestion points: The assistant fetches product data, ratings, and long-form reviews from external websites and specialized shopping platforms (Phase 3 and Phase 8).
- Boundary markers: The instructions do not include specific boundary markers or "ignore previous instructions" warnings to prevent the agent from being influenced by malicious prompts hidden within product reviews.
- Capability inventory: The skill uses web search and the vendor's specialized analysis tools (analyze_related_search_queries, search_queries_vs_search_results). It does not have high-privilege capabilities like file system writing or arbitrary command execution.
- Sanitization: There is no evidence of sanitization, validation, or filtering logic applied to the external content before it is parsed by the agent.
Audit Metadata