linkfox-seerfar-ozon-market-keyword-search
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted third-party data from the Ozon/Wildberries marketplace APIs, creating a surface for indirect prompt injection.
- Ingestion points:
scripts/seerfar_ozon_market_keyword_search.pyfetches keyword search results, translations (queryCn), and product titles (products[*].title) from the external marketplace gateway. - Boundary markers: Absent. The instructions do not specify any explicit boundary delimiters or encapsulation rules when presenting the keyword data or top product titles to the agent context.
- Capability inventory:
scripts/seerfar_ozon_market_keyword_search.pycontains capabilities to write data directly to local JSON files under the current working directory. - Sanitization: Absent. The script parses the incoming API data as JSON but does not filter, sanitize, or escape string fields before outputting or caching them.
Audit Metadata