skills/nexscope-ai/nexscope-ecommerce-skills/ecommerce-ozon-product-detail-search/Gen Agent Trust Hub
ecommerce-ozon-product-detail-search
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches product details, including titles and seller names, from the external Ozon marketplace. Because this content is generated by third parties, it could contain malicious instructions designed to manipulate the agent's behavior.
- Ingestion points: Data is ingested through the
seerfar/ozon/productDetailSearchAPI response, which is then processed and printed byscripts/ozon_product_detail_search.py. - Boundary markers: The script output lacks explicit delimiters or boundary markers to separate the untrusted external data from the skill's own instructions.
- Capability inventory: The skill's Python script is capable of performing network requests (
urllib.request.urlopen) and writing data to the local file system (within thenexscope/directory tree). - Sanitization: The skill does not appear to perform sanitization, filtering, or escaping of the API response data before presenting it to the agent's context.
Audit Metadata