skills/nexscope-ai/nexscope-ecommerce-skills/ecommerce-patent-abstract-image-data/Gen Agent Trust Hub
ecommerce-patent-abstract-image-data
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes patent metadata and image paths retrieved from the PatSnap database via an external API call.
- Ingestion points: The
scripts/patent_abstract_image_data.pyscript fetches content from the${NEXSCOPE_PROXY_BASE}/api/v1/tools/research/zhihuiya/abstractImageendpoint. - Boundary markers: There are no explicit instructions or delimiters defined in
SKILL.mdto warn the agent against executing instructions potentially embedded in the external API data. - Capability inventory: The skill uses
urllibfor network communications and writes JSON results to the local filesystem within thenexscope/directory. - Sanitization: The script parses the response using standard JSON libraries but does not perform additional sanitization or filtering on the retrieved strings before they are processed by the agent.
Audit Metadata