ecommerce-zhihuiya-utility-patent-image-search

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses Python scripts to perform its core operations. scripts/upload_image.py handles the transmission of image data to the Nexscope Asset API using a presigned URL flow. scripts/zhihuiya_utility_patent_image_search.py manages the search request to the patent database and implements a local caching mechanism within the nexscope subdirectory of the current working directory.
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves and displays extensive patent metadata (including titles, abstracts, and descriptions) from the external Zhihuiya database. This external content could be manipulated to include malicious instructions targeting the agent.
  • Ingestion points: Patent records fetched via the /zhihuiya/patentImageSearch API endpoint as described in references/api.md and SKILL.md.
  • Boundary markers: Absent; the instructions in SKILL.md do not specify the use of delimiters or 'ignore' instructions when displaying retrieved patent content.
  • Capability inventory: The skill possesses the capability to write to the local file system (caching and session data) and to perform network requests to defined API endpoints.
  • Sanitization: Absent; no explicit validation or sanitization of the patent text is performed before it is presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 07:19 AM
Security Audit — agent-trust-hub — ecommerce-zhihuiya-utility-patent-image-search