linkfox-zhihuiya-fulltext-image
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
scripts/zhihuiya_fulltext_image.pyto interact with the patent data service. - [EXTERNAL_DOWNLOADS]: The skill makes network requests to
https://tool-gateway.linkfox.com/zhihuiya/fulltextImageandhttps://skill-api.linkfox.com/api/v1/public/feedback. These endpoints are owned by the vendor and are necessary for the skill's primary function. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests data from external API responses. Ingestion points: Data from the Zhihuiya API is processed in
scripts/zhihuiya_fulltext_image.py. Boundary markers: TheSKILL.mdfile provides clear instructions to format output as structured tables and explicitly forbids the fabrication of results. Capability inventory: The skill uses Python scripts to perform network operations. Sanitization: Parameters are validated for presence and length, and API responses are parsed as structured JSON.
Audit Metadata