patent-mining-agent
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill implements a mandatory '挖掘前置确认卡' (CP-0) which requires the user to manually confirm the mining type, mode, and technical lines before any analytical blocks are executed. This effectively prevents the agent from performing autonomous actions without explicit user consent.
- [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface through the use of
web.searchandpatent.searchtools to gather technical information. It mitigates potential risks by mandating that all retrieved data be clearly labeled with its source (e.g., [网页检索], [Patsnap检索]) and strictly prohibits the agent from fabricating patent data or information. - [EXTERNAL_DOWNLOADS]: The documentation directs users to the official vendor developer portal at
open.zhihuiya.comfor configuring Model Context Protocol (MCP) services. This URL belongs to the skill's author (Patsnap/Zhihuiya) and is used for legitimate configuration purposes.
Audit Metadata