opportunities
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill references the ECharts visualization library from the JSDelivr CDN and retrieves web fonts from Google Fonts in its HTML templates. These are well-known services used for UI rendering and are treated as safe dependencies.
- [COMMAND_EXECUTION]: Instructions strictly prohibit the use of Python or the creation of script files. The skill operates solely through authorized MCP tool calls and localized file-writing, limiting the capability to execute arbitrary or dangerous system commands.
- [DATA_EXFILTRATION]: Network activity is restricted to patent data acquisition via Patsnap's own MCP services. No mechanisms were found that would allow the exfiltration of sensitive user data, environment variables, or private files to unauthorized external endpoints.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied technical directions to build search queries. This input is handled as data within a predefined workflow, minimizing the surface area for user input to override the agent's behavioral logic.
Audit Metadata