scan-emerging-ai-technologies-rd
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from patent searches and scientific literature via MCP tools. Potential risks are mitigated by explicit instructions to escape untrusted text and avoid injecting raw user HTML during the report production process. Boundary markers like evidence IDs are also used to maintain data integrity.
- [DYNAMIC_EXECUTION]: The report template (references/reference_ui_template.html) uses client-side JavaScript for rendering data. The implementation is secure, utilizing textContent and replaceChildren to avoid XSS, and includes a validation layer that checks data types and values before rendering.
- [EXTERNAL_DOWNLOADS]: The skill references marketplace pages for MCP tools (e.g., advanced_patent_search) on the patsnap.com domain. These are legitimate vendor resources required for the skill's operation and do not involve untrusted third-party connections.
Audit Metadata