analyze-patent-search-results-ip

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes patent records and search artifacts from local CSV files and external MCP tools. * Ingestion points: Data is read from candidate_pool.csv, core_recall.csv, and retrieved via Patsnap's patent search and briefing MCPs. * Boundary markers: Strict artifact schemas are defined for JSON and CSV outputs. * Capability inventory: The agent writes analysis results to local files and generates an offline HTML snapshot. * Sanitization: The instructions explicitly require escaping all retrieved values and validating URLs during the HTML report generation phase.
  • [EXTERNAL_DOWNLOADS]: The skill relies on external Patsnap MCP services for patent data retrieval. * The skill references official tools like advanced_patent_search and patent_briefing hosted on the vendor's marketplace at open.patsnap.com.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 01:23 AM