create-fto-screening-report-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 user-supplied Word documents (.docx) to extract technical features and metadata, which are then used to generate patent search queries and final reports. This ingestion of untrusted data creates a surface for indirect prompt injection.
- Ingestion points: The
extract_docx_inputfunction inscripts/run_generic_fto_report.pyparses the body and tables of the input document. - Capability inventory: The skill possesses the capability to perform network requests to the Patsnap API using user-provided credentials and write JSON, HTML, and DOCX files to the local file system.
- Sanitization: The
scripts/render_report.pyscript employs HTML escaping for all dynamic values in the report output. - Boundary markers: The skill includes a mandatory human-review step where all search expressions must be explicitly approved before execution, and results are labeled as screening-level concern rather than definitive legal conclusions.
- [EXTERNAL_DOWNLOADS]: The skill makes network calls to the official Patsnap REST API (https://connect.patsnap.com) for patent searching and claim retrieval. This behavior is consistent with the skill's stated purpose and originates from the service vendor's official domain.
Audit Metadata