high-value-patent-package
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied search queries and patent data retrieved from external Patsnap APIs, creating an indirect prompt injection surface.
- Ingestion points: User-provided patent search query; API responses from connect.zhihuiya.com including patent titles, AI-generated summaries, and legal event data (found in scripts/hv_1_fetch.py, scripts/hv_2_numeric.py, scripts/hv_3_legal.py, scripts/hv_5_display.py).
- Boundary markers: No explicit prompt delimiters are used for the query, but the skill mitigates output risks by sanitizing data during report generation.
- Capability inventory: Performs HTTP requests to vendor APIs and drawing URLs; writes JSON checkpoints, HTML reports, and Word documents to the local file system (found in scripts/hv_7_html_a.py, scripts/hv_8_word.py).
- Sanitization: Uses html.escape for HTML report generation and utilizes the python-docx library for structured Word document creation to ensure output integrity.
Audit Metadata