ecommerce-patent-description-data-translation

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves external patent descriptions, which could potentially contain malicious instructions intended to influence the agent's behavior.
  • Ingestion points: The Python script scripts/patent_description_data_translation.py fetches translated patent specification text from an external Zhihuiya API endpoint.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the fetched content as data rather than instructions.
  • Capability inventory: The skill possesses network communication capabilities and is intended to be used within a context that involves displaying or processing patent text.
  • Sanitization: The script returns the API response directly as JSON without any filtering, escaping, or validation of the retrieved patent text.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 05:40 AM
Security Audit — agent-trust-hub — ecommerce-patent-description-data-translation