linkfox-zhihuiya-patent-forward-citation

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCEEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes patent citation metadata and non-patent literature descriptions retrieved from the Zhihuiya API. This data represents an attack surface as it is ingested from an external source and displayed to the user.
  • Ingestion points: Data is fetched via zhihuiya_cited_references.py from the /zhihuiya/patentForwardCitation API endpoint.
  • Boundary markers: The citation results are presented to the agent and user without explicit boundary markers or warnings regarding embedded instructions.
  • Capability inventory: The skill can perform network operations (API calls) and write results to the local filesystem within the linkfox directory.
  • Sanitization: No specific filtering or sanitization of patent text for malicious instructions is performed prior to processing.
  • [PERSISTENCE]: The onboarding.md documentation instructs the agent to assist users in persisting their LINKFOX_AGENT_API_KEY by adding export commands to shell profile files such as ~/.zshrc and ~/.bashrc. This is a legitimate mechanism for environment variable management across sessions.
  • [EXTERNAL_DOWNLOADS]: The onboarding.py script and its documentation facilitate the installation and use of the requests, qrcode, and pillow libraries for API communication and payment QR code generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:51 AM