identify-patent-white-space-ip
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows secure patterns for AI agents, including requiring explicit user confirmation before proceeding to critical analytical stages. All external references point to the author's official domains.
- [INDIRECT_PROMPT_INJECTION]: The skill contains a vulnerability surface as it processes external patent datasets and user-supplied context (Stage 0). However, the risk is mitigated by explicit boundary markers and instructions to separate retrieved evidence from inference. The process includes mandatory user confirmation points (Stage 1 and Stage 3) before executing downstream tasks. The capability inventory is restricted to vendor-specific MCP tools and localized HTML report generation. These safeguards ensure that user oversight is maintained throughout the data processing lifecycle.
- [DATA_EXFILTRATION]: All external tool references and URLs point to the author's own infrastructure. The skill does not access or attempt to transmit sensitive local system files, environment variables, or credentials.
Audit Metadata