external-tech-acquisition
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external technical data throughout its workflow.
- Ingestion points: Technical descriptions, application scenarios, and patent identifiers provided by the user in SKILL.md triggers and inputs.
- Boundary markers: The instructions do not implement explicit delimiters or 'ignore' warnings to isolate user-supplied technical data from the agent's analytical logic.
- Capability inventory: The skill calls multiple high-capability MCP tools (patent.search, patent.fetch) and integrates with several internal analytical sub-skills (e.g., patent-asset-grading, tech-transfer-target-discovery) to generate decision-support reports.
- Sanitization: No input validation or sanitization mechanisms are defined for the metadata retrieved from external databases or the technical descriptions provided by the user before they are processed by the model.
Audit Metadata