patent-application-evaluation-assistant
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data such as technical disclosures and research materials, along with external content fetched via web searches to perform its analysis. The instructions lack explicit boundary markers or directions for the agent to ignore potentially adversarial instructions embedded within these inputs.\n
- Ingestion points: Technical disclosures, technical schemes, and external references entered by the user or fetched via search tools as described in SKILL.md.\n
- Boundary markers: The skill lacks explicit delimiters to separate user data from instructions or warnings to ignore embedded commands.\n
- Capability inventory: The skill uses web search tools and a Python interpreter to render files to the local session directory.\n
- Sanitization: There is no mention of sanitization or validation of external content before it is processed by the agent context.\n- [DYNAMIC_EXECUTION]: The skill utilizes a Python tool to programmatically render analysis reports into self-contained HTML files.\n
- Execution method: The agent generates a report based on its findings and then invokes the Python tool to format and save this content as an HTML file in the @session directory.\n
- Source: The data processed for rendering is derived from user-supplied materials and external search results obtained during the evaluation process.
Audit Metadata