tooluniverse-clinical-trial-design
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions contain behavioral guidelines (e.g., 'MANDATORY', 'DO NOT show tool outputs') intended to ensure the quality and formatting of the generated clinical report. These are standard instructional patterns and do not attempt to subvert safety filters or bypass system constraints.
- [EXTERNAL_DOWNLOADS]: The skill utilizes the ToolUniverse library to query public scientific databases. These sources are reputable academic and governmental entities, including Harvard Medical School's Zitnik Lab, the FDA, and the NIH.
- [DATA_EXFILTRATION]: No sensitive local files are accessed. The skill focuses on aggregating public clinical data into a structured report for the user's research purposes.
- [COMMAND_EXECUTION]: Python code examples provided in the documentation are for illustrative purposes regarding the use of the ToolUniverse API and do not contain any patterns for arbitrary system command execution or shell injection.
Audit Metadata