ml-researcher
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely descriptive and educational, focusing on frameworks for clinical ML experiment design. It does not include any scripts, executable commands, or configurations that interact with the host system or network.
- [SAFE]: External references point to well-known, reputable sources including medical journals (BMJ, Radiology:AI) and official government documentation (FDA). These are used strictly for informational purposes.
- [SAFE]: There are no indicators of prompt injection, obfuscation, or data exfiltration. The instructions guide the agent on how to structure its reasoning for specific biomedical problems without attempting to bypass safety protocols or access sensitive data.
- [NO_CODE]: The skill does not define any allowed tools or platform-level configuration and contains no shell commands or scripting logic, significantly reducing its attack surface.
Audit Metadata