openclaw-medical-skills
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill incorporates detailed 'Safety Boundary' and 'Output Standards' instructions that explicitly guide the agent to provide medical disclaimers, handle crises appropriately, and avoid definitive diagnoses, effectively governing agent behavior in high-risk clinical contexts.
- [COMMAND_EXECUTION]: Preset ripgrep (
rg) command patterns are provided for navigating internal reference files. These are restricted to local documentation and serve as a discovery mechanism for capabilities. - [EXTERNAL_DOWNLOADS]: The skill directs the agent to utilize primary, well-known biomedical sources such as PubMed, PMC, ClinicalTrials.gov, and regulatory agencies like the FDA and EMA for evidence-based responses.
- [DATA_EXFILTRATION]: Robust privacy guidelines are included, instructing the agent to de-identify personal health data and maintain local execution for analysis to protect sensitive information.
Audit Metadata