clinical-decision-support
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for research governance and documentation only. It includes rigorous 'Hard Safety Boundaries' and refusal rules for any patient-specific care, diagnosis, or triage tasks. All bundled scripts are offline-only, requiring no network access or external credentials, and exclusively handle local JSON/CSV/Markdown files. The code includes dedicated validation logic (
_common.py,ensure_no_person_level_keys) to reject common PHI (Protected Health Information) keys like 'patient_id', 'ssn', and 'medical_record_number' at runtime. The documentation and templates strongly emphasize human review, regulatory non-conformance for bedside use, and data de-identification practices. No malicious patterns or data exfiltration vectors were identified.
Audit Metadata