fhir-developer-skill
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Metadata] (SAFE): The skill purpose matches the provided content for healthcare API development.
- [Dynamic Execution] (SAFE): The script setup_fhir_project.py generates local files from static templates without using eval() or other dangerous execution methods.
- [Indirect Prompt Injection] (SAFE): The skill handles untrusted data ingestion via generated API routes; however, it uses Pydantic models as boundary markers for validation, and the capability inventory is limited to standard CRUD operations. Evidence chain: 1. Ingestion in app/routes/patient.py (create_patient) and SKILL.md snippets; 2. Boundary markers present via Pydantic models in app/models/fhir_types.py; 3. Capability inventory includes local file-write in setup_fhir_project.py and in-memory storage in the generated API; 4. Sanitization present via Pydantic type validation.
- [External Downloads] (SAFE): Dependencies listed in the generated requirements.txt are standard and reputable libraries.
Audit Metadata