fhir-basics
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides Python snippets that use
subprocess.runto executecurlfor network requests, which is a documented workaround for sandbox proxy constraints. - [EXTERNAL_DOWNLOADS]: The skill interacts with the well-known SMART on FHIR public test server (r4.smarthealthit.org) for synthetic data retrieval.
- [INDIRECT_PROMPT_INJECTION]: Data is ingested from external FHIR API endpoints (SKILL.md). No explicit prompt boundary markers are included in the instructions. The skill capability inventory includes
subprocess.runfor network calls. Sanitization is performed through standard JSON parsing viajson.loads.
Audit Metadata