clinical-reports
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions attempting to override agent behavior or bypass safety filters were found. The skill's guidance is focused strictly on medical documentation standards.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file path access, or network communication patterns (e.g., curl, fetch, requests) were detected. The scripts process data locally and only output results to the console or user-specified paths.
- [Remote Code Execution] (SAFE): The skill does not download or execute remote code. All scripts rely on Python's standard library for regex-based validation and file management.
- [Indirect Prompt Injection] (LOW): The skill has a surface for indirect injection as it ingests clinical reports. However, it lacks the high-privilege capabilities (like arbitrary command execution or network access) required to exploit such an injection. The evidence chain includes: (1) Ingestion via local file reading in validation scripts; (2) Use of structured templates as boundaries; (3) Capabilities limited to text processing and file copying; (4) No explicit sanitization beyond regex filtering.
Audit Metadata