prepare
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses local health files such as data/allergies.json and data/health-feeling-logs.json to provide the user with a medical summary. This behavior is the primary function of the skill and does not involve network exfiltration or unauthorized external access.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from health logs into the agent's prompt context without explicit boundary markers.
- Ingestion points: Reads health data from data/health-feeling-logs.json and data/allergies.json to generate summaries.
- Boundary markers: Absent. The data from the files is interpolated into the summary output format.
- Capability inventory: The skill possesses Read and Write permissions for local file operations.
- Sanitization: There is no evidence of sanitization or filtering for the content within the health records before it is processed by the agent.
Audit Metadata