family-health-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple local JSON files, including
data/family-health-tracker.json,data/hypertension-tracker.json, anddata/diabetes-tracker.json. It does not define boundary markers or explicit sanitization steps for this content, which could allow malicious instructions embedded in the health records to influence the agent's behavior. - Ingestion points: Files
data/family-health-tracker.json,data/hypertension-tracker.json,data/diabetes-tracker.json, anddata/profile.json(Step 2). - Boundary markers: None identified. The agent is instructed to read and validate but not to ignore instructions inside the data.
- Capability inventory: Uses
Read,Write,Grep, andGlobtools. - Sanitization: No validation or sanitization of the health data content is described beyond basic JSON structure checks.
- [DATA_EXPOSURE]: The skill's primary purpose involves reading and processing sensitive personal health information (PHI) and genetic risk data. While no network exfiltration tools are used, the handling of such sensitive data within a skill that also possesses
Writecapabilities increases the risk profile if the agent's logic is subverted.
Audit Metadata