sesion-mental-health-platform
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to download the project source code from an external GitHub repository (
https://github.com/fahad-hamid/psique-workflow-clinic.git). - [COMMAND_EXECUTION]: The installation process involves executing shell commands to clone the repository, install dependencies via
npm install, and perform database migrations usingnpx prisma. - [PROMPT_INJECTION]: The AI service implementation in
ai/claude.service.tspresents an indirect prompt injection surface by interpolating patient history and session notes directly into clinical assistant prompts without sanitization or boundary markers. - Ingestion points:
sessionNotes,patientHistory, andconversationHistoryinai/claude.service.tsare sourced from the database and potentially external WhatsApp messages. - Boundary markers: Absent; data is directly concatenated into the model's message content.
- Capability inventory: The skill manages clinical records, sends automated WhatsApp messages, and interfaces with the AFIP invoicing system.
- Sanitization: No visible escaping or validation logic is applied to untrusted text before it is processed by the AI models.
Audit Metadata