ShowAndTell (YC F24) — agentic threat model
ShowAndTell operates in a high-risk healthcare environment, handling sensitive Patient Health Information (PHI) and interacting directly with patients. Its integration with Patient Management Systems and automated messaging capabilities present significant data privacy and clinical safety risks if compromised.
OWASP AIVSS score rationale
| Autonomy of Action | 0.50 | |
| Goal-Driven Planning | 0.40 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.60 | |
| Persistent Memory | 0.70 | |
| Contextual Awareness | 0.80 | |
| Dynamic Identity | 0.20 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.60 | |
| Opacity & Reflexivity | 0.50 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — likely relies on commercial LLMs to generate patient-facing explanations and treatment plans. Main threats include prompt injection leading to the generation of incorrect medical/dental advice or leakage of system instructions.
Not certain from the listing — ingests patient data and treatment plans from existing patient management systems. Main threats include unauthorized exfiltration of sensitive Patient Health Information (PHI) and data poisoning of the context window.
Not certain from the listing — orchestrates automated follow-up and re-care messaging. Main threats include insecure tool integration with communication APIs (SMS/Email) and lack of strict input validation before executing PMS database queries.
Not certain from the listing — likely hosted as a cloud-based SaaS. Main threats include insecure storage of Patient Management System (PMS) API keys and credentials, and potential lack of strict multi-tenant isolation.
Not certain from the listing — no mention of clinical safety guardrails or hallucination detection. Main threats include undetected generation of harmful or misleading medical advice sent directly to patients.
Not certain from the listing — operates in the dental/healthcare space, necessitating strict HIPAA compliance and Business Associate Agreements (BAAs). Main threats include lack of comprehensive audit logs for patient interactions and unauthorized access to PHI.
Not certain from the listing — primarily functions as a standalone agent integrating with a practice's PMS. Main threats include cascading failures if the underlying PMS API changes or if third-party communication channels are compromised.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).