Saffron Health — agentic threat model
Saffron Health presents a high-risk profile due to its end-to-end automation of specialist referrals, which handles highly sensitive Protected Health Information (PHI) and interacts with clinical systems without documented human-in-the-loop safeguards.
OWASP AIVSS score rationale
| Autonomy of Action | 0.80 | |
| Goal-Driven Planning | 0.70 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.60 | |
| Contextual Awareness | 0.70 | |
| Dynamic Identity | 0.40 | |
| Multi-Agent Interactions | 0.20 | |
| Non-Determinism | 0.50 | |
| 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 utilizes clinical or general-purpose LLMs to parse referral requests. Risks include prompt injection attacks that could alter clinical intent, patient data, or destination specialists.
Not certain from the listing — must ingest and process highly sensitive Protected Health Information (PHI) from EHRs. Risks include unauthorized PHI exposure, lack of data lineage, and potential leakage of patient records through model context or vector databases.
Not certain from the listing — orchestration likely manages multi-step workflows (extraction, matching, scheduling). Risks include insecure tool execution when writing back to EHRs or scheduling systems, and state manipulation of active referrals.
Not certain from the listing — requires highly secure, HIPAA-compliant hosting infrastructure. Risks include insecure API integrations with hospital networks and lack of strict network isolation for the agent's execution environment.
Not certain from the listing — clinical workflows demand rigorous audit logging and drift detection. Risks include a lack of observability into why a specific specialist was selected or failure to log failed referral transmissions.
Not certain from the listing — must adhere to HIPAA, HITECH, and potentially SOC2/HITRUST. The listing does not specify any access controls, encryption standards, or regulatory compliance frameworks.
Not certain from the listing — may interact with external insurance portals, scheduling APIs, or other clinic agents. Risks include cascading failures if external booking systems are compromised or return unexpected payloads.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.