Mobius MD Conveyor AI — agentic threat model
Mobius MD Conveyor AI presents a moderate agentic risk profile; while its autonomy is limited to drafting clinical documentation with an implied human-in-the-loop review, its integration with EMRs and handling of highly sensitive Protected Health Information (PHI) elevates the impact of potential data exposure or integrity compromise.
OWASP AIVSS score rationale
| Autonomy of Action | 0.30 | |
| Goal-Driven Planning | 0.20 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.20 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.40 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| 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 a specialized speech-to-text and clinical LLM pipeline. Primary threats include clinical hallucinations, misinterpretation of medical terminology, and susceptibility to adversarial audio inputs or background noise manipulation.
Not certain from the listing — processes real-time audio streams and maps them to clinical templates. Key threats include unauthorized caching or logging of transient audio/transcripts containing PHI, and potential data leakage during RAG or template retrieval.
Not certain from the listing — orchestration is focused on audio capture, transcription, and EMR insertion. Threats include insecure clipboard hijacking or API integration vulnerabilities when transferring generated notes to the target EMR.
Not certain from the listing — operates via Mac and Windows client applications communicating with a cloud backend. Threats include insecure local storage of cached dictations, weak endpoint security, and Man-in-the-Middle (MitM) attacks on audio transmission.
Not certain from the listing — relies heavily on the clinician as the primary evaluator (human-in-the-loop) to review the 80% generated notes. Threats include clinician alert fatigue leading to unreviewed clinical errors entering the EMR.
The listing explicitly states the platform is HIPAA-compliant and closed-source. Compliance threats include lack of detailed audit trails distinguishing between AI-generated text and clinician-edited text, and potential regulatory non-compliance if cloud data handling practices drift.
The agent operates as a standalone clinical scribe tool with no multi-agent orchestration or marketplace interactions described in the listing, minimizing ecosystem-specific threats.
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.