Nora AI — agentic threat model
Nora AI presents a low-to-moderate agentic risk profile, primarily characterized by the handling of sensitive personal data (resumes and voice recordings) rather than autonomous action execution. The main security concerns center on data privacy, secure document parsing, and prompt injection vulnerabilities that could manipulate evaluation feedback.
OWASP AIVSS score rationale
| Autonomy of Action | 0.30 | |
| Goal-Driven Planning | 0.40 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.20 | |
| Persistent Memory | 0.30 | |
| Contextual Awareness | 0.50 | |
| Dynamic Identity | 0.10 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.60 | |
| Opacity & Reflexivity | 0.40 |
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 third-party LLMs and speech-to-text/text-to-speech models. Primary threats include prompt injection to bypass interview constraints, model bias in candidate evaluation, and adversarial inputs designed to break the conversational flow.
Not certain from the listing — processes uploaded resumes and voice inputs against a database of 100,000+ questions. Key threats include malicious document uploads exploiting resume parsers, data exfiltration of sensitive candidate PII, and potential poisoning of the question database.
Not certain from the listing — orchestrates the interview flow and adapts questions dynamically. Threats include session hijacking, state manipulation to skip interview stages, and prompt injection that forces the agent to output system prompts or pre-defined questions.
Not certain from the listing — likely hosted on standard cloud infrastructure with web/voice API endpoints. Threats include insecure storage of audio recordings and resumes, and vulnerabilities in third-party document processing or audio rendering libraries.
Not certain from the listing — evaluates answer quality and delivery. Threats include evaluation gaming (candidates finding specific prompt bypasses to receive perfect scores) and a lack of guardrails to detect and log abusive or toxic user inputs during voice sessions.
Not certain from the listing — handles highly sensitive PII (resumes) and voice data. Threats include non-compliance with GDPR/CCPA regarding biometric/voice data retention, lack of clear data deletion mechanisms for users, and weak access controls on user profiles.
The listing describes Nora AI as a standalone mock interview tool with no multi-agent or marketplace integrations. Ecosystem threats are currently negligible, though future integrations with Applicant Tracking Systems (ATS) could introduce supply chain risks.
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.