Quollie AI — agentic threat model
Quollie AI presents a moderate-to-high risk profile due to its autonomous interaction with external users (website visitors) and write-access integrations (calendar booking, email nurturing). A compromise could lead to automated phishing, calendar spam, and exposure of sensitive lead/revenue data.
OWASP AIVSS score rationale
| Autonomy of Action | 0.80 | |
| Goal-Driven Planning | 0.60 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.60 | |
| Persistent Memory | 0.70 | |
| Contextual Awareness | 0.60 | |
| Dynamic Identity | 0.20 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.70 | |
| Opacity & Reflexivity | 0.60 |
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 (e.g., GPT-4) to drive natural, personalized conversations. Primary threats include prompt injection from website visitors designed to hijack the conversation or extract system instructions.
Not certain from the listing — stores conversation history, lead behavior, and revenue tracking metrics. Threats include unauthorized access to customer PII and lead data stored in its database or vector stores.
Orchestrates multi-step workflows including lead qualification, calendar syncing, and email nurturing. Insecure tool integration is a major threat, where prompt injection could trigger unauthorized calendar bookings or malicious email dispatches.
Not certain from the listing — deployed as a closed-source SaaS platform. Threats include the exposure of API keys for integrated calendar and email services, and potential container/host compromise.
Not certain from the listing — provides a dashboard for tracking conversations and revenue, but lacks explicit details on real-time guardrails, prompt filtering, or anomaly detection for malicious inputs.
Not certain from the listing — as a closed-source paid tool handling customer contact details and calendar access, it requires robust access controls and GDPR/CCPA compliance, though none are explicitly detailed.
Not certain from the listing — operates primarily as a standalone horizontal agent, but its integration with external email and calendar ecosystems introduces trust boundaries that can be exploited if those external services are compromised.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).