Rep AI — agentic threat model

Not certain from the listing — likely relies on commercial or fine-tuned LLMs to drive conversational sales. Key threats include prompt injection attacks that could trick the agent into offering unauthorized discounts or misrepresenting store policies.

Not certain from the listing — integrates with e-commerce product catalogs and customer databases to provide contextual support. Threats include unauthorized access to customer PII or order history through conversational data exfiltration.

Not certain from the listing — orchestrates proactive behavioral triggers and executes order management tasks. Insecure tool integration could allow attackers to manipulate API calls to modify orders or trigger unauthorized actions.

Not certain from the listing — deployed as a closed-source SaaS widget embedded on e-commerce storefronts. Threats include widget-based cross-site scripting (XSS) or compromise of the hosting infrastructure leading to supply chain attacks on client websites.

Not certain from the listing — offers analytics and data-driven insights to merchants, but specific real-time guardrails or conversational drift monitoring are not detailed. This creates potential blind spots for abusive or brand-damaging interactions.

Not certain from the listing — as a closed-source commercial tool handling customer transactions, compliance with PCI-DSS and GDPR is critical, but specific security certifications or access control mechanisms are not detailed in the listing.

Not certain from the listing — primarily operates as a standalone concierge interacting with human shoppers and e-commerce platforms (e.g., Shopify) rather than a multi-agent ecosystem, limiting horizontal agent-to-agent threats.