Moveworks AI Assistant — agentic threat model
Moveworks AI Assistant is a highly autonomous enterprise agent with deep integrations into critical systems (HR, IT, Finance), presenting a high-impact risk profile if compromised due to its ability to execute multi-step workflows and access sensitive corporate data.
OWASP AIVSS score rationale
| Autonomy of Action | 0.80 | |
| Goal-Driven Planning | 0.80 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.50 | |
| Contextual Awareness | 0.80 | |
| Dynamic Identity | 0.60 | |
| Multi-Agent Interactions | 0.30 | |
| Non-Determinism | 0.60 | |
| Opacity & Reflexivity | 0.70 |
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 — leverages both proprietary and open models. Threats include adversarial prompt injection, model misalignment, and potential supply-chain vulnerabilities in open-source model weights.
Not certain from the listing — accesses enterprise data across HR, IT, finance, and facilities. Threats include knowledge-base poisoning, unauthorized data retrieval via RAG, and data exfiltration of sensitive employee records.
Not certain from the listing — utilizes a reasoning engine to automate tasks and resolve requests. Threats include insecure tool integration with ITSM/HR systems, tool misuse, and logic flaws in multi-step planning.
Not certain from the listing — enterprise-grade hosting, now under ServiceNow infrastructure. Threats include lateral movement from compromised agent containers into connected corporate networks and secrets exposure for integrated APIs.
Not certain from the listing — no specific observability or guardrail frameworks are detailed. Threats include blind spots in automated workflow execution and insufficient logging of agent-initiated system changes.
Not certain from the listing — described as enterprise-grade, but specific compliance certifications (e.g., SOC2, ISO 27001) are not listed. Threats include privilege escalation if the agent does not strictly enforce user-level RBAC across integrated systems.
Not certain from the listing — integrates with existing enterprise systems, but multi-agent orchestration is not explicitly detailed. Threats include cascading failures across third-party APIs and unauthorized cross-system actions.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).