AgentReadyHomeAgent Listing

← xpander AI

xpander AI — agentic threat model

7.1AIVSS 7.1 · High

xpander.ai presents a high-impact risk profile due to its ability to connect AI agents directly to sensitive, private enterprise systems and execute multi-step workflows. While its VPC-contained AI Gateways offer strong network-level isolation, the broad access to internal APIs requires rigorous application-layer guardrails to prevent unauthorized tool execution.

OWASP AIVSS score rationale

AIVSS = (CVSS_Base + AARS) × Mitigation_Factor, where AARS = (10 − CVSS_Base) × (Factor_Sum / 10) × ThM
CVSS base 8.5AARS uplift 0.9Factor sum 6.0/10Threat ×1.0Mitigation ×0.75
Autonomy of Action
0.80
Goal-Driven Planning
0.80
Self-Modification
0.10
Dynamic Tool Use
0.90
Persistent Memory
0.50
Contextual Awareness
0.60
Dynamic Identity
0.70
Multi-Agent Interactions
0.50
Non-Determinism
0.60
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.

L1 · Foundation Models⚠ not certain from listing

Not certain from the listing — The platform is model-agnostic and does not specify the underlying foundation models used, leaving threats like model-specific adversarial vulnerabilities or alignment gaps dependent on the customer's choice of LLM.

L2 · Data Operations⚠ not certain from listing

Not certain from the listing — While the platform connects to private enterprise systems, the specific handling of RAG pipelines, vector databases, and data lineage controls is not detailed.

L3 · Agent Frameworks✓ mapped

The automatically generated Agent Graph System manages function calling and tool integration. The primary threat here is tool misuse or indirect prompt injection leading to unauthorized actions across connected enterprise systems.

L4 · Deployment & Infrastructure✓ mapped

Strong focus on secure deployment via VPC-contained AI Gateways and on-prem execution, mitigating internet-facing exposure. However, threats remain regarding lateral movement within the VPC if an agent is compromised.

L5 · Evaluation & Observability⚠ not certain from listing

Not certain from the listing — Although the platform enforces 'managed and accurate function calling', there is no explicit mention of built-in evaluation frameworks, continuous monitoring, or real-time guardrail logging.

L6 · Security & Compliance (cross-cutting)✓ mapped

Provides architectural controls for enterprise security (VPC containment, private system gateways), but specific compliance certifications (e.g., SOC2, ISO 27001) or fine-grained authorization policies are not detailed in the listing.

L7 · Agent Ecosystem✓ mapped

Designed to build multi-step agents connecting to 'any system'. This creates an ecosystem of interconnected integrations where a vulnerability in one connected system or agentic interface could cascade across the entire enterprise graph.

MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).