trellis 2 — agentic threat model
Trellis 2 is a specialized generative AI tool for 3D asset creation with low agentic risk, primarily exposed to standard SaaS vulnerabilities, intellectual property theft, and model evasion/stealing rather than autonomous execution threats.
OWASP AIVSS score rationale
| Autonomy of Action | 0.10 | |
| Goal-Driven Planning | 0.10 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.00 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.10 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| 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 — likely uses a proprietary or fine-tuned 3D diffusion/reconstruction model. Threats include adversarial image inputs causing generation failures, model extraction/stealing of the proprietary 3D weights, and prompt injection to bypass content filters.
Not certain from the listing — requires large 3D datasets for training. Threats include copyright/provenance issues of training data, and potential data exfiltration of user-uploaded images used for Image-to-3D generation.
Not certain from the listing — likely does not use an agentic orchestration framework, but rather a standard web API orchestrating model inference. Low risk of tool misuse or memory poisoning due to lack of agentic features.
Not certain from the listing — hosted closed-source service. Threats include GPU instance compromise, API abuse, and insecure export file generation (e.g., malicious 3D file formats like glTF/OBJ containing exploits).
Not certain from the listing — likely lacks specialized LLM guardrails, relying on standard input filtering for prompt safety and basic output validation for 3D mesh integrity.
Not certain from the listing — standard SaaS risks. Needs robust authentication and access controls to protect user-generated intellectual property (3D assets) and payment details.
The listing indicates this is a standalone vertical 3D generator with no multi-agent or marketplace ecosystem features, minimizing ecosystem-level cascading risks.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).