Wan 2.7 AI — agentic threat model
Wan 2.7 AI is a low-autonomy video generation tool with minimal agentic risk, primarily vulnerable to model-level exploits like jailbreaking for deepfakes and data privacy risks regarding uploaded assets.
OWASP AIVSS score rationale
| Autonomy of Action | 0.10 | |
| Goal-Driven Planning | 0.10 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.10 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.20 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| 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.
Uses proprietary video generation models (Wan 2.7). Primary threats include adversarial prompt injection to bypass safety filters (generating NSFW, violent, or copyrighted content) and potential model stealing of closed-source weights.
Not certain from the listing — likely ingests user-uploaded images and text prompts. Threats include data exfiltration of private user assets and potential data poisoning if user uploads are used to fine-tune future model iterations.
Not certain from the listing — the platform functions as a generation pipeline rather than a complex agentic framework. Orchestration threats are limited to prompt manipulation altering the generation pipeline's parameters.
Not certain from the listing — likely hosted on cloud GPU infrastructure. Threats include GPU resource exhaustion (denial of service) and standard web application vulnerabilities on the hosting platform.
Not certain from the listing — likely relies on input/output content moderation filters. Gaps in observability could allow users to generate and export policy-violating deepfakes without detection.
Not certain from the listing — no compliance certifications (e.g., SOC2, GDPR) or explicit copyright/watermarking policies are detailed in the public directory.
The agent operates as a standalone vertical SaaS tool with no described multi-agent coordination, marketplace integrations, or external agent-to-agent trust boundaries.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).