Wan 2.7 — agentic threat model
Wan 2.7 is a generative video tool with very low agentic risk, as it lacks planning, tool execution, and autonomous capabilities, though it carries risks related to deepfake generation, model abuse, and resource exhaustion.
OWASP AIVSS score rationale
| Autonomy of Action | 0.10 | |
| Goal-Driven Planning | 0.00 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.00 | |
| Persistent Memory | 0.00 | |
| Contextual Awareness | 0.10 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.70 | |
| Opacity & Reflexivity | 0.80 |
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.
The core of Wan 2.7 is a text-to-video and image-to-video foundation model. Primary threats include adversarial prompt injection to bypass safety filters (generating NSFW, violent, or copyrighted content), model extraction/stealing, and output misalignment.
Not certain from the listing — details on the training dataset, video/image ingestion pipelines, or RAG/vector stores are not provided. Potential threats include training data poisoning and copyright/IP infringement from ingested training data.
Not certain from the listing — Wan 2.7 appears to be a single-turn generation model rather than an agentic framework. It lacks explicit orchestration, planning, memory, or tool-calling capabilities.
Not certain from the listing — the hosting infrastructure for the online tool is unspecified. Threats include GPU resource exhaustion (DoS) due to heavy video rendering demands and potential container escape if user-uploaded images are processed insecurely.
Not certain from the listing — no details are provided regarding output moderation, guardrails, or logging. Gaps here could allow the generation of deepfakes, CSAM, or copyrighted material without detection.
Not certain from the listing — compliance frameworks (e.g., GDPR, EU AI Act regarding deepfakes) and user authentication/authorization controls are not detailed.
Not certain from the listing — there is no evidence of multi-agent coordination or marketplace integration. The tool operates as a standalone generator.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).