qwen-image-2

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly accepts external image URIs in "reference_images" (see Examples and Input Options), so the app fetches and consumes arbitrary public images which the model interprets during editing—exposing it to untrusted third-party content that could carry indirect prompt instructions.