generate-reference-inspired-brand-worlds

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

• Third-party content exposure detected (medium risk: 0.65). SKILL.md:16-21 requires loading and inspecting every user-supplied image (including the “primary visual-language reference”) and using it as an image input for the LLM-controlled imagegen prompt spine, meaning externally authored free-form text contained in those images (e.g., logos/wordmarks) can be read/included in the runtime prompt context via the vision/image pipeline.