meitu-visual-me

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly accepts and ingests arbitrary user- or URL-hosted images/videos (e.g., "Image input supports both local paths and URLs" and many workflows like swap-bg, style-remix, image-generate/image-edit/image-to-video/motion-transfer in SKILL.md and references/models.md) and then reads/extracts composition/style/person details from those references to build prompts and choose commands, so untrusted third-party content can directly influence prompt construction and subsequent tool actions.