ai-video-generation

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

• Third-party content exposure detected (high risk: 0.80). The skill’s primary workflow examples (SKILL.md) show running infsh apps with arbitrary external resources (e.g., "image_url", "audio_url", "video_url" like "https://portrait.jpg", "https://speech.mp3", "https://..."), which means it fetches and ingests untrusted public third‑party media that the agent will read/interpret as part of generating videos and thus could carry embedded instructions that affect outputs or follow‑on actions.