iopho-seedance-prompts

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

• Third-party content exposure detected (high risk: 1.00). The skill's Execute section and "10. 与 iopho 管线集成" explicitly instruct the agent to search, download, and analyze public videos (e.g., /iopho-searching-videos ... --platform bilibili and /iopho-getting-videos https://bilibili.com/...), and then to read and reference those untrusted user-generated videos (e.g., "参考@视频 1" for motion, audio, style) as part of prompt construction, so third‑party content can materially influence agent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's execution steps explicitly download external reference videos at runtime (e.g., /iopho-getting-videos https://bilibili.com/BV...) and those fetched videos are then used as @video references that directly control prompt construction, so the https://bilibili.com/BV... URL is a runtime dependency that influences prompts.