video-request-architect

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

• Third-party content exposure detected (high risk: 0.80). The skill bundle includes runtime code that fetches arbitrary public web content (e.g., _postplus_shared/.../scripts/download_videos_from_manifest_with_ytdlp.mjs which invokes yt_dlp on item.sourceUrl, plus network_runtime.mjs requestText/requestJson/requestBytes and hosted_media_generation_bridge.downloadHostedMediaFile) and the SKILL's workflow explicitly uses and binds reference videos/images as inputs, so untrusted third‑party, user‑generated content can be downloaded and then read/interpreted as part of the request-building workflow.