huashu-seedance

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

• Third-party content exposure detected (medium risk: 0.30). 运行时会读取并解析用户/调用方提供的 scene.json（包含如 shots[].action、meta.sceneDesc 等自由文本），经 engine/make_prompt.mjs/engine/build.mjs 直接进入生成的 prompt；因此 outsider 可通过提交可控的 scene.json 字段注入文本。

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

• Potentially malicious external URL detected (high risk: 0.70). The renderer templates import and load remote JavaScript at runtime from https://cdn.jsdelivr.net/npm/three@0.181.2/+esm which will be fetched and executed in-process (Three.js is required for rendering), so this is a runtime remote-code dependency.