ai-ad-creative

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

• Third-party content exposure detected (medium risk: 0.65). 该技能的运行时输入支持 materialUrl（公开可访问的文章/页面/竞品内容链接），平台在执行时会读取该 URL 的页面正文作为可读文本进入 LLM 上下文，从而引入“公共网页内容/任意 URL”这一外部自由文本来源的间接提示注入风险。

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

• Potentially malicious external URL detected (high risk: 0.80). scripts/run.py makes runtime API calls to the external service at https://ai-skills.ai (DEFAULT_BASE_URL) — POSTing to /api/execute — so execution and model prompts/responses are controlled by that remote endpoint.