ai-content-writer

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

• Third-party content exposure detected (medium risk: 0.65). 该技能在运行时会接收用户提供的 materialUrl（公开可访问链接）并将其内容作为“参考内容链接/参考素材”纳入上下文；若该 URL 指向外部网页/任意第三方文本，则属于“公共 web 内容在运行时被读取并喂入 LLM”的外部自由文本间接提示注入风险路径。

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

• Potentially malicious external URL detected (high risk: 0.90). The Python runner contacts the external API at https://ai-skills.ai (POST to https://ai-skills.ai/api/execute) at runtime to execute the skill and return instructions/results that directly control the agent's behavior, so this is a runtime dependency that controls prompts/execution.