xiaohongshu-cover

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs the agent to help the user set their REDFOX_API_KEY by generating shell/PowerShell commands that include the user's API key value (e.g., export REDFOX_API_KEY=<值>, SetEnvironmentVariable(..., "你的API Key", ...)) which requires the LLM to handle/insert secret values verbatim into output.

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

• Third-party content exposure detected (medium risk: 0.30). 在步骤4的“AI批量图像分析与特征提取”中，Agent 会从第三方红狐数据接口返回的 JSON 里提取 coverUrl 并逐张访问这些 URL 进行图像理解（其中封面上的文字/内容属于外部作者生成的自由文本/视觉信息）。