review-analysis

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs reading an API key from .mcp.json and inserting it into curl requests as ?key=${API_KEY}, which requires the agent to access and embed the secret verbatim in generated commands/outputs.

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

• Third-party content exposure detected (medium risk: 0.30). 运行时在 Step 4/解析阶段通过 product_reviews（SSE）读取 Sorftime 返回的差评文本并将 result.content[0].text 解析成评论数组供 LLM 分析（攻击者可通过提交恶意/注入式评论内容污染该数据源，而工作流会在未先选择特定评论项的情况下直接摄取）。