crm-customer-import

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt requires reading a user_key from a local file or asking the user to paste it and then embedding that secret verbatim into API request bodies/curl commands, which exposes secrets to the agent and its outputs.

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

• Third-party content exposure detected (high risk: 0.75). 运行时会调用 Apollo 的 /skill/search_customer_profile，其返回的 customers 列表（公司名/联系人/邮箱/地址/remark 等自由文本字段）会被拼入对话上下文并展示给用户，从而把“外部数据源（Apollo/公开数据库检索结果）”的自由文本带入 LLM 上下文。