ray-nodecheck

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs keeping subscription URLs (which may contain tokens) and other sensitive fields in the output report, requiring the agent to surface secret-bearing values verbatim and thus creating an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.65). SKILL.md 的 Step 1-5 运行时通过 curl/ssh/探针日志/订阅URL把公网返回内容、他人生成的探针日志与订阅文本（以及节点清单里“用户提供”的 server/ssh_host 对应的可读输出）读入并用于LLM推理与表述，因此存在“运行时外部自由文本（网络/日志/订阅）→LLM上下文”的间接提示注入风险。