portfolio-flow

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires returning the default administrator account password verbatim after deployment (「默认密码: PortfolioFlow_{随机6位}」/“返回访问 URL + 默认管理员账号密码”), which forces the agent to handle and output a secret value directly.

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

• Third-party content exposure detected (high risk: 0.85). 运行时 LLM 上下文会被“访客自由提问”注入：前台将用户输入的 question 通过 fetch('/api/assistant', {question: q}) 发送到 templates/functions/api/assistant.js，该函数在 answerWithAI() 中把 question 拼进 DeepSeek prompt（访客的问题：${question}）作为 LLM 输入。

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)