arkweb-design-alignment

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs the agent to Grep+Read code and to display or forward verbatim "代码中实际的实现" and document excerpts (e.g., "{当前代码的实现}" in the coder template), so any hard-coded secrets in code or docs would be output or relayed exactly, creating an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). 运行时 Step 1 会并行读取 requirement.md / design.md / spec.md / analysis.md（并在代码仓库中 Grep + Read 源文件），这些输入文档在“模式 B”中可由用户提供任意路径，因此外部作者可通过提交这些文档内容来注入自由文本并被LLM摄取。