sdk-security-audit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill mandates providing verifiable code evidence (file:line, code excerpts via Read/Grep) and preserving original identifiers for each reported issue but gives no guidance to redact secrets, so if secrets exist in scanned code the LLM may be forced to output them verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). 该 skill 的运行时工作流在“远程仓库审计”场景中会基于用户提供的仓库 URL 克隆并读取仓库源码文件来进行扫描与 AST/Read/Grep 二次校验，从而让外部（第三方作者）的自由文本/代码内容进入模型上下文；尽管是代码文本而非聊天式描述，但仍属于可由外部作者提交并被立即读取的内容路径。