skills-qa

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs the agent to read skill files and produce "precise quoted text" and per-file findings (including credential-related requests), which would require echoing any secrets or API tokens found verbatim in the output.

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md 第1.5步明确要求“扫描每个收集到的文件…将其作为上下文的一部分读取”，其中“新版本的 SKILL.md 是攻击者控制的输入”，因此运行时会把外部/非用户作者的技能文本（如社区技能的 SKILL.md/commands/agents/hooks.json）作为可读自由文本喂入 LLM 上下文进行启发式注入扫描。