brand-voice-review

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). The links point to a small/unknown domain (ai-skills.ai) and a personal GitHub repo (allinherog-star) that host runnable code (a Python runner) rather than official vendor downloads; there are no direct .exe/.msi links but running code from an untrusted/small personal repo is potentially risky and warrants caution.

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

• Third-party content exposure detected (medium risk: 0.65). 该技能在运行时会把用户提供的 materialUrl（公开网页/社媒/公开文档链接）对应的页面正文作为可读文本注入上下文进行“contextComposer”处理，因此存在来自外部网页的自由文本进入 LLM 的间接提示注入风险。