naval-perspective

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). Most links are to articles, transcripts and GitHub repos, but the package/install instruction (npx skills add ...) and multiple GitHub repos from a single, not‑widely‑known user plus personal hosts (S3, bookai.top, huasheng.ai) mean these are potentially risky to run or install without vetting—there are no obvious .exe links but executing remote npm/GitHub code could deliver malware.

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

• Third-party content exposure detected (high risk: 0.75). 该Skill的运行时工作流要求“必须使用工具（WebSearch等）获取真实信息”，这会在运行时抓取/读取公开网页或其他外部文本并把其摘要/内容喂入LLM上下文，从而存在“外部作者自由文本→LLM上下文”的间接提示注入风险。

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

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