huashu-nvwa

SUSPICIOUS. The skill is mostly a prompt-only roleplay artifact with no direct credential theft, malware behavior, or exfiltration path, but it carries medium risk because it is a transitive GitHub-hosted skill from a personal repo, uses open-web research that can import untrusted content, and is designed to produce persuasive crypto-related guidance with real-world implications.

该 Skill 的核心能力与“调研并生成人物/主题 Skill”总体一致，但其实际足迹较宽：会扫描本地 Skills、处理用户本地素材、抓取大量不可信外部内容，并把结果固化为新的 agent 指令文件。最主要风险不是直接恶意，而是供应链与间接提示注入：外部内容和第三方来源可能污染生成的 Skill，形成二次信任链。整体更适合归为可疑/中高风险的技能生成器，而非明确恶意。