Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only Li Bai persona and creative-thinking skill with broad activation wording, but no code execution, data access, network behavior, or persistence beyond normal skill installation.
Install this if you want a stylized Chinese Li Bai perspective for creative, rhetorical, or mindset work. Be aware that broad phrases about wanting to be more free-spirited may trigger the persona unexpectedly; switch out with phrases like “退出角色” or “切回正常” when you need neutral, factual help.
Referenced artifact was not completely inspected
本目录为**自包含** Skill 包：仅需保留 `SKILL.md` 即可被 Cursor 识别；`references/` 为可选调研附录。发布时可将整个 `libai-perspective` 文件夹打包 zip，或复制到任意项目的 `.cursor/skills/`、或用户目录 `~/.cursor/skill
The README instructs activation via broad natural-language phrases such as “用李白的视角” and “太白模式,” and the metadata indicates even looser conversational triggers may invoke the skill. This can cause unintended activation during ordinary discussion, leading the agent to adopt a stylized persona or decision framework when the user did not explicitly request it, which is a prompt-scope and behavior-boundary issue rather than code execution risk.
The skill’s trigger conditions include broad emotional phrases like wanting to live more freely or not be worn down by mundane affairs, which are common in ordinary conversations and can activate the persona unexpectedly. Because the skill changes response framing and role behavior, accidental activation can steer users into an unsolicited persona mode, increasing the chance of misaligned advice in sensitive or unrelated contexts.
The skill mandates responding in Li Bai’s first-person voice by default once activated, without asking for user consent on language, tone, or roleplay mode. This can override user expectations and reduce clarity, particularly for users who did not explicitly request Chinese persona roleplay or who need neutral, factual assistance.
No suspicious patterns detected.