Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed Chinese narrative-style response skill with no code, file access, network activity, persistence, or credential handling.
Install this if you want the assistant to use a Chinese, story-like tone for emotional, creative, or life-topic conversations. It does not add technical privileges or data access, but it may change the style and language of ordinary conversational replies more often than expected.
The description hard-codes Chinese-language behavior without indicating that output should follow the user's language preference or that the locale restriction is intentional and limited. This can cause unexpected language switching, reduce usability, and interfere with downstream safety or instruction-following expectations when users interact in other languages.
The activation criteria are broad enough to trigger on many ordinary emotional or reflective exchanges, which can cause the assistant to adopt this stylistic persona when the user did not explicitly request it. In a conversational system, overbroad auto-activation can override user intent, degrade reliability, and crowd out more appropriate skills or safer default behavior in sensitive contexts.
No suspicious patterns detected.
