Security audit
Security checks for vulnerabilities and agentic risk
This is a simple Chinese literary style-writing prompt skill with no code, privileged access, persistence, or hidden behavior evident.
Install this if you want a Chinese-language literary style assistant. Be aware it may respond in Chinese and may activate on broad author-style requests; no local data access, command execution, credentials, or persistent changes were evident in the artifact.
The skill metadata and behavior are entirely Chinese-oriented without stating fallback behavior or requiring user opt-in for language selection. In multilingual environments this can override user expectations and produce responses in Chinese even when the surrounding conversation is in another language, reducing usability and potentially causing miscommunication.
The trigger list includes broad phrases such as '按某某作家风格写' and other style-centric cues that can activate the skill with minimal constraint. This can cause unintended routing or invocation in loosely related conversations, leading to incorrect behavior, surprise language/style transformation, or policy bypass of more specific skills through overmatching.
No suspicious patterns detected.
