Security audit
Security checks for vulnerabilities and agentic risk
This is a simple gift-recommendation skill with no executable code, persistence, credential use, or hidden data handling.
Safe to install from a security perspective. Users should know it is mainly written for Chinese gift-advice workflows and may activate broadly on gift recommendation requests; review suggestions for cultural fit and budget accuracy before acting on them.
The skill description is broad enough to trigger on many ordinary gift-related conversations without clear boundaries for when the skill should or should not activate. Over-broad invocation can cause unintended routing, reduce user control, and expose the model to follow this skill in contexts where another skill or general assistant behavior would be more appropriate.
The skill content and examples strongly default to Chinese, and the description does not clearly instruct the agent to preserve or ask for the user's preferred language. This can lead to mismatched-language responses, degraded usability, and confusion, especially in multilingual environments where invocation may occur from English prompts.
No suspicious patterns detected.
