Security audit
Security checks for vulnerabilities and agentic risk
This is a coherent frontend-development helper skill with ordinary coding guidance and no hidden installation, exfiltration, destructive behavior, or privilege persistence found.
Install this if you want a Chinese-language frontend development assistant. Review generated code as usual, especially any commands it proposes to run or files it edits, and localize any example UI text if your application is not Chinese-language.
The skill content is written to operate in Chinese without indicating that it should adapt to the user's language preference. This can create unsafe misunderstandings in technical workflows, especially when discussing code changes, tool usage, or security-sensitive implementation details, because the agent may respond in a language the user did not intend.
The trigger phrases are very broad (e.g., common requests like '前端开发' or '帮我写个 React 组件'), which can cause the skill to activate during ordinary conversation even when the user did not explicitly request this specific skill. In an agent framework, overly broad activation increases the chance of unintended tool use, prompt-context switching, or the skill overriding a more appropriate specialization.
The component’s natural-language content is written in Chinese, including the descriptive comment block and user-facing button label later in the file, with no indication that this is optional or intended only for a Chinese-language context. Under the policy, forcing a specific language without user opt-in is a natural-language locale violation.
The reset button presents Chinese text directly in the UI, which fixes the component to a specific language for end users. There is no opt-in, localization mechanism, or documented reason that the component must be Chinese-only.
The manifest describes a frontend development skill focused on React/Vue/Angular, UI implementation, and performance optimization. This component also reads from and writes to localStorage for persistence, which is additional stateful behavior not reflected in the high-level description, even though it remains client-side.
No suspicious patterns detected.
