chinese-code-review

Replace commands with suggestions and questions, for example phrasing issues as '建议考虑用 X，因为 Y' instead of direct orders. Attach a priority label like [必须修复] or [建议修改] so the tone can stay mild while the severity stays clear.

Use four standard labels: [必须修复] for security or logic bugs blocking merge, [建议修改] for performance and maintainability issues, [仅供参考] for optional style suggestions, and [问题] when you need the author to explain intent.

Use Chinese for business logic explanations, complex algorithm notes, and internal TODOs; use English for identifiers, error messages, logs, and open-source projects. Keep a space between Chinese and English text and retain technical terms like debounce in English.

Follow the format type(scope): description, using types like feat, fix, refactor, and perf with a Chinese description, for example 'fix(支付): 修复微信支付回调重复处理的问题'. Add body details and issue references like Closes #1234 when needed.

Frame feedback as questions or learning observations rather than corrections, such as asking why recursion was chosen over iteration and raising the stack overflow concern. Code review should be about the code, not the person's seniority.

It should only run when the user explicitly invokes /chinese-code-review, not automatically from context. It also does not replace automated linters, since style issues like indentation should be handled by ESLint, Prettier, or gofmt.