# Vibe Coding CN

A Chinese-language guide to "vibe coding", describing a standard workflow of prompts, skills, context management and quality gates for pairing with an AI coding agent from an idea to a finished product.

This is a Chinese-language tutorial on vibe coding, AI pair-programming, organized around a standard workflow of prompt design, reusable skills, context management, quality gates and an engineering feedback loop, from a raw idea to a shipped product.

Its core arguments describe vibe coding as a goal-driven, constrained, verifiable loop: a vague requirement is clarified into a versioned goal baseline, then an agent repeatedly observes the current state, identifies the gap to the goal, chooses an action, gathers verification evidence, and accepts, corrects, rolls back or switches strategy. Related arguments cover what a model's output can directly or indirectly reach, why AI-generated results should be treated as unverified candidates that need an independent review session rather than self-confirmation, and why unfinished business logic should be assembled from mature existing tools and repositories rather than written from scratch.

It is released under the MIT license.

Website: https://github.com/tradecatlabs/vibe-coding-cn