Tech Leader Says I Can't Use Codex, Claude Code, Vibe Coding—Can I Learn?

1. vibe-coding-cn: Run the AI coding workflow first

For users already experimenting with Claude Code, Codex, or Cursor but feeling stuck in a back‑and‑forth loop, the vibe-coding-cn repository serves as a checklist‑style memo. It reminds you to treat the AI as a collaborator that follows a development process rather than a chat bot, covering task definition, context provision, testing, code review, and committing changes. Project URL: https://github.com/tradecatlabs/vibe-coding-cn

2. easy-vibe: Build the first prototype quickly

The easy-vibe project is placed early in the learning path because it avoids overwhelming newcomers with the full stack. It narrows the goal to a small, demonstrable product, letting AI generate a task list from a single sentence, produce a runnable UI, and iterate on a few interaction rounds. It is suitable for product, operations, design, or students who have only asked ChatGPT questions before. Online tutorial: https://datawhalechina.github.io/easy-vibe/zh-cn/ Project URL: https://github.com/datawhalechina/easy-vibe

3. Vibe Vibe: Chinese entry point for absolute beginners

vibe-vibe

(a Datawhale project) assumes the user is not a programmer. It teaches Vibe Coding as a transition from “Coder” to “Commander”, emphasizing clear goal articulation, task decomposition, and iterative refinement of AI output before tackling syntax, tooling, or system design. Online address: https://www.vibevibe.cn Project URL: https://github.com/datawhalechina/vibe-vibe

4. roadmap.sh Vibe Coding Tutorial: Build a small CLI app with Claude Code

This tutorial walks you through creating a command‑line task manager that stores data in tasks.json and is implemented with TypeScript and Node.js. The project is intentionally tiny, allowing you to focus on writing a clear requirement, reviewing the plan, executing commands, and fixing errors without being distracted by full‑stack concerns.

Tutorial URL: https://roadmap.sh/vibe-coding/tutorial

5. Microsoft Learn: Follow the official Copilot Agent workflow

If you spend most of your time in VS Code with GitHub Copilot, the Microsoft Learn material integrates Vibe Coding into the Copilot Agent flow. It guides you from requirement definition and prompt crafting to generating product specs, wireframes, environment setup, and finally a prototype application.

Course URL: https://learn.microsoft.com/en-us/training/modules/introduction-vibe-coding/

6. 鱼皮 AI Guide: A navigation hub for missing pieces

The “Fish‑skin AI Guide” aggregates tutorials on Vibe Coding, Prompt engineering, large‑model workflows, Cursor, Claude Code, Codex, Copilot, RAG, MCP, Spring AI, LangChain, and more. It acts as a searchable index: look up a keyword (e.g., Claude Code) and jump directly to the relevant guide.

Online address: https://ai.codefather.cn Project URL: https://github.com/liyupi/ai-guide

7. AIGuide: Engineering judgment for the second half

AIGuide

focuses on the post‑generation stage: deciding whether AI‑generated code can be merged, handling interface validation, permission scopes, test execution, and long‑term maintenance. It is recommended once you have a functional AI‑generated feature but feel you are merely directing the model to pile code.

Project URL: https://github.com/Snailclimb/AIGuide

How to Choose the Right Resource

For absolute beginners, start with Vibe Vibe or easy-vibe to produce the first tangible prototype and gather feedback. Then, if Claude Code is already installed, follow the small CLI tutorial on roadmap.sh to practice the full requirement‑to‑implementation loop.

When moving to real projects, adopt vibe-coding-cn for a comprehensive workflow checklist. If you encounter gaps, consult the Fish‑skin AI Guide. Finally, once AI‑generated code becomes part of a production system, use AIGuide to evaluate engineering concerns such as cost, permissions, testing, and maintenance.