cellinlab/awesome-agentic-ai-zh

awesome-agentic-ai-zh is a structured learning roadmap for people who want to learn how to build and use AI agents, written mainly in Traditional Chinese with Simplified Chinese and English versions also available. Rather than being a piece of software, it is a curated guide: a learning map, more than 145 hand picked resources, and small hands on exercises meant to take someone from knowing nothing about large language models to being able to design systems with multiple cooperating agents. The material is organized into eight stages split across two tracks. Track A, called CLI Power User, is for people who want to use existing command line AI agent tools such as Claude Code, Codex, or Gemini CLI more effectively, without necessarily building their own agent. Track B, called Agent Builder, is for people who want to build an agent from scratch, covering topics like agent frameworks, memory and retrieval, and running multiple agents together. Both tracks share early foundational stages covering basic Python, git, and prompt writing, and later share two hub stages covering the Claude Code ecosystem and agent interfaces like computer use and browser use. Each stage links to written material and includes one to five small practice exercises, described as 70 to 150 lines of starter code with parallel examples using both a local model tool called Ollama and the Anthropic SDK, plus simple automated tests. Beyond the two main tracks, the project also offers five shorter paths aimed at specific groups such as researchers, developers, teachers, knowledge workers, and everyday chat tool users who want a lighter introduction without the full track. The whole project is released under the MIT license, so all content can be reused freely. Estimated timelines given in the README suggest Track A takes about eight to ten weeks and Track B takes at least sixteen to twenty two weeks of part time study. The full README is longer than what was shown.