ljx1230

ljx1230 / agent-tutorial

Public

agent教程

33
0
89% credibility
Found May 26, 2026 at 33 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A Chinese-language educational tutorial that teaches how to build AI agents progressively through 9 hands-on demos, starting from basic AI chat and building up to a coding agent with human approval workflows.

How It Works

1
💡 Discover the Tutorial

You hear about AI agents and find a step-by-step tutorial on Bilibili that promises to show you how they work from scratch.

2
🤖 Send Your First Message to an AI

You run a simple program that sends a question to an AI and get back a thoughtful answer - just like chatting with a smart assistant.

3
🧠 Watch the AI Remember Your Conversation

You try again and notice the AI remembers what you said before - it keeps track of your back-and-forth chat naturally.

4
🔧 Give the AI Tools to Use

You teach the AI to actually create files on your computer, not just talk about it - it can now do tasks instead of just answering questions.

5
🔄 See the AI Think Through Problems

The AI now shows its reasoning step by step: thinking about what to do, taking action, then observing what happened before deciding the next step.

6
Choose Your Learning Path
💻
Build a Coding Agent

Create an AI that reads your code, finds bugs, and makes edits safely in a controlled workspace

🔗
Build a Workflow Agent

Create an AI that follows fixed steps like a recipe: classify, inspect, plan, apply, verify, then report

7
🛡️ Add Human Safety Checks

Before the AI makes any changes to your files, it shows you the plan and waits for your approval - you stay in control.

🎉 Your Agent is Ready

You've built a complete AI assistant that can understand tasks, use tools, follow workflows, and ask for your approval before making changes.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 33 to 33 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agent-tutorial?

This is a Chinese-language tutorial that teaches you how to build AI agents from scratch. Starting with basic LLM API calls, it progressively adds capabilities like conversation memory, tool calling, planning, and ReAct-style reasoning loops. By the end, you have a working coding agent that can read and modify files in a restricted workspace, plus a workflow agent with human approval gates. The project uses Python and the DeepSeek API, with nine numbered demos that build on each other like a staircase.

Why is it gaining traction?

The tutorial's main hook is its incremental approach. Instead of throwing you into LangChain or an existing framework, it shows you exactly how each agent capability works under the hood. You see message history handling, tool schema design, and runtime loops implemented in plain Python before any abstraction happens. The progression from "just call the API" to "framework with human-in-the-loop approval" makes the complexity feel earned rather than magical.

Who should use this?

Python developers who understand basic API calls but want to see how agents actually work. If you have used Copilot or similar tools and wondered "how does it decide when to edit files versus just answering questions," this gives you the mental model. It is less useful if you need production-ready code or want to integrate with existing infrastructure.

Verdict

At 33 stars, this is a small, personal tutorial project with solid pedagogical structure but limited community validation. The 0.899% credibility score reflects its early stage. Worth working through if you want to understand agent fundamentals, but do not adopt it as a production dependency.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.