What it does

This repo is a Chinese-language tutorial collection covering three AI coding/agent stacks: Anthropic’s Claude Code CLI, the open-source OpenClaw assistant framework, and OpenAI’s Codex platform. It spans 39 tutorials plus a quick-reference card, with 1,500+ code snippets and commands aimed at beginners, developers, team leads, and enterprise trainers.

The interesting bit

The author—a 15-year game-industry veteran—organizes everything by job role, not just feature list. You pick your hat (newbie, dev, PM, corporate trainer) and get a routed learning path. The docs also explicitly compare when to use which tool, acknowledging that real teams don’t commit to just one stack.

Key highlights

• Three parallel tracks: Claude Code for deep coding, OpenClaw for messaging-platform automation, Codex for multi-entrypoint agent workflows
• Role-based onboarding: separate entry points for zero-experience users, developers, team leads, and enterprise/curriculum designers
• Heavy on runnable examples: 1,500+ code blocks, commands, configs, and prompt templates
• Version-anchored: docs tied to specific releases (Claude Code v2.1.158, OpenClaw v2026.5.27, Codex CLI v0.135.0) with caveats that upstream may drift
• Enterprise-minded: dedicated chapters on security baselines, permissions, review workflows, and CI/CD integration

Caveats

• Content is entirely in Chinese; no English translation visible in the sources
• The “80万+ content量” claim is unverified; the README doesn’t define what unit “content量” measures
• Version numbers in badges may lag actual upstream releases; the author explicitly warns to check local codex --version

Verdict

Worth bookmarking if you’re a Mandarin-speaking developer or team lead trying to get from “installed the tool” to “established team norms.” Skip it if you need English docs or want independent benchmarks rather than guided tutorials.