Learn Docker & K8s Game
学习Docker与K8s的游戏
Skill by
ara.so — Daily 2026 Skills collection.
An open-source, AI-driven interactive learning game that teaches Docker, Linux, networking, and Kubernetes through a story-driven simulation. No web app, no video courses — just you, your AI editor, a terminal, and the chaotic coffee startup NoCappuccino Inc.
技能由
ara.so提供 — 2026年度每日技能合集。
这是一款开源的AI驱动交互式学习游戏，通过故事驱动的模拟场景教授Docker、Linux、网络知识和Kubernetes。无需网页应用，无需视频课程——只有你、你的AI编辑器、终端，以及状况百出的咖啡创业公司NoCappuccino Inc。
The game runs entirely inside your AI editor. Markdown prompt files in this repo act as the game engine. When a user says "let's play," the AI reads
(or the editor-specific entry point), becomes "Sarah" (the mentor character), and walks the learner through real Docker/K8s scenarios using their actual terminal.
User says "let's play"
→ AI reads AGENTS.md + engine/rules.md + engine/narrator.md
→ AI becomes Sarah, the senior DevOps mentor
→ Story begins: Dave broke staging, fix it with containers
→ Lessons → Challenges → verify.sh → next chapter
游戏完全运行在你的AI编辑器内部。本仓库中的Markdown提示文件充当游戏引擎。当用户输入"let's play"时，AI会读取
（或对应编辑器的专属入口文件），化身导师角色"Sarah"，通过用户的真实终端引导学习者完成真实的Docker/K8s场景任务。
用户输入"let's play"
→ AI读取AGENTS.md + engine/rules.md + engine/narrator.md
→ AI化身为资深DevOps导师Sarah
→ 故事开启：Dave搞坏了预发布环境，用容器修复它
→ 课程学习 → 挑战任务 → verify.sh验证 → 进入下一章节
bash
git clone https://github.com/ericboy0224/learn-docker-and-k8s.git
cd learn-docker-and-k8s
Open the directory in your AI editor, then type: "let's play"
bash
git clone https://github.com/ericboy0224/learn-docker-and-k8s.git
cd learn-docker-and-k8s
在你的AI编辑器中打开该目录，然后输入："let's play"
| Requirement | Chapters | Notes |
|---|
| Docker | Ch 1–7 | Required |
| Docker Compose v2 | Ch 1–7 | Required |
| kubectl | Ch 6–7 | Optional |
| kind | Ch 6–7 | Optional |
| AI Editor | All | Claude Code, Cursor, Windsurf, Copilot, Cline, Codex, Gemini CLI |
| 依赖 | 适用章节 | 说明 |
|---|
| Docker | 第1–7章 | 必须 |
| Docker Compose v2 | 第1–7章 | 必须 |
| kubectl | 第6–7章 | 可选 |
| kind | 第6–7章 | 可选 |
| AI编辑器 | 全部章节 | Claude Code、Cursor、Windsurf、Copilot、Cline、Codex、Gemini CLI |
bash
bash engine/environment-check.sh
This verifies Docker, Docker Compose, and optional Kubernetes tools are installed and running.
bash
bash engine/environment-check.sh
该脚本会验证Docker、Docker Compose以及可选的Kubernetes工具是否已安装并正常运行。
Editor Entry Points
编辑器入口点
Each AI editor reads a different config file automatically:
| Editor | Entry File |
|---|
| Claude Code | |
| Cursor | |
| Windsurf | |
| Cline / Roo Code | |
| GitHub Copilot | .github/copilot-instructions.md
|
| Gemini CLI | |
| All others | |
The AI editor will read the appropriate file on startup and load game context from
and
.
不同的AI编辑器会自动读取对应的配置文件：
| 编辑器 | 入口文件 |
|---|
| Claude Code | |
| Cursor | |
| Windsurf | |
| Cline / Roo Code | |
| GitHub Copilot | .github/copilot-instructions.md
|
| Gemini CLI | |
| 其他编辑器 | |
AI编辑器会在启动时读取对应的文件，并从
和
目录加载游戏上下文。
Type these naturally in your AI editor chat:
/play — start or resume the game
/env-check — verify Docker/K8s setup before starting
/progress — view save file (.player/progress.yaml)
/hint — progressive hint (3 levels: nudge → direction → near-answer)
/verify — check your challenge solution
/next — advance to next lesson or challenge
/skip-to <N> — jump to chapter N (triggers a quiz gate)
/cleanup — remove all learn-* Docker resources safely
You can also speak naturally:
"I'm stuck on the port mapping"
你可以在AI编辑器的聊天框中自然输入这些命令：
/play — 开始或继续游戏
/env-check — 开始前验证Docker/K8s环境配置
/progress — 查看存档文件(.player/progress.yaml)
/hint — 逐级提示（3个等级：轻微引导 → 方向指引 → 接近答案）
/verify — 检查你的挑战任务解决方案
/next — 进入下一课或下一个挑战
/skip-to <N> — 跳转到第N章（会触发测验门槛）
/cleanup — 安全移除所有learn-*开头的Docker资源
你也可以用自然语言提问：
Ch1 → Ch2 → Ch3 → Ch4 → Ch5 → Ch6 → Ch7
| Chapter | Title | Core Skills |
|---|
| 1 | 📦 It Works on My Machine | containers, images, port mapping |
| 2 | 🏋️ The 2GB Espresso | multi-stage builds, layer caching, .dockerignore |
| 3 | 💾 The Vanishing Beans | volumes, bind mounts, persistence |
| 4 | 🔌 The Silent Grinder | DNS, bridge networks, isolation |
| 5 | 🎼 The Symphony of Steam | Docker Compose, health checks, secrets |
| 6 | ⎈ The Giant Roaster | Pods, Deployments, Services, self-healing |
| 7 | 🔥 The Great Latte Leak | rolling updates, Secrets, HPA, chaos triage |
Linux fundamentals (namespaces, cgroups, mounts) and networking (DNS, NAT, subnets, iptables) are taught contextually throughout — no dedicated lecture needed.
第1章 → 第2章 → 第3章 → 第4章 → 第5章 → 第6章 → 第7章
| 章节 | 标题 | 核心技能 |
|---|
| 1 | 📦 我本地运行没问题啊 | 容器、镜像、端口映射 |
| 2 | 🏋️ 2GB大小的浓缩咖啡 | 多阶段构建、层缓存、.dockerignore |
| 3 | 💾 消失的咖啡豆 | 卷、绑定挂载、数据持久化 |
| 4 | 🔌 沉默的研磨机 | DNS、桥接网络、隔离 |
| 5 | 🎼 蒸汽交响乐 | Docker Compose、健康检查、密钥 |
| 6 | ⎈ 巨型烘焙机 | Pod、Deployment、Service、自愈 |
| 7 | 🔥 拿铁大泄漏 | 滚动更新、Secret、HPA、故障排查 |
Linux基础（命名空间、cgroups、挂载）和网络知识（DNS、NAT、子网、iptables）会在整个课程中结合上下文讲解——无需单独的理论课。
learn-docker-and-k8s/
├── AGENTS.md # Universal AI entry point
├── CLAUDE.md # Claude Code entry + skill definitions
├── GEMINI.md # Gemini CLI entry point
├── .cursorrules # Cursor entry point
├── .clinerules # Cline/Roo Code entry point
├── .windsurfrules # Windsurf entry point
├── .github/
│ └── copilot-instructions.md # GitHub Copilot entry point
│
├── engine/
│ ├── rules.md # Teaching vs challenge mode rules
│ ├── narrator.md # Story, characters (Sarah/Dave/Marcus), tone
│ ├── validation.md # How AI should verify challenge solutions
│ ├── environment-check.sh # Pre-flight Docker/K8s check
│ └── cleanup.sh # Remove all learn-* resources
│
├── curriculum/
│ ├── ch01-containers/
│ │ ├── README.md # Chapter story + learning objectives
│ │ ├── lessons/ # 3 teaching lessons per chapter
│ │ ├── challenges/ # Hands-on tasks + verify.sh scripts
│ │ └── quiz.md # Skip-level assessment questions
│ ├── ch02-image-optimization/
│ ├── ch03-persistence/
│ ├── ch04-networking/
│ ├── ch05-compose/
│ ├── ch06-k8s-intro/
│ └── ch07-k8s-production/
│
└── .player/
└── progress.yaml # Save file — AI reads/writes this
learn-docker-and-k8s/
├── AGENTS.md # 通用AI入口点
├── CLAUDE.md # Claude Code入口 + 技能定义
├── GEMINI.md # Gemini CLI入口点
├── .cursorrules # Cursor入口点
├── .clinerules # Cline/Roo Code入口点
├── .windsurfrules # Windsurf入口点
├── .github/
│ └── copilot-instructions.md # GitHub Copilot入口点
│
├── engine/
│ ├── rules.md # 教学模式和挑战模式的规则
│ ├── narrator.md # 故事、角色（Sarah/Dave/Marcus）、语气设定
│ ├── validation.md # AI验证挑战解决方案的规则
│ ├── environment-check.sh # 启动前Docker/K8s环境检查脚本
│ └── cleanup.sh # 移除所有learn-*开头的资源
│
├── curriculum/
│ ├── ch01-containers/
│ │ ├── README.md # 章节故事 + 学习目标
│ │ ├── lessons/ # 每章3节教学课
│ │ ├── challenges/ # 实操任务 + verify.sh验证脚本
│ │ └── quiz.md # 跳级评估问题
│ ├── ch02-image-optimization/
│ ├── ch03-persistence/
│ ├── ch04-networking/
│ ├── ch05-compose/
│ ├── ch06-k8s-intro/
│ └── ch07-k8s-production/
│
└── .player/
└── progress.yaml # 存档文件 — AI会读写该文件
The Game Engine: Key Files
游戏引擎：核心文件
Defines the two game modes:
- Teaching mode — AI explains concepts, answers questions freely
- Challenge mode — AI gives only progressive hints; never reveals the answer directly
定义了两种游戏模式：
- 教学模式 — AI解释概念，自由回答问题
- 挑战模式 — AI仅提供逐级提示，永远不会直接透露答案
Defines character voices and story tone:
- Sarah — friendly senior DevOps mentor, uses coffee metaphors
- Dave — CTO who breaks things and says "just restart it"
- Marcus — PM who sets impossible deadlines ("demo is at 3")
定义了角色语调和故事风格：
- Sarah — 友好的资深DevOps导师，喜欢用咖啡相关的比喻
- Dave — 总搞坏东西的CTO，口头禅是"重启一下就好了"
- Marcus — 总提不可能 deadlines的PM（"3点就要演示"）
Tells the AI how to run and interpret
scripts — it will execute them in the terminal and parse output to determine pass/fail.
告知AI如何运行和解读
脚本——它会在终端中执行这些脚本，解析输出判断是否通过。
The save file. The AI manages this automatically:
.player/progress.yaml (example structure)
.player/progress.yaml（示例结构）
player:
name: ""
started_at: ""
current:
chapter: 1
lesson: 2
challenge: 1
completed:
chapters: []
challenges: []
hints_used: 0
player:
name: ""
started_at: ""
current:
chapter: 1
lesson: 2
challenge: 1
completed:
chapters: []
challenges: []
hints_used: 0
Challenge Verification Scripts
挑战验证脚本
Each challenge has a
that the AI runs to check the learner's work:
每个挑战都有对应的
，AI会运行该脚本检查学习者的完成情况：
curriculum/ch01-containers/challenges/verify.sh (example pattern)
curriculum/ch01-containers/challenges/verify.sh（示例）
#!/bin/bash
set -e
echo "🔍 Checking Chapter 1 Challenge..."
#!/bin/bash
set -e
echo "🔍 正在检查第1章挑战..."
Check container is running
检查容器是否正在运行
if docker ps --filter "name=learn-api" --filter "status=running" | grep -q "learn-api"; then
echo "✅ Container 'learn-api' is running"
else
echo "❌ Container 'learn-api' not found or not running"
exit 1
fi
if docker ps --filter "name=learn-api" --filter "status=running" | grep -q "learn-api"; then
echo "✅ 容器'learn-api'正在运行"
else
echo "❌ 未找到容器'learn-api'或容器未运行"
exit 1
fi
if docker inspect learn-api | grep -q '"HostPort": "8080"'; then
echo "✅ Port 8080 is mapped correctly"
else
echo "❌ Port mapping incorrect — expected 8080:8080"
exit 1
fi
if docker inspect learn-api | grep -q '"HostPort": "8080"'; then
echo "✅ 端口8080映射正确"
else
echo "❌ 端口映射错误 — 预期为8080:8080"
exit 1
fi
Check endpoint responds
检查端点是否响应
if curl -sf
http://localhost:8080/health > /dev/null; then
echo "✅ API is responding on port 8080"
else
echo "❌ API not responding — check the container logs"
exit 1
fi
echo ""
echo "🎉 Challenge complete! Sarah is proud of you."
if curl -sf
http://localhost:8080/health > /dev/null; then
echo "✅ API在8080端口响应正常"
else
echo "❌ API未响应 — 请检查容器日志"
exit 1
fi
echo ""
echo "🎉 挑战完成！Sarah为你感到骄傲。"
Real Docker Commands Used in the Game
游戏中使用的真实Docker命令
The game teaches these patterns through hands-on challenges:
Ch1 — Run a container with port mapping
第1章 — 运行带端口映射的容器
docker run -d --name learn-api -p 8080:8080 my-api-image
docker run -d --name learn-api -p 8080:8080 my-api-image
Ch2 — Multi-stage build to reduce image size
第2章 — 多阶段构建缩小镜像体积
docker build -t learn-api:optimized .
docker images learn-api # compare sizes
docker build -t learn-api:optimized .
docker images learn-api # 对比体积
Ch3 — Named volume for persistence
第3章 — 命名卷实现数据持久化
docker volume create learn-beans-data
docker run -d -v learn-beans-data:/app/data --name learn-db postgres:15
docker volume create learn-beans-data
docker run -d -v learn-beans-data:/app/data --name learn-db postgres:15
Ch3 — Bind mount for development
第3章 — 开发环境绑定挂载
docker run -d -v $(pwd)/src:/app/src --name learn-dev my-app
docker run -d -v $(pwd)/src:/app/src --name learn-dev my-app
Ch4 — Custom bridge network
第4章 — 自定义桥接网络
docker network create learn-coffee-net
docker run -d --network learn-coffee-net --name learn-api my-api
docker run -d --network learn-coffee-net --name learn-db postgres:15
docker network create learn-coffee-net
docker run -d --network learn-coffee-net --name learn-api my-api
docker run -d --network learn-coffee-net --name learn-db postgres:15
Ch5 — Docker Compose
第5章 — Docker Compose
docker compose up -d
docker compose ps
docker compose logs -f api
docker compose down -v
docker compose up -d
docker compose ps
docker compose logs -f api
docker compose down -v
Ch6 — Kubernetes basics
第6章 — Kubernetes基础
kubectl apply -f curriculum/ch06-k8s-intro/manifests/
kubectl get pods -l app=learn-api
kubectl rollout status deployment/learn-api
kubectl apply -f curriculum/ch06-k8s-intro/manifests/
kubectl get pods -l app=learn-api
kubectl rollout status deployment/learn-api
Ch7 — Horizontal Pod Autoscaler
第7章 — 水平Pod自动扩缩容
kubectl apply -f curriculum/ch07-k8s-production/hpa.yaml
kubectl get hpa learn-api-hpa
kubectl apply -f curriculum/ch07-k8s-production/hpa.yaml
kubectl get hpa learn-api-hpa
Example Docker Compose (Ch5 Pattern)
Docker Compose示例（第5章模式）
curriculum/ch05-compose/docker-compose.yml (game example)
curriculum/ch05-compose/docker-compose.yml（游戏示例）
version: "3.9"
services:
api:
build: ./api
container_name: learn-api
ports:
- "8080:8080"
environment:
- DATABASE_URL=postgresql://user:${DB_PASSWORD}@db:5432/nocappuccino
depends_on:
db:
condition: service_healthy
networks:
- learn-coffee-net
db:
image: postgres:15-alpine
container_name: learn-db
volumes:
- learn-beans-data:/var/lib/postgresql/data
environment:
- POSTGRES_PASSWORD=${DB_PASSWORD}
- POSTGRES_DB=nocappuccino
healthcheck:
test: ["CMD-SHELL", "pg_isready -U postgres"]
interval: 10s
timeout: 5s
retries: 5
networks:
- learn-coffee-net
volumes:
learn-beans-data:
networks:
learn-coffee-net:
driver: bridge
> Note: `DB_PASSWORD` is read from a `.env` file — never hardcoded.
---
version: "3.9"
services:
api:
build: ./api
container_name: learn-api
ports:
- "8080:8080"
environment:
- DATABASE_URL=postgresql://user:${DB_PASSWORD}@db:5432/nocappuccino
depends_on:
db:
condition: service_healthy
networks:
- learn-coffee-net
db:
image: postgres:15-alpine
container_name: learn-db
volumes:
- learn-beans-data:/var/lib/postgresql/data
environment:
- POSTGRES_PASSWORD=${DB_PASSWORD}
- POSTGRES_DB=nocappuccino
healthcheck:
test: ["CMD-SHELL", "pg_isready -U postgres"]
interval: 10s
timeout: 5s
retries: 5
networks:
- learn-coffee-net
volumes:
learn-beans-data:
networks:
learn-coffee-net:
driver: bridge
> 注意：`DB_PASSWORD`会从`.env`文件读取——永远不要硬编码。
---
Example Kubernetes Manifests (Ch6–7 Pattern)
Kubernetes清单示例（第6-7章模式）
curriculum/ch06-k8s-intro/manifests/deployment.yaml
curriculum/ch06-k8s-intro/manifests/deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: learn-api
labels:
app: learn-api
spec:
replicas: 2
selector:
matchLabels:
app: learn-api
template:
metadata:
labels:
app: learn-api
spec:
containers:
- name: api
image: learn-api:latest
ports:
- containerPort: 8080
env:
- name: DB_PASSWORD
valueFrom:
secretKeyRef:
name: learn-db-secret
key: password
readinessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 5
periodSeconds: 10
apiVersion: v1
kind: Service
metadata:
name: learn-api-svc
spec:
selector:
app: learn-api
ports:
- port: 80
targetPort: 8080
type: ClusterIP
apiVersion: apps/v1
kind: Deployment
metadata:
name: learn-api
labels:
app: learn-api
spec:
replicas: 2
selector:
matchLabels:
app: learn-api
template:
metadata:
labels:
app: learn-api
spec:
containers:
- name: api
image: learn-api:latest
ports:
- containerPort: 8080
env:
- name: DB_PASSWORD
valueFrom:
secretKeyRef:
name: learn-db-secret
key: password
readinessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 5
periodSeconds: 10
apiVersion: v1
kind: Service
metadata:
name: learn-api-svc
spec:
selector:
app: learn-api
ports:
- port: 80
targetPort: 8080
type: ClusterIP
Remove all game-created Docker resources safely (uses
prefix convention):
bash
bash engine/cleanup.sh
Or via AI editor:
This removes containers, images, volumes, and networks prefixed with
. It never touches your other Docker resources.
安全移除所有游戏创建的Docker资源（遵循
前缀命名规范）：
bash
bash engine/cleanup.sh
或者在AI编辑器中输入：
该操作会移除所有
前缀的容器、镜像、卷和网络，不会影响你其他的Docker资源。
Key contribution areas:
主要贡献方向：
1. New challenges → curriculum/chXX-*/challenges/
1. 新增挑战 → curriculum/chXX-*/challenges/
2. New chapters → curriculum/ch08-helm/, ch09-argocd/, etc.
2. 新增章节 → curriculum/ch08-helm/、ch09-argocd/等
3. Fix verify.sh → make sure exit codes are correct (0=pass, 1=fail)
3. 修复verify.sh → 确保退出码正确（0=通过，1=失败）
4. New AI editor → add entry point file + update AGENTS.md
4. 适配新的AI编辑器 → 新增入口文件并更新AGENTS.md
5. Translations → keep technical terms (Docker, kubectl) in English
5. 翻译 → 技术术语（Docker、kubectl）保留英文
Each chapter follows this structure:
chXX-topic/
├── README.md # Story hook + learning objectives
├── lessons/
│ ├── 01-concept.md # Teaching content
│ ├── 02-concept.md
│ └── 03-concept.md
├── challenges/
│ ├── README.md # Challenge instructions
│ └── verify.sh # Auto-verification script (chmod +x)
└── quiz.md # 5 questions for /skip-to gate
chXX-主题/
├── README.md # 故事引子 + 学习目标
├── lessons/
│ ├── 01-concept.md # 教学内容
│ ├── 02-concept.md
│ └── 03-concept.md
├── challenges/
│ ├── README.md # 挑战说明
│ └── verify.sh # 自动验证脚本（需授予执行权限chmod +x）
└── quiz.md # /skip-to跳转门槛的5道测验题
"Docker daemon not running"
sudo systemctl start docker
sudo usermod -aG docker $USER # then log out and back in
**"verify.sh permission denied"**
```bash
chmod +x curriculum/ch01-containers/challenges/verify.sh
"Port already in use"
sudo systemctl start docker
sudo usermod -aG docker $USER # 执行后退出并重新登录
**"verify.sh权限被拒绝"**
```bash
chmod +x curriculum/ch01-containers/challenges/verify.sh
"端口已被占用"
Find what's using port 8080
查找占用8080端口的进程
docker ps # check if a previous learn-* container is still running
bash engine/cleanup.sh
**"AI isn't staying in character"**
Remind it: `"You are Sarah from NoCappuccino. Read engine/narrator.md and continue the game."`
**"Progress file is corrupted"**
```bash
rm .player/progress.yaml
docker ps # 检查是否有之前的learn-*容器还在运行
bash engine/cleanup.sh
**"AI没有保持角色设定"**
提醒它：`"你是NoCappuccino的Sarah，读取engine/narrator.md继续游戏。"`
**"进度文件已损坏"**
```bash
rm .player/progress.yaml
Then: /play to restart
然后输入：/play 重新开始
**kubectl not found for Ch6–7**
```bash
**第6-7章提示kubectl未找到**
```bash
Install kind for local Kubernetes
安装kind用于本地运行Kubernetes
brew install kind kubectl # macOS
brew install kind kubectl # macOS
Create a local cluster
创建本地集群
kind create cluster --name learn-k8s
kind create cluster --name learn-k8s