2. SKILL.md 格式

2.1 基本结构

---
name: deployment
description: Deploy applications to production using Docker and Kubernetes
triggers:
  - deploy
  - deployment
  - production
  - k8s
  - kubernetes
version: 1.0.0
author: team@example.com
---

# Deployment Skill

## When to Use

Use this skill when you need to:
- Deploy applications to production
- Update existing deployments
- Rollback deployments
- Check deployment status

## Prerequisites

- Docker installed
- kubectl configured
- Access to production cluster
- CI/CD pipeline passing

## Workflow

### 1. Pre-deployment Checks

\`\`\`bash
# Run tests
npm test

# Check linting
npm run lint

# Build locally to verify
npm run build
\`\`\`

### 2. Build Docker Image

\`\`\`bash
# Build image
docker build -t myapp:${VERSION} .

# Tag for registry
docker tag myapp:${VERSION} registry.example.com/myapp:${VERSION}

# Push to registry
docker push registry.example.com/myapp:${VERSION}
\`\`\`

### 3. Update Kubernetes Deployment

\`\`\`bash
# Update deployment
kubectl set image deployment/myapp myapp=registry.example.com/myapp:${VERSION}

# Wait for rollout
kubectl rollout status deployment/myapp

# Verify pods
kubectl get pods -l app=myapp
\`\`\`

### 4. Post-deployment Verification

\`\`\`bash
# Check health endpoint
curl https://myapp.example.com/health

# Check logs
kubectl logs -l app=myapp --tail=50

# Run smoke tests
npm run test:smoke
\`\`\`

## Rollback Procedure

If deployment fails:

\`\`\`bash
# Rollback to previous version
kubectl rollout undo deployment/myapp

# Verify rollback
kubectl rollout status deployment/myapp
\`\`\`

## Common Issues

### Issue: Image pull error

**Symptom**: Pods stuck in ImagePullBackOff

**Solution**:
1. Verify image exists: `docker pull registry.example.com/myapp:${VERSION}`
2. Check registry credentials: `kubectl get secret regcred`
3. Verify image tag is correct

### Issue: Health check failing

**Symptom**: Pods restarting repeatedly

**Solution**:
1. Check application logs: `kubectl logs <pod-name>`
2. Verify environment variables: `kubectl describe pod <pod-name>`
3. Test health endpoint locally

## References

- [Kubernetes Deployment Guide](./references/k8s-deployment.md)
- [Docker Best Practices](./references/docker-best-practices.md)
- [Rollback Procedures](./references/rollback-procedures.md)

2.2 YAML Frontmatter 详解

---
# 必需字段
name: skill-name                    # 唯一标识符，kebab-case
description: Short description      # 简短描述，用于自动检测

# 触发器（可选但推荐）
triggers:                           # 关键词列表
  - keyword1
  - keyword2
  - phrase with spaces

# 元数据（可选）
version: 1.0.0                      # 语义化版本
author: email@example.com           # 作者
tags:                               # 标签
  - deployment
  - kubernetes
  - production

# 依赖（可选）
dependencies:                       # 依赖的其他 skills
  - docker-skill
  - testing-skill

# 权限（可选）
permissions:                        # 需要的权限
  - network
  - git_write

# 配置（可选）
config:                             # 可配置项
  registry: registry.example.com
  cluster: production
---

---