Cloud & Deploy Skills
Cloud infrastructure provisioning, serverless deployment, edge computing, and platform-as-a-service integrations. Deploy to any major cloud provider directly from CLI-JAW with a single natural-language command.
Overview
Cloud & Deploy skills transform CLI-JAW into a unified deployment control plane. Instead of memorizing CLI flags for each platform, describe what you want in plain language and the appropriate skill handles provider-specific configuration, build steps, and deployment orchestration.
"Vercel에 배포해줘" or "Deploy this to Cloudflare Workers" and CLI-JAW selects the right skill, detects your framework, and runs the full deployment pipeline.
Skill Catalog
| Skill | Provider | Description |
|---|---|---|
cloudflare-deploy | Cloudflare | Workers, Pages, KV, R2, D1, AI, WAF, Tunnel, Spectrum. Full Wrangler integration. |
agents-sdk | Cloudflare | Build stateful AI agents on Workers: Durable Objects state, WebSocket real-time, scheduled tasks, MCP servers. |
durable-objects | Cloudflare | Stateful coordination primitives: chat rooms, multiplayer, booking systems, SQLite storage, alarms. |
aws-skills | AWS | Lambda, S3, EC2, ECS, CloudFront, API Gateway, DynamoDB, SQS, SNS, IAM policy generation. |
terraform | Multi-cloud | Infrastructure-as-code: plan, apply, state management, module authoring, drift detection. |
vercel-deploy | Vercel | Zero-config deploys for Next.js, SvelteKit, Nuxt, Remix. Preview URLs, edge functions, environment variables. |
netlify-deploy | Netlify | Jamstack deploys, serverless functions, form handling, identity, edge handlers. |
render-deploy | Render | Web services, static sites, cron jobs, databases, private services, blueprints. |
mcp-builder | Any | Build and deploy Model Context Protocol servers for tool integrations. |
jupyter-notebook | Local / Remote | Notebook execution, kernel management, cell manipulation, output capture. |
Framework Detection
Cloud skills automatically detect your project framework and apply the optimal build configuration. Supported frameworks:
| Framework | Detection Signal | Default Target |
|---|---|---|
| Next.js | next.config.* | Vercel / Cloudflare Pages |
| Nuxt | nuxt.config.* | Vercel / Netlify |
| SvelteKit | svelte.config.js | Vercel / Cloudflare Pages |
| Remix | remix.config.* | Vercel / Cloudflare Workers |
| Astro | astro.config.* | Netlify / Cloudflare Pages |
| Hono | hono in deps | Cloudflare Workers |
| Express / Fastify | express / fastify in deps | Render / AWS Lambda |
| Static HTML | index.html at root | Cloudflare Pages / Netlify |
Cloudflare Ecosystem
Three dedicated skills cover the Cloudflare platform in depth. They bias toward retrieval from official Cloudflare docs over pre-trained knowledge.
cloudflare-deploy
The umbrella skill for all Cloudflare products. Handles Workers, Pages, storage (KV, D1, R2), AI (Workers AI, Vectorize), networking (Tunnel, Spectrum), security (WAF, DDoS), and infrastructure-as-code via Terraform or Pulumi.
# Deploy a Worker
"이 코드를 Cloudflare Worker로 배포해줘"
# Create a KV namespace and bind it
"KV 네임스페이스 만들고 Worker에 바인딩해줘"
# Set up R2 bucket with custom domain
"R2 버킷 생성하고 커스텀 도메인 연결해줘"
agents-sdk
Build AI agents that run on Cloudflare Workers with durable state, real-time WebSocket communication, and scheduled task execution.
# Create a stateful chat agent
"Cloudflare Agents SDK로 채팅 에이전트 만들어줘"
# Add WebSocket real-time support
"실시간 WebSocket 통신 추가해줘"
# Build an MCP server on Workers
"MCP 서버를 Workers에 만들어줘"
durable-objects
Stateful coordination for multiplayer, booking, rate limiting, and more. Covers RPC methods, SQLite storage, alarms, WebSockets, and Vitest testing.
# Create a multiplayer game room
"Durable Objects로 멀티플레이어 게임룸 만들어줘"
# Add SQLite storage to a DO
"DO에 SQLite 스토리지 추가해줘"
AWS Skills
The aws-skills module covers the most commonly used AWS services through natural-language commands.
"Lambda에 S3 읽기 권한 추가해줘" and CLI-JAW generates the minimal IAM policy document, attaches it to the execution role, and validates the change.
# Create a Lambda function
"AWS Lambda 만들어줘, Node.js 20 런타임으로"
# Set up API Gateway + Lambda
"API Gateway로 REST API 만들고 Lambda에 연결해줘"
# Deploy to ECS with Fargate
"Docker 이미지 빌드해서 ECS Fargate에 배포해줘"
# Create S3 bucket with CloudFront CDN
"S3 정적 호스팅이랑 CloudFront 설정해줘"
| Service | Supported Operations |
|---|---|
| Lambda | Create, deploy, update, add triggers, set env vars, configure layers |
| S3 | Create bucket, set CORS, enable static hosting, lifecycle rules |
| EC2 | Launch instances, security groups, key pairs, AMI management |
| ECS | Task definitions, services, Fargate, ECR image push |
| CloudFront | Distributions, origins, cache policies, invalidations |
| API Gateway | REST and HTTP APIs, routes, integrations, authorizers |
| DynamoDB | Table creation, GSI, query/scan, TTL configuration |
| SQS / SNS | Queue/topic creation, subscriptions, dead-letter queues |
Terraform
The terraform skill provides infrastructure-as-code across any cloud provider. It generates HCL, manages state, and handles the plan/apply lifecycle.
# Generate Terraform for a full-stack app
"Next.js 앱을 위한 Terraform 설정 만들어줘 — Vercel + PlanetScale"
# Plan and apply changes
"Terraform plan 돌려줘"
"변경사항 확인했으니 apply 해줘"
# Import existing resources
"기존 AWS 리소스 Terraform state로 import 해줘"
terraform plan first and requires explicit confirmation before apply. Destructive changes (resource deletion) trigger an additional warning with a diff summary.
Platform-as-a-Service
Vercel, Netlify, and Render skills provide streamlined deployments for web applications with minimal configuration.
vercel-deploy
# Deploy with preview URL
"Vercel에 배포해줘"
# Set environment variables
"Vercel에 DATABASE_URL 환경변수 설정해줘"
# Promote preview to production
"프리뷰 배포를 프로덕션으로 올려줘"
netlify-deploy
# Deploy a static site
"Netlify에 빌드하고 배포해줘"
# Configure serverless functions
"Netlify Functions 추가해줘"
# Set up form handling
"Netlify Forms로 문의 폼 연결해줘"
render-deploy
# Deploy a web service
"Render에 Express 서버 배포해줘"
# Create a cron job
"Render에 매일 자정에 실행되는 크론잡 만들어줘"
# Set up a managed database
"Render PostgreSQL 데이터베이스 만들어줘"
MCP Builder
The mcp-builder skill creates Model Context Protocol servers that expose tools, resources, and prompts to any MCP-compatible client.
# Create an MCP server
"Notion API를 MCP 서버로 감싸줘"
# Add tools to an existing server
"MCP 서버에 검색 도구 추가해줘"
# Deploy MCP server to Workers
"MCP 서버를 Cloudflare Workers에 배포해줘"
Jupyter Notebook
The jupyter-notebook skill manages notebook execution, kernel lifecycle, and cell-level editing for data science and experimentation workflows.
# Execute a notebook
"이 노트북 전체 실행해줘"
# Add a cell and run it
"pandas로 CSV 로딩하는 셀 추가해줘"
# Convert notebook to HTML report
"노트북을 HTML 보고서로 변환해줘"
Configuration
Cloud skills use credentials from standard provider configuration. No additional CLI-JAW config is needed if you already have provider CLIs set up.
| Provider | Credential Source | Required Setup |
|---|---|---|
| Cloudflare | ~/.wrangler/ or CLOUDFLARE_API_TOKEN | wrangler login |
| AWS | ~/.aws/credentials or env vars | aws configure |
| Vercel | ~/.vercel/ | vercel login |
| Netlify | ~/.netlify/ | netlify login |
| Render | RENDER_API_KEY | API key from dashboard |
| Terraform | Provider-specific | Per-provider auth |
Common Patterns
Multi-stage deployment
Combine skills for complex workflows. CLI-JAW chains the steps automatically:
# Build, test, and deploy
"테스트 돌리고 통과하면 Vercel에 배포해줘"
# Multi-provider setup
"프론트엔드는 Cloudflare Pages, 백엔드는 AWS Lambda로 배포해줘"
Environment promotion
# Promote staging to production
"스테이징 확인했으니 프로덕션으로 승격해줘"
# Rollback
"이전 버전으로 롤백해줘"
Infrastructure + deploy in one flow
# Full-stack provisioning
"Terraform으로 인프라 구성하고 앱도 배포해줘"
Error Handling
Cloud skills provide structured error recovery:
- Auth failures — Detects expired tokens and prompts re-authentication with the correct provider CLI command.
- Build failures — Parses build logs, identifies the root cause, and suggests fixes before retrying.
- Quota limits — Reports which resource limit was hit and suggests cleanup or plan upgrade.
- Network timeouts — Automatic retry with exponential backoff for transient failures.