Sandbox Stable logo

Sandbox Stable

OrganizationPopular
cloudflare
sandbox-stable

Build or maintain Cloudflare Sandbox apps on the stable @cloudflare/sandbox package. Use sandbox-next for preview apps and sandbox-migrate-to-next for stable-to-preview migrations.

Overview

Publishercloudflare
Repositoryskills
Skill namesandbox-stable
Stars
2.8K
Forks
277
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by cloudflare on GitHub. Read the source before you install it.

Installation

Install the Sandbox Stable AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/cloudflare/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/sandbox-stable .claude/skills/sandbox-stable
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Sandbox Stable in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Sandbox Stable on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Sandbox Stable is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Sandbox SDK — stable package

Isolated Linux environments on Cloudflare Containers, driven from Workers.

Prefer the main Sandbox docs and installed stable types over memory. This skill is a gate, a contract, and a retrieval map—not a full manual.

This line is the current stable default npm package. The main Sandbox documentation describes it. Existing apps can stay here and keep shipping.

We recommend new projects on @cloudflare/sandbox@next with sandbox-next. When you can, plan a move with sandbox-migrate-to-next so you are ready when 1.0 becomes the stable release. Do not force that port unless the user asks.

1. Gate — confirm the package line

Before writing code, inspect the app:

CheckMust match
npm dependencyDefault @cloudflare/sandbox (not @next / preview tags)
Container imageMatching stable image (not cloudflare/sandbox:next)
If you find…Action
@cloudflare/sandbox@next or a next imageStop. Load sandbox-next.
User wants to port to 1.0 / @nextStop. Load sandbox-migrate-to-next. Do not half-apply preview APIs on a stable package.
Only cleaning deprecated stable APIsStay here; use the 2026 deprecation guide. That is not a move to @next.

Never mix a stable Worker package with an @next container image (or the reverse).

Skills install: Agent setup · cloudflare/skills

2. Contract — non-negotiables

  • await sandbox.exec(command) takes a command string and resolves when the command finishes, with buffered stdout / stderr / exitCode (and related fields).
  • Long-running and streaming work use the stable command APIs (startProcess, execStream, and related helpers)—not the @next single-handle model. Open the Commands docs; do not invent @next output() handles on stable.
  • Sessions can preserve working directory and environment across commands (default session / enableDefaultSession, createSession). See Sessions docs when state must carry across calls.
  • Interactive browser terminals often use sandbox.terminal(request) and session/xterm helpers on stable—not preview createTerminal unless the package is @next.
  • Prefer RPC transport when using tunnels or large/binary streaming. HTTP/WebSocket transports are deprecated (cleanup guide below).
  • Files, mounts, ports, tunnels, backups, lifecycle, and interpreter: use main docs for signatures; trust installed stable types.
  • Non-secret config in sandbox env; live credentials in the Worker. Use outbound handlers when processes call external APIs.
  • Production preview hostnames need wildcard DNS on a custom domain when using those URL patterns.
  • Do not apply @next argv/process.output() APIs while the dependency is still stable.
  • Self-deployed bridge stays on the stable package and image. Bridge

Minimal shape:

ts
import { getSandbox, proxyToSandbox, Sandbox } from "@cloudflare/sandbox";

export { Sandbox };

const sandbox = getSandbox(env.Sandbox, "user-123");
const result = await sandbox.exec('python3 -c "print(2 + 2)"');
// result.stdout, result.exitCode, result.success

3. Retrieve — open the doc for the task

Fetch the page before implementing. Installed stable types win over guesses.

You need to…Open
OrientSandbox overview
First Worker, template, DockerGet started
exec, streaming, background processesCommands API · Execute commands · Background processes · Streaming output
Sessions / shell state across commandsSessions concept · Sessions API
getSandbox options, sleep, destroyLifecycle API · Sandbox options
Env varsEnvironment variables
FilesFiles API · Manage files · File watching
Buckets / mountsStorage API · Mount buckets
BackupsBackups API · Backup and restore
Ports, preview URLs, exposePorts API · Expose services
TunnelsTunnels API
Proxy / Workers connectionsProxy requests · Workers connections
Browser / PTY terminalTerminal API · Terminal concept · Browser terminals
Code interpreterInterpreter API · Code execution
Git in the sandboxGit workflows
Secrets / egressOutbound traffic
WebSocketsWebSocket connections
Docker-in-DockerDocker in Docker
Production deployProduction deployment
Containers conceptContainers
How-to indexGuides
API indexAPI reference
Deprecated APIs while staying on stable2026 deprecation guide
Self-deployed bridgeBridge · Bridge HTTP API
Examples (stable/main)examples on GitHub
New work on 1.0 previewsandbox-next · 1.0 preview
Port existing app to @nextsandbox-migrate-to-next · Migrate

Deprecated-API cleanup (stay on stable)

Update package + matching image first, then follow the guide. Typical search:

sh
rg 'SANDBOX_TRANSPORT|transport:|exposePort\(|enableDefaultSession|execStream\(|readFileStream|writeFileStream'

This path does not switch you to @next.

4. Before you ship

  • Worker package and container image on the same stable line
  • Typecheck against installed stable types
  • No live secrets in sandbox env
  • If using deprecated transports/helpers, finish or track 2026 deprecation cleanup
  • When the team is ready for 1.0, use sandbox-migrate-to-next—do not force cutover unprompted

Frequently asked questions

What does the Sandbox Stable AI skill do?

Build or maintain Cloudflare Sandbox apps on the stable @cloudflare/sandbox package. Use sandbox-next for preview apps and sandbox-migrate-to-next for stable-to-preview migrations.

Why use Sandbox Stable on TypingMind?

Because you install it once and use it with any model. Sandbox Stable is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Sandbox Stable in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/cloudflare/skills/tree/main/skills/sandbox-stable. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Sandbox Stable?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Sandbox Stable?

As many as you like. As long as a model supports skills, you can use Sandbox Stable with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Sandbox Stable AI skill free?

Yes. It is published on GitHub by cloudflare under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇