skills/sandbox/SKILL.md
Deploy and test applications in Buddy Sandbox cloud environments. Use when user asks about "deploy app", "create sandbox", "test in cloud", "isolated environment", "remote environment", "run app in sandbox", or mentions deploying, testing, or running applications in cloud sandboxes.
npx skillsauth add sztwiorok/buddy-plugin sandboxInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
On-demand cloud environments for deploying and testing applications with public HTTP/TCP endpoints.
Ubuntu 24.04.4 LTS Sandboxes come with the following software preinstalled:
No need to install these — they're ready to use immediately after sandbox creation.
If any bdy command returns Token not provided or Not logged in, ask the user to run bdy login in a separate terminal (interactive browser auth — AI cannot do it).
Basic creation:
bdy sandbox create -i my-app --resources 2x4 --wait-for-running
Options:
1x2, 2x4, 4x8, 8x16, 12x24 (CPUxRAM)--wait-for-configured - wait until install commands completebdy sandbox cp ./src my-app:/app > /dev/null 2>&1
Redirect output to /dev/null - without it, file copy floods stdout and breaks execution.
Commands run in background by default.
With --wait for operations that must complete before next step, logs are visible in output:
bdy sandbox exec command my-app "cd /app && npm install" --wait
Without --wait for long-running processes (servers):
bdy sandbox exec command my-app "cd /app && npm start"
Check status and logs:
bdy sandbox exec list my-app # list commands
bdy sandbox exec logs my-app <command-id> # view logs
bdy sandbox exec logs my-app <command-id> --wait # wait for completion and show logs
Creating config files: Write locally, then copy - avoids escaping issues with heredoc through exec:
# 1. Write file locally
cat > config.json << 'EOF'
{"host": "localhost", "port": 3000}
EOF
# 2. Copy to sandbox
bdy sandbox cp config.json my-app:/app/ > /dev/null 2>&1
bdy sandbox endpoint add my-app -n web -e 3000
Application MUST bind to 0.0.0.0 (not 127.0.0.1 or localhost) - otherwise endpoint won't work.
With authentication:
bdy sandbox endpoint add my-app -n web -e 3000 --auth BASIC --username admin --password secret
Check endpoint URL:
bdy sandbox endpoint list my-app
Reverse proxy: Apps run behind reverse proxy. Handle X-Forwarded-Proto for correct HTTPS detection:
// Express.js
app.set('trust proxy', true);
// PHP
if ($_SERVER['HTTP_X_FORWARDED_PROTO'] === 'https') $_SERVER['HTTPS'] = 'on';
# Copy file from sandbox
bdy sandbox cp my-app:/app/result.txt ./result.txt > /dev/null 2>&1
# Copy directory from sandbox
bdy sandbox cp my-app:/app/output ./output > /dev/null 2>&1
When destination already exists, use --merge or --replace:
# Replace existing file/directory
bdy sandbox cp my-app:/app/results ./results --replace > /dev/null 2>&1
# Merge into existing directory
bdy sandbox cp my-app:/app/results ./results/ --merge > /dev/null 2>&1
Tip: Run bdy sandbox cp --help for detailed examples of merge/replace behavior with files and directories.
tools
Deploy apps to Buddy sandboxes, publish artifacts, create tunnels, manage domains, configure distributions, and run CI/CD pipelines using `bdy` CLI. Triggers on: deploy, sandbox, tunnel, expose, artifact, static site, publish, domain, distribution, pipeline, CI/CD.
development
This skill should be used when the user asks to "expose localhost", "create tunnel", "share local server", "test webhooks locally", "tunnel to localhost", "make local dev accessible", "public URL for localhost", or mentions exposing local services, testing webhooks, or sharing development servers.
development
Deploy static websites to Buddy Packages. Use when user asks to "deploy static site", "host static files", "publish website", "upload build artifacts", or mentions deploying pre-built HTML/CSS/JS files, SPA builds, or static exports.
data-ai
Delegate AI tasks to Claude agents running in isolated Buddy Sandbox environments. Use when user asks to "delegate task", "run task in sandbox", "parallel agent execution", "isolated AI task", "YOLO mode", "sub-agent", "compare solutions", "multi-agent", or mentions running AI tasks in sandboxes.