/deploy

Deploy application to Buddy Sandbox. Use the sandbox skill for deployment procedures and best practices.

Arguments:

• $1: Optional sandbox name (default: generate from directory/timestamp)
• $2: Optional path to deploy (default: current directory)

Workflow

1. Verify Authentication

Check authentication with:

bdy workspace ls

If fails, inform user: "Run bdy login in a separate terminal (AI agents cannot perform interactive login). See bdy-auth skill for details."

2. Analyze Project

Examine the directory to understand:

• What application(s) need deployment
• Required dependencies and runtime
• How to start the application
• What port(s) to expose

If user specified a subdirectory in $2, deploy from that path. Otherwise use current directory.

Note: Repository may contain multiple deployable applications (e.g., blog + admin panel). Ask user which to deploy if unclear.

3. Deploy Using Sandbox Skill

Use the sandbox skill procedures to:

• Create sandbox with appropriate resources
• Copy application files with bdy sandbox cp --silent
• Install dependencies
• Start application in detached mode
• Expose HTTP endpoint

The sandbox skill contains complete procedures for deployment.

4. Display Results

Show:

• Sandbox identifier
• Public URL
• How to view logs: bdy sandbox command logs [id] --last
• Management commands (stop, destroy)

Important Notes

• Always use --silent with bdy sandbox cp
• Applications must bind to 0.0.0.0, not 127.0.0.1
• Ubuntu 24.04 requires venv for Python pip packages
• Endpoints served via HTTPS proxy

Error Handling

If deployment fails:

• Show error from bdy CLI
• Check logs with bdy sandbox command logs [id] --last
• For auth errors: direct user to run bdy login in separate terminal
• For name conflicts: suggest unique name or destroy existing sandbox