1-hour workshop: Fail startup clearly when dependencies or services do not start

Problem

The learner setup depends entirely on the startup scripts, but failure handling can report a misleading success state. In app/scripts/start-app.sh, npm install is not checked before npm run dev starts. In app/scripts/start-app.ps1, a failed Node dependency installation emits a warning and continues. Both scripts wait a fixed five seconds and then print both service URLs without checking that Flask and Astro are ready.

This makes common setup failures look like browser or workshop problems and gives independent learners little actionable feedback.

Suggested change

Make both startup scripts stop on dependency installation failures and verify service/process readiness before announcing success.

Acceptance criteria

• A failed Python or Node dependency install exits nonzero with a clear message.
• Failure of either Flask or Astro stops the other managed process.
• URLs are printed only after both services are ready.
• Cleanup still works on Ctrl+C and partial startup failures.
• macOS/Linux and Windows paths have equivalent behavior.

Start with app/scripts/start-app.sh and app/scripts/start-app.ps1, tracing dependency installation, the fixed five-second wait, service URL output, and cleanup paths. Run both startup scripts while simulating failed Python or Node installation and Flask or Astro startup; done means failures are nonzero and clear, managed processes stop together, URLs appear only after both services are ready, and Ctrl+C or partial startup cleanup still works.