Great question — answering as the UphillTech team:
I just tried this with the prompt from your question and model GPT-4.1 (Preview). For me it's working. I had to watch for confirmation prompts in the terminal (e.g. npx express-generator server --no-view asked if it was ok to proceed ) and in GitHub Copilot chat I had to say "ok" for it to take the next step at different times. It wasn't super clear that this is what needed to happen, though. So maybe it's not so much a model selection issue as much as the UX not making it easy enough for the user to find where they need to intervene.
Full answer by Thomas Mechelke: https://stackoverflow.stackexchange.com/a/79581562 (CC BY-SA 4.0).
Need it done for your business? UphillTech.