Use Copilot agent mode to finish a web app
Jonas used GitHub Copilot with Claude Sonnet to turn a game-score idea into a working app.
- What they did
- Jonas first tried a GitHub tutorial, but that effort stalled. He then opened a fresh repo for a web app that would keep score for his Partners game group, added a rough outline and his existing JSON records, and handed the work to VS Code agent mode running Claude Sonnet. The agent assembled the score entry flow, a past-results table, stats charts, avatar mood pop-ups, bilingual Danish/English support, and dark mode, while Jonas made only small edits himself.
- How it went
- He finished with an app that worked and put it live on Digital Ocean App Platform. The earlier tutorial path fell apart because the agent kept reverting its own changes, and the hover pop-up ran into similar problems. He also still had leftover agent-generated files to clean up.
- Tips
- Break the job into stages and save working commits often; Jonas wished he had done that more consistently. If the agent starts losing its editor connection or claiming its own port is occupied, expect extra debugging time.
Try it yourself with GitHub Copilot
In [repo or folder], use Copilot agent mode with Claude Sonnet to build [small complete feature or app]. First, give me a brief plan and the files you expect to touch. Then work step by step: implement one piece, run it or test it, fix any issues, and wait for my approval before any irreversible action such as deleting files, sending data, making payments, or submitting anything. At the end, summarize what you changed, what still needs cleanup, and how to run or verify it.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Set up layered agent code reviews
A team had Claude Code write most code, then used multiple review layers to catch bugs before merge.
Talk your way to software for an Apple IIe
Benj Edwards used Claude to discuss a program idea and have it turned into code for an Apple IIe setup.
Use spare AI quota to draft reviewable PRs
A founder pointed an agent at goals and task lists so it could turn leftover Claude and Codex quota into pull requests for review.
Use Copilot CLI subagents for parallel terminal work
A commenter used GitHub Copilot CLI to split repo work across multiple subagents, run commands in parallel, and review the results live.
Split development work into phases before using Copilot
A developer plans features in stages first, then feeds the approved steps to GitHub Copilot instead of asking it to make all architecture decisions at once.
Plan work in a PRD, then assign the build
A developer asks a stronger model for a plan, turns it into tasks, and routes implementation to lower-cost agents.