Here we'll take the Issues you created at the end of Part 6 and have the CLI implement them in parallel via /fleet.
13.0 โ (If you're not on a Codespace) Install Copilot CLI
The workshop Codespace already has Copilot CLI pre-installed, so you can skip this step. Only if you're working locally and don't have it yet, install it with one of the following:
# npm (all platforms / requires Node.js 22+)
npm install -g @github/copilot
# Homebrew (macOS / Linux)
brew install copilot-cli
# Install script (macOS / Linux)
curl -fsSL https://gh.io/copilot-install | bashKey commands
| Command | Description |
| ------------- | ------------------------------------------------------ |
| /help | List available interactive commands |
| /model | Pick the AI model (Claude / GPT, etc.) |
| Shift+Tab | Switch mode (plan / agent) |
| /ide | Connect to an IDE workspace |
| /allow-all | Allow tool execution without confirmation (Autopilot) |
| /fleet | Run multiple agents in parallel |
| /tasks | Check the status of running fleet tasks |
| /context | Show context window token usage |
| /compact | Summarize conversation history to free context |
| /chronicle | Analyze session history and usage |
| /resume | Resume the previous session |
13.1 โ Launch Copilot CLI
First, grab the latest code. The Cloud Agent and Code Review should have merged PRs by now, so update your local branch too:
git checkout main
git pullThen launch Copilot CLI from the VS Code terminal:
copilot13.2 โ Pick a model and enable Autopilot
Inside the CLI, run:
/allow-all/modelPick the most capable model available (e.g., Claude Opus 4.6).
13.3 โ Implement Issues in parallel with /fleet
Have the CLI implement the 3โ5 Issues you created in Part 6 in parallel. The prompt also asks for a rubber-duck review on each implementation at the end.
/fleet Take every open Issue in this repository labeled `bug` or `enhancement` and
assign one fleet agent per Issue to implement them in parallel.
Each agent should follow this flow:
1. Create a branch matching the Issue number (`fix/<number>-<slug>`)
2. Implement the fix
3. Verify it runs with `pnpm dev`
4. Commit and open a PR (include `Closes #<number>` in the body)
Once all PRs are open, run a rubber-duck review on each implementation.โฑ ~10โ15 minutes
The CLI spawns multiple fleet agents, each working on its own branch through implementation and PR creation autonomously. Progress streams in your terminal, and the rubber-duck review is printed at the end.
To check progress during and after execution:
/tasksThis lists the status of each fleet agent (running / done / failed) along with the current step.
13.4 โ Analyze your usage with Chronicle
Finally, look back on your Copilot CLI usage during the workshop with Chronicle:
/experimentalchronicle