Copilot CLI is distributed as a GitHub CLI extension. Extensions are installed with a single command and kept up to date with gh extension upgrade.

Install the Extension

Run the following command to install the GitHub Copilot CLI extension:

gh extension install github/gh-copilot

Verify the Installation

Confirm the extension is installed and check which version is active:

gh copilot --version

You should see output similar to:

gh-copilot version 1.x.x (yyyy-mm-dd)

View Available Commands

Explore what Copilot CLI can do:

gh copilot --help

gh extension upgrade gh-copilot

Authenticate (if prompted)

On first use, Copilot CLI may prompt you to grant additional permissions. Follow the on-screen instructions to complete the OAuth flow in your browser. After authentication, return to your terminal — you are ready to start using Copilot CLI.

gh copilot suggest converts a plain-language description into a shell command. You describe what you want to accomplish, and Copilot returns a command you can copy, run, or refine.

Basic Usage

gh copilot suggest "<describe the task in plain English>"

Choosing a Shell Target

You can tell Copilot which shell environment you are targeting with the --target flag:

gh copilot suggest --target shell "list all files modified in the last 7 days"
gh copilot suggest --target git "show a compact commit log for the last 10 commits"
gh copilot suggest --target gh "list open pull requests in the current repo"

Supported targets are shell, git, and gh. When omitted, Copilot infers the best target from your description.

Interactive Response

After Copilot suggests a command, it presents an interactive menu:

• Copy command to clipboard — Paste it into your terminal manually
• Explain command — Get a plain-language explanation before running it
• Execute command — Run the command immediately in your current shell
• Revise command — Provide additional context to refine the suggestion

Hands-on Exercises

1. Ask Copilot to suggest a command to list all files modified in the last 7 days:

   gh copilot suggest "list all files modified in the last 7 days"

2. Ask for a command to find all files larger than 100 MB in your home directory:

   gh copilot suggest "find all files larger than 100MB in my home directory"

3. Use the --target git flag to ask for a Git command:

   gh copilot suggest --target git "show a compact graph of the last 10 commits across all branches"

4. Try the interactive Revise command option to add more specificity to one of the suggestions

gh copilot explain takes any shell command and returns a clear, plain-language explanation of what it does — including each flag and argument. This is invaluable when you encounter unfamiliar commands in scripts, documentation, or Stack Overflow answers.

Basic Usage

gh copilot explain "<command to explain>"

Hands-on Exercises

1. Explain an interactive Git rebase:

   gh copilot explain "git rebase -i HEAD~3"

2. Explain a complex find command:

   gh copilot explain "find . -type f -name '*.log' -mtime +30 -delete"

3. Explain a ps pipeline:

   gh copilot explain "ps aux | grep node | awk '{print $2}' | xargs kill"

4. Paste a command from a script you are unfamiliar with and ask Copilot to explain it

Piping Output to Explain

You can pipe command output directly to Copilot explain for contextual help:

cat error.log | gh copilot explain

This is useful for understanding error messages or log output without copying and pasting manually.

Typing gh copilot suggest or gh copilot explain every time is verbose. Setting up short shell aliases — ghcs (GitHub Copilot Suggest) and ghce (GitHub Copilot Explain) — makes Copilot CLI feel like a natural part of your workflow.

Automatic Alias Setup

Copilot CLI can generate the alias configuration for your shell automatically:

gh copilot alias -- bash

Replace bash with your shell (zsh, fish, or powershell). The command prints shell-specific configuration to standard output.

Bash / Zsh

Add the aliases to your shell configuration file:

# Add to ~/.bashrc or ~/.zshrc
eval "$(gh copilot alias -- bash)"

For Zsh, replace bash with zsh:

# Add to ~/.zshrc
eval "$(gh copilot alias -- zsh)"

After editing the file, reload your shell configuration:

source ~/.bashrc    # or ~/.zshrc

Fish Shell

For Fish, add the configuration to your Fish config file:

# Run once to add to your Fish config
gh copilot alias -- fish | source

To make it permanent, add the following line to ~/.config/fish/config.fish:

gh copilot alias -- fish | source

PowerShell

For PowerShell on Windows, add the following to your PowerShell profile ($PROFILE):

Invoke-Expression (gh copilot alias -- powershell | Out-String)

Verify the Aliases Work

Open a new terminal session (or reload your shell config) and test the aliases:

ghcs "list all running Docker containers"
ghce "docker ps -a --format 'table {{.Names}}\t{{.Status}}'"

Now it is time to put everything together. Complete all three tasks below using only Copilot CLI suggestions — no manual searches or documentation. For each task, use ghcs (or gh copilot suggest) to generate the command, then use ghce (or gh copilot explain) to verify you understand what it does before running it.

Task 1: Repository Hygiene

Use Copilot CLI to find and remove all .DS_Store files from a Git repository (without deleting the files from disk, only from Git tracking).

• Hint: You will need a git command and a follow-up shell command
• Use ghcs --target git for the Git-specific part

ghcs --target git "remove all .DS_Store files from git tracking without deleting them from disk"

Task 2: Log Analysis

Your application generates log files in /var/log/app/. Use Copilot CLI to:

1. Count the number of ERROR lines across all .log files in that directory
2. Extract the 10 most recent unique error messages, sorted by frequency

• Use ghcs for each sub-task separately
• Use ghce to understand the pipeline before running it

Task 3: Docker Cleanup

Use Copilot CLI to generate commands that:

1. List all Docker containers that exited more than 24 hours ago
2. Remove all stopped containers, unused networks, dangling images, and build cache in a single command

• Always use ghce to review destructive commands before executing them

Reflection

After completing the three tasks, consider:

• How did using Copilot CLI compare to your usual workflow for looking up commands?
• Which task benefited most from the --target flag?
• Did any suggestion require refinement? What additional context helped?

Make the most of GitHub Copilot CLI with these tips and best practices gathered from real-world use.

Writing Effective Prompts

• Be specific — Include file types, time ranges, and output format in your description. "list log files" is less effective than "list all .log files in /var/log modified in the last 24 hours, sorted by size"
• Mention the tool or ecosystem — Saying "a git command to …" or "a Docker command to …" helps Copilot choose the right tool
• Use the --target flag — Setting --target git or --target gh scopes suggestions and makes them more accurate
• Iterate with "Revise" — The first suggestion may not be exactly what you need. Use the interactive Revise option to add constraints or change the approach

Safety and Verification

• Always explain before running — Particularly for destructive commands (rm -rf, docker system prune, database drops). Use ghce to review what a command does before executing it
• Prefer dry-run flags — Many tools support --dry-run or -n flags. Ask Copilot to include them: "list the files that would be deleted, without actually deleting them"
• Review multi-piped commands carefully — Long pipelines can have unintended side effects. Break them into stages if you are unsure

Integrating Into Your Workflow

• Pair with your shell history — After Copilot suggests a command and you run it successfully, it is added to your shell history for easy recall with the up arrow
• Use it for learning — Ask Copilot to suggest a command you already know, then use explain to see if there is a more efficient variant you were not aware of
• Combine with scripts — Use Copilot CLI interactively to prototype shell scripts, then copy the final commands into a .sh file
• Use the aliases — Once you have set up ghcs and ghce, using Copilot CLI becomes as natural as running any other command

Useful Reference Commands

# Install / update
gh extension install github/gh-copilot
gh extension upgrade gh-copilot

# Suggest a shell command
gh copilot suggest "your task description"
ghcs "your task description"                 # with alias

# Suggest with explicit target
gh copilot suggest --target git "your task"
gh copilot suggest --target gh "your task"

# Explain a command
gh copilot explain "command to explain"
ghce "command to explain"                    # with alias

# Set up aliases (bash/zsh/fish/powershell)
eval "$(gh copilot alias -- bash)"

Additional Resources

• GitHub Copilot CLI Reference Manual
• Using GitHub Copilot in the command line
• GitHub CLI — cli.github.com
• gh-copilot extension on GitHub

🎉 Congratulations!

You have completed the Copilot CLI workshop activity! You can now:

• Install and authenticate the GitHub Copilot CLI extension
• Use gh copilot suggest to get AI-powered shell command suggestions
• Use gh copilot explain to understand unfamiliar commands
• Use short aliases (ghcs / ghce) for faster access
• Apply Copilot CLI to real-world terminal tasks safely and effectively

Next Steps

Continue your Path 3 journey by exploring the other extended capabilities:

• Copilot SDK — Build your own Copilot extensions
• Agent Skills — Define reusable Copilot behaviors
• Sub-Agents — Coordinate multi-agent workflows
• Copilot Memory — Persist context across conversations