Estimated time: 45–60 minutes
In this module, you will go deeper into the commands available in GitHub Copilot CLI. You will learn how to work with shell commands, scripts, and Git operations, and how to give Copilot enough context to get useful responses.
One of the most important ideas when working with any AI assistant is context — the information the tool has available when generating a response.
GitHub Copilot CLI works within your terminal session. It does not automatically read your project files unless you provide them explicitly. The quality of suggestions you receive depends heavily on how clearly you describe your situation.
Less useful prompt:
How do I fix the error?
More useful prompt:
I'm getting a "permission denied" error when running my Python script on Ubuntu. The file is executable but I'm running it as a normal user.
Be specific about your operating system, the tool or language you are using, and any constraints you have.
gh copilot explain accepts any shell command as input and returns a plain-language breakdown.
gh copilot explain "chmod 755 script.sh"gh copilot explain "find . -name '*.py' | xargs grep -l 'import os'"If you see a command in documentation with unfamiliar flags, paste the whole thing:
gh copilot explain "curl -fsSL https://example.com/install.sh | bash"Copilot will break down each component, which helps you understand what you are running before executing it.
Here is what referencing a file in your context looks like:
gh copilot suggest generates shell commands from natural language descriptions.
gh copilot suggest "compress the images folder into a zip file excluding hidden files"gh copilot suggest "undo the last commit but keep the changes staged"gh copilot suggest "rename all .txt files in the current directory to .md"gh copilot suggest "show which process is using port 3000"After Copilot suggests a command, you can:
- Run it directly by copying and pasting
- Ask for a variation by describing what is different about your situation
- Ask for an explanation of the suggested command before running it
A practical workflow is to use suggest to get a command and then immediately use explain to understand it before running it.
For example:
gh copilot suggest "delete all Docker containers that are stopped"
# Copilot suggests: docker rm $(docker ps -aq -f status=exited)
gh copilot explain "docker rm $(docker ps -aq -f status=exited)"
# Copilot explains what each part doesThis gives you confidence that the command does what you expect.
Copilot CLI is particularly useful for Git workflows where you need to remember the exact syntax for less common operations.
Writing commit messages:
gh copilot suggest "write a commit message for changes that add user authentication using JWT tokens"Undoing mistakes:
gh copilot suggest "I accidentally pushed to main, how do I revert the last push"Branch operations:
gh copilot suggest "create a new branch from a specific commit hash"Rebasing:
gh copilot explain "git rebase -i HEAD~3"
By default, each gh copilot command starts a fresh session with no memory of previous requests. If you are working through a multi-step problem, include the relevant context in each prompt.
For example, if you asked Copilot to suggest a deployment script in one command and now want to extend it, include a brief description of what the script already does in your next prompt rather than assuming Copilot remembers it.
When working on a project, you may need Copilot to understand how multiple files relate to each other.
You can provide this context manually by describing the structure in your prompt:
gh copilot suggest "I have a Flask app where routes.py handles HTTP requests and services.py contains the business logic. In services.py, how should I structure a function that validates user input before saving to the database?"You can use Copilot CLI to help write and understand shell scripts.
Generating a script:
gh copilot suggest "write a bash script that backs up the current directory to a timestamped folder"Explaining an existing script:
If you have a script you inherited or downloaded, use explain to understand what it does before running it:
gh copilot explain "$(cat deploy.sh)"Note: Be cautious about passing long or sensitive scripts directly into the explain command. Review the content first and redact any credentials or tokens.
Here are patterns that consistently produce better results:
| Pattern | Example |
|---|---|
| State your OS or environment | "On Ubuntu 22.04, ..." |
| Include the language or tool | "Using Python 3.11, ..." |
| Describe the goal, not just the error | "I want to X, but I'm seeing Y" |
| Give constraints | "without installing additional packages" |
| Specify output format | "output as a single command I can copy and run" |
Exercise 2.1
Use gh copilot explain on a command you have never used or that you find confusing. Write a one-sentence summary of what you learned.
Exercise 2.2
Use gh copilot suggest to get a Git command for one of the following tasks (your choice):
- Squash the last 3 commits into one
- Show a visual graph of the commit history
- Find which commit introduced a specific line of code
Run the command on a test repository.
Exercise 2.3
Try chaining suggest and explain. Get a command suggestion for any task, then use explain on the output. Notice whether the explanation changes how confident you are about running it.
In this module you:
- Learned how context affects the quality of responses
- Used
explainto understand commands in depth - Used
suggestfor shell tasks, Git operations, and scripting - Understood how sessions work and how to provide multi-file context
- Practised writing more specific prompts
Continue to Module 03 — Development Workflows.