The situation
An AI coding tool mentions a terminal, a command, or the current folder. That can sound like a separate technical world when all you wanted was to make progress on your project.
The idea in one paragraph
The command line is a text-based control surface for your computer. You enter a small instruction, the computer runs it in a chosen folder, and it reports what happened. For AI-assisted work, its value is not clever commands: it gives you a clear, repeatable place to inspect a project and run the checks the project already provides.
| Part | Plain meaning | Useful question |
|---|---|---|
| Folder | The project location | Am I in the right project? |
| Command | One requested action | What will this do? |
| Output | The computer's reply | Did it finish or fail? |
How it actually works
A terminal starts in one folder and runs commands relative to it. A command may only display information, or it may change files, install software, contact a service, or publish something. Read the command and its output before treating either as routine. An agent can use the same interface, but its confidence does not make the action safer or the result correct.
What this changes for you
- Confirm the project folder before asking an agent to act.
- Prefer the project's documented commands over invented replacements.
- Pause when a command would change, delete, send, or publish something you did not expect.
Where it breaks
Terminal output can be terse, and copied commands can assume a different operating system, project, or permission level. Do not paste a command just because it looks familiar. Ask what it does, what it can change, and how to verify the result.
Terms used on this page
Command line: A text interface for asking a computer to perform an action.
Terminal: The application window that gives access to a command line.
Working folder: The folder a command treats as its starting location.
Read next
Continue to Instruction files.