Skip to content
explainroo
Working with AI agents

How to Use Claude Code

Updated 3 min read

With explainroo How to Prompt a Coding Agent to Make an Explainer Video
On this page
  1. Install Claude Code and start a session
  2. Give Claude a clear task
  3. Review changes before relying on them
  4. Keep useful sessions and project guidance

If you need help changing or understanding a software project, Claude Code works with the project files on your computer. It is Anthropic’s coding assistant for the terminal and supported development tools. It can inspect code, edit files, run shell commands, and help with multi-file tasks.

Start with a supported Claude plan or a Claude Console account; a supported cloud provider is another option. Install Claude Code and launch it from your project folder. Describe one task clearly, review the proposed changes, and run the relevant tests. The official quickstart covers the setup and basic workflow.

Install Claude Code and start a session

You need an eligible account to use Claude Code. A free claude.ai plan does not include access. Supported options are Claude Pro, Max, Team, and Enterprise plans, a Claude Console account, or supported cloud providers such as Amazon Bedrock, Google Cloud’s Agent Platform, and Microsoft Foundry. Check the setup requirements before installing. They include supported operating systems, at least 4 GB of memory, an internet connection, and a compatible shell.

Install Claude Code using the instructions for your operating system. The official setup page covers macOS, Windows, and Linux, including package-manager options. After installation, open a terminal in the project directory and run:

claude

On first launch, Claude Code prompts you to sign in. If the ANTHROPIC_API_KEY environment variable is set, it asks you to approve that key instead. To switch accounts or sign in again during a session, enter /login.

Start by asking Claude to explain something specific in the project, such as:

Trace how this project handles user sign-in. Explain the main files and entry points without changing anything.

To give Claude an initial prompt from the command line, use claude "your request". The CLI reference also lists commands to continue or resume a conversation, or run a prompt without an interactive session. Inside Claude Code, enter /help to see available commands.

Give Claude a clear task

A clear, focused request helps Claude find the right code and makes its changes easier to review. State the goal, describe any constraints, and say what result you expect. For example:

Find why the password reset test fails. Explain the cause, make the smallest fix, and run the relevant tests.

For a larger change, ask Claude to outline a plan before it edits files. Check that plan for missing requirements or incorrect assumptions, then ask it to proceed. Break broad requests into smaller tasks, such as locating the code path first, making a targeted change next, and running tests afterward.

Claude Code reads project files as needed, so you do not have to paste the whole codebase into a prompt. For project-specific guidance, add a CLAUDE.md file at the repository root. Use it to record details such as the project structure, test commands, and coding conventions. Claude Code reads this file at the start of a session, so it can follow the same instructions across tasks.

Review changes before relying on them

Claude Code can edit files and run commands, so check what it did before using the result. Review the diff, confirm that the changes match the request, and run the relevant tests. If a test fails, share the failure output and ask Claude to investigate it.

Permission behavior depends on the session’s mode and configuration. Read each request to run a command or change a file. Approve only actions you understand and want Claude to perform. Avoid --dangerously-skip-permissions for local work: It disables permission prompts. Claude can then read or write files, execute commands, and commit without asking.

Avoid exposing sensitive files and credentials to Claude Code whenever possible. Before working in a repository that contains secrets, check which project files Claude Code can access.

Keep useful sessions and project guidance

Use /resume to continue a previous conversation. Use claude -c to continue the most recent conversation from the terminal, or claude -r to resume a session by its ID or name. Start a new session for a separate task to keep its instructions and discussion focused. To troubleshoot installation or settings problems, run claude doctor. It checks installation health and settings-file errors.

Claude Code also works in supported IDEs, the desktop app, and other interfaces. The quickstart describes the available ways to use it. Choose the terminal when you want direct access to a project directory and shell commands; use an IDE extension if you prefer to work alongside your editor.

Free and open source

Let your AI agent make the video

explainroo lets a coding agent like Claude Code or Codex make narrated explainer videos and product demos. The voice, the word timing and the rendering run on your own computer, with no API key and no cost per video. Optional outside services use your own key.

Copy this into your AI agent

Make me a short explainer video about [your topic]. Use explainroo for it: run npx --yes explainroo docs and follow what it prints.

Getting started Example videos GitHub

Made with explainroo: How explainroo makes a video

More on working with ai agents

All guides on working with ai agents