Start here: your first terminal task
Goal. Install Claude Code, point it at a real project folder, and hand the agent a whole task — turning a data file into a working, runnable script.
claude "Read measurement_data.csv and write a Python script that plots a heatmap of sample vs condition with seaborn. Save it as heatmap.png and run it."
Install Claude Code, then run claude inside your project folder and type your task in plain English — in any language you like. Claude reads your project files itself; you don't paste them in.
- 1Install. macOS/Linux:
curl -fsSL https://claude.ai/install.sh | bash. Windows PowerShell:irm https://claude.ai/install.ps1 | iex. (Orbrew install --cask claude-code/winget install Anthropic.ClaudeCode.) Native installs update themselves in the background. - 2Open a terminal in your project folder (
cd ~/lab-data) and runclaude. You log in on first use with a Claude Pro or Max account. There is no free tier. - 3Type the task above. Claude finds the right file, shows you the changes it proposes, asks for approval, then writes and runs the script — and reads the traceback and fixes the bug if it errors.
You'll see. A heatmap.png produced by a Python script that Claude wrote, ran, and fixed if it errored — all from one plain-English task in your terminal.
Cost. Claude Code needs a paid Claude plan — Pro ($17/mo annual, $20/mo monthly) is the entry point and includes Claude Code. There is no free tier.
Takeaway. Claude Code takes a whole task, not a single line — describe the outcome and it reads your project, writes real code, runs it, and iterates on failures.