Claude Code for Non-Programmers

In this one-day course, students will be introduced to using Claude Code, which is the agentic command-line interface (CLI) tool that is used to interact directly with local filesystems, build features, and automate routine technical tasks in the Claude Code (https://support.anthropic.com/en/articles/10636361-what-is-claude-code) environment. The course covers the fundamental terminal mechanics required to launch the tool, but the emphasis is on using Claude Code to analyze and organize local projects without extensive programming experience. Students will learn how to configure project guardrails, navigate multi-mode execution cycles, control token budgets, and safely coordinate parallel background sub-agents. This instructor-led course consists of a combination of lecture, demonstrations, and hands-on labs. Software Used During the Class The standard lab setup for this class consists of a PC or Mac with a terminal application (such as Terminal, PowerShell, or Git Bash), a modern text editor or IDE (such as VS Code), and Claude Code installed natively or via a package manager.

Getting started with Claude Code Exploring the Claude Code terminal interface and architecture Moving from conversational interfaces to an active filesystem agent Navigating directories and standard command-line survival basics Authenticating your workstation using the login protocol Launching your first interactive session and running simple text queries Managing context, budgets, and the workspace Understanding token windows, data transmission, and pricing tiers Monitoring billing spend and token boundaries using the usage command Reducing context overhead and visual text clutter with compact mode Clearing terminal history noise to reset focus during long sessions Tracking real-time background tool executions with the statusline indicator Reviewing project-wide insights and developer habit recommendations Controlling execution modes and interaction styles Alternating between core execution states using terminal shortcut keys Reviewing individual file modifications securely within Normal Mode Scoping structural requirements before editing code using Plan Mode Accelerating batch processing and bulk edits with Auto-Accept Mode Evaluating the safety and security trade-offs of skipping explicit permissions Applying Learning style and Explanatory style to capture engineering logic Project configuration and memory boundaries Initializing permanent workspace rules and tracking tools Enforcing team preferences and formatting rules via a local memory file Safeguarding codebase integrity with checkpoint tracking and system rollbacks Reverting broken multi-file edits cleanly without losing local manual updates Purging local operational histories and internal data caches safely Structuring global automation parameters versus project-specific instructions Scripting automation and background sub-agents Executing single-line updates directly from scripts using non-interactive pipe commands Introduction to multi-agent tasking and spawning background sub-workers Monitoring active worker progress using diagnostic session attachments and logs Controlling parallel execution costs to prevent runaway token exhaustion Coordinating independent, decoupled AI interactions with Delegate Mode Best practices for preventing conflict and chaos in autonomous multi-agent systems