SECTION · CLAUDE CODE GUIDE (BACKGROUND REFERENCE)

Claude Code Guide

A 4-group learning path for understanding Claude Code from scratch — the platform on which MoAI-ADK's three core values (Tokenomics, Agentic Loop Engineering, Agentic Harness) stand.

Foundations — Understanding Claude Code

From how Claude Code works to interactive usage, slash commands, tools, and the settings directory — the fundamentals that form the foundation of the agentic harness.

Context and Memory

The context window, memory, prompt caching, and checkpointing for carrying long tasks forward reliably and economically — the technical foundation of MoAI-ADK tokenomics.

Extensibility — Skills, Hooks, MCP, Plugins

The four extension points that widen Claude Code's capabilities (skills, hooks, MCP, plugins) — a concept-first tour of the materials for building an agentic harness.

Agents and Automation

Orchestration primitives like subagents, agent teams, and dynamic workflows, plus worktrees, goal-directed execution, scheduled tasks, large codebases, and best practices — the platform foundation of agentic loop engineering.

Info
Belongs to: Agentic Harness · Agentic Loop Engineering

Claude Code extensions

This section is a learning path for understanding Claude Code, Anthropic’s terminal CLI, from scratch. It is a guide for developers who are new to Claude Code, and for anyone who wants a precise grasp of the foundation MoAI-ADK operates on.

Background reference
This page is background material on Claude Code itself, the platform MoAI-ADK runs on. MoAI-ADK’s own features are covered in the sections above it in the sidebar.

Claude Code is a coding agent that runs in the terminal — it reads and modifies code, executes commands, and works through conversation with the developer. MoAI-ADK is an orchestration layer that runs on top of Claude Code, and its three core values — Tokenomics (Token Economics), Agentic Loop Engineering (recursive self-learning), and the Agentic Harness — are all built on the fundamental Claude Code mechanisms covered in this section. You cannot design tokenomics without knowing the context window and prompt caching, you cannot understand agentic loops without knowing subagents and /goal, and you cannot build a harness without knowing skills, hooks, and MCP.

Info
One-line summary: This section is where you learn Claude Code itself — the tool (platform). How MoAI-ADK builds its three core values on this foundation is covered in the Core Concepts and Advanced Learning sections.

Learning Flow

flowchart TD
    A[Foundations
How it works and how to use it] --> B[Context and Memory
Tokens, memory, caching]
    B --> C[Extensibility
Skills, hooks, MCP, plugins]
    C --> D[Agents and Automation
Subagents, teams, workflows]

Start with the Foundations group to learn how Claude Code’s agentic loop turns, then build your token-economy fundamentals with context and memory management. Next, gather the harness ingredients in Extensibility, and finally advance to autonomous execution loops in Agents and Automation.

Contents — Four Groups Mapped to the Three Core Values

Document	Description	Connected MoAI core value
Foundations	How Claude Code works and basic usage	Shared foundation of all three core values
Context and Memory	Managing tokens, context, memory, caching, checkpoints	Tokenomics
Extensibility	Extending capabilities with skills, hooks, MCP, plugins	Agentic Harness
Agents and Automation	Subagents, teams, workflows, autonomous execution	Agentic Loop Engineering

Completing the four groups in order gives you an understanding of the entire Claude Code platform. From there, move on to the MoAI-ADK Core Concepts section to see how SPEC-based development and token-efficient design combine on top of this foundation.