Architecture Walkthrough
Read the guided chapters from startup flow to tool dispatch, permissions, memory, MCP, streaming, and commands.
Learning Lab
Claude Code Internals
A structured, bilingual tutorial for studying Claude Code as a real CLI agent: architecture, loop design, tools, permissions, memory, MCP, streaming, and hands-on Python labs.
Quick Start
Build a minimal agent, add a tool, and wire a streaming chat loop through short tutorials.
Hands-on Labs
Pair each chapter with a runnable Python experiment that recreates one production-shaped agent pattern.
References
Keep the source-map notes, pattern cheatsheets, diagrams, and glossary close while reading.