Claude Code Unpacked

Ask DeepWikiYFeatured on Hacker News

What actually happens when you type a message into Claude Code? The agent loop, 50+ tools, multi-agent orchestration, and unreleased features, mapped straight from the source.

0+
Files
0K+
Lines of Code
0+
Tools
0+
Commands
Start exploring↓
01

The Agent Loop

From keypress to rendered response, step by step through the source.

Watch what happens when you send a message to Claude Code

1User Input

src/components/TextInput.tsx

User types a message or pipes input through stdin

● claude-code
$

Keyboard input comes from Ink's TextInput component. In non-interactive mode, it reads from piped stdin instead.

1/11
02

Architecture Explorer

Click around the source tree to explore what's inside.

Tools & Commands
Core Processing
UI Layer
Infrastructure
Support & Utilities
Personality & UX
utils/564 filescomponents/389 filescommands/189 filestools/184 filesservices/130 fileshooks/104 filesink/96 filesbridge/31 filesconstants/21 filesskills/20 filescli/
03

Tool System

Every built-in tool Claude Code can call, sorted by what it does.

File Operations
6 tools
Execution
3 tools
Search & Fetch
4 tools
Agents & Tasks
11 tools
Planning
5 tools
MCP
4 tools
System
11 tools
Experimental
8 tools

Click a tool to see details and source code
04

Command Catalog

Every slash command available in Claude Code, sorted by what it does.

Setup & Config12
Daily Workflow24
Code Review & Git13
Debugging & Diagnostics23
Advanced & Experimental23

Click a command to see details and source code
05

Hidden Features

Stuff that's in the code but not shipped yet. Feature-flagged, env-gated, or just commented out.

Click a feature to explore