# Windy3f3f3f3f/how-claude-code-works

Deep dive into Claude Code internals — architecture, agent loop, context engineering, and more. / 深入解析 Claude Code 源码：架构、Agent 循环、上下文工程、工具系统等

Repository: https://github.com/Windy3f3f3f3f/how-claude-code-works
Canonical: https://ross.abutalabs.com/products/how-claude-code-works
Homepage: https://windy3f3f3f3f.github.io/how-claude-code-works/#/
License: MIT
License Family: permissive
Topics: ai, ai-agent, anthropic, architecture, claude, claude-code, coding-agent, deep-dive, llm, source-code-analysis
Last push: 2026-08-17T08:27:42+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 35, longevity 11
- inputs: {"age_days": 155, "days_push": 16, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3529, forks 695 (observed 2026-08-28T04:08:08.802007+00:00)

## What it is
A set of educational architecture notes and deep-dive documentation analyzing the leaked source code of Claude Code, Anthropic's AI coding agent. It covers the agent loop, context engineering, tool system, and permission system across 21 chapters, with an online docs site.

## Use cases
- understand how claude code works internally
- learn agent loop architecture for coding agents
- study context engineering and context compaction techniques
- build my own ai coding agent
- learn how tool execution and permission systems work in ai agents
- reverse engineer a production llm agent system

## When to choose
- you want to study a production-grade agentic coding system's architecture
- you're building your own coding agent and want design references
- you want to use Claude Code more effectively by understanding its internals

## When to avoid
- you need official Anthropic documentation or guarantees of implementation accuracy
- you want runnable software rather than reading material
- you need a stable reference — the analyzed source evolves quickly

## Facets
- artifact type: learning-resource
- maturity: active
- function: documentation, developer-tools
- domain: large-language-models, developer-tools, tutorials
- platform: -
- tags: claude-code, source-code-analysis, architecture-notes, agentic-coding, context-engineering, reverse-engineering, educational, ai-agents, web-server

## Member repositories
- Windy3f3f3f3f/how-claude-code-works (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:08.802007+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:35:07.460364+00:00, confidence not recorded.
  - readme: https://github.com/Windy3f3f3f3f/how-claude-code-works (fetched 2026-08-28T04:08:08.802007+00:00, sha 968011c7c5c0)
  - homepage: https://windy3f3f3f3f.github.io/how-claude-code-works/#/ (fetched 2026-08-29T09:29:23.420908+00:00, sha aaafc1d7d40e)
- Data as of 2026-08-30T08:39:29.467469+00:00.