# claude-code-best/claude-code

原汁原昧 Claude Code 可运行,可构建, 可调试版; 生产级工程化, 企业级可靠性; 安全无毒, 内存泄露修复

Repository: https://github.com/claude-code-best/claude-code
Canonical: https://ross.abutalabs.com/products/claude-code-best-claude-code
Homepage: https://ccb.agent-aura.top/
Language: TypeScript
License Family: other
Last push: 2026-08-24T00:33:28+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 93, longevity 11
- inputs: {"age_days": 155, "days_push": 10, "days_rel": 44, "gap_med": 0, "n_releases_24m": 52}
- flags: young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 22224, forks 16480 (observed 2026-08-28T04:11:32.475560+00:00)

## What it is
An open-source, buildable and debuggable reimplementation of Anthropic's Claude Code terminal-native agentic coding system, written in TypeScript and running on Bun. It replicates the agentic loop, tool system, permissions, and adds features like Goal-driven orchestration, Artifacts hosting, and multi-agent workflows.

## Use cases
- run an AI coding agent in the terminal
- let AI read, edit and run code in my repo
- self-host a Claude Code compatible CLI
- understand how Claude Code works internally
- automate multi-step coding tasks with an agent
- use Claude Code features without an Anthropic login

## When to choose
- you want a terminal-native agentic coding assistant you can build and debug yourself
- you need Claude Code compatibility with extra enterprise features like Goal and Artifacts
- you want to study a production-grade agentic loop architecture

## When to avoid
- you need official Anthropic support or licensing (this repo has no license)
- you prefer IDE-integrated assistants like Cursor or Copilot
- you are uncomfortable with a reverse-engineered clone of proprietary software

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, cli, developer-tools, llm-inference, mcp
- domain: developer-tools, large-language-models
- platform: cli, bun, cross-platform, windows
- tags: claude-code, agentic-coding, terminal-ai, reverse-engineered, coding-assistant, ai-agents, command-line, nodejs, macos, linux

## Member repositories
- claude-code-best/claude-code (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:32.475560+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 2026-08-29T16:57:41.477210+00:00, confidence not recorded.
  - readme: https://github.com/claude-code-best/claude-code (fetched 2026-08-28T04:11:32.475560+00:00, sha a620eb8fa8e9)
  - homepage: https://ccb.agent-aura.top/ (fetched 2026-08-29T07:56:15.394039+00:00, sha 16ca58bb08e3)
  - site_page: https://ccb.agent-aura.top/docs/tools/what-are-tools (fetched 2026-08-29T07:56:15.413901+00:00, sha 80b943259c32)
  - site_page: https://ccb.agent-aura.top/docs/tools/file-operations (fetched 2026-08-29T07:56:15.415943+00:00, sha f00fc367fc51)
  - site_page: https://ccb.agent-aura.top/docs/tools/shell-execution (fetched 2026-08-29T07:56:15.417828+00:00, sha a0d137505b5d)
  - site_page: https://ccb.agent-aura.top/docs/introduction/architecture-overview (fetched 2026-08-29T07:56:15.406019+00:00, sha 6ee6aac668e9)
  - site_page: https://ccb.agent-aura.top/docs/conversation/the-loop (fetched 2026-08-29T07:56:15.408043+00:00, sha 4af60cfd1efc)
  - site_page: https://ccb.agent-aura.top/docs/conversation/streaming (fetched 2026-08-29T07:56:15.410147+00:00, sha 6e6ebf8777cb)
  - site_page: https://ccb.agent-aura.top/docs/conversation/multi-turn (fetched 2026-08-29T07:56:15.411904+00:00, sha 968a686141e0)
  - site_page: https://ccb.agent-aura.top/docs/introduction/why-this-whitepaper (fetched 2026-08-29T07:56:15.403800+00:00, sha f2e28264d3fa)
- Data as of 2026-08-30T08:39:29.467469+00:00.