oboard/claude-code-rev

Runnable ClaudeCode source code

📦 What is claude-code-rev?

A community‑run reconstruction of the Claude Code source tree. The original repository is closed‑source, but enough of its JavaScript/TypeScript code survived in source‑maps that volunteers could piece together a working copy. The project ships a CLI that can run Claude‑style AI agents locally, using inference back‑ends that expose an OpenAI‑compatible API (GitHub Models, GitHub Copilot). It is essentially a runnable, open‑source façade of Anthropic’s Claude Code runtime.

✅ Core capabilities

Feature What it does
Restored CLI bun run dev starts the original Claude Code command‑line interface, showing the full command tree.
Multi‑provider inference Supports two providers that forward chat requests to GitHub‑hosted endpoints:
• github-models OpenAI‑compatible GitHub Models API.
• github-copilot GitHub Copilot account‑backed endpoint, currently able to run Claude‑family models.
Model picker Inside the CLI you can run /provider and /model to switch providers and choose a specific model (e.g., claude‑opus‑4.6).
Skill bundles Restored claude‑api and verify skill definitions are now usable reference docs rather than empty placeholders.
Compatibility shims Chrome MCP and Computer‑Use MCP layers expose a realistic tool catalog and return structured fallback responses, allowing the agent’s tool‑use loop to function.

🛠️ Getting started

# 1️⃣ Install the required runtime
#    • Bun 1.3.5+ (https://bun.sh)
#    • Node 24+ (for type definitions)

bun install               # installs dependencies from package.json
bun run dev               # launches the restored Claude Code CLI
bun run version           # prints the reconstructed version string

Selecting a provider & model

# Use GitHub Models (OpenAI‑compatible)
bun run dev --settings '{"provider":"github-models"}'
# Or pick a concrete model
bun run dev --settings '{"provider":"github-models"}' --model "openai/gpt-4.1"

# Use GitHub Copilot (Claude models)
bun run dev --settings '{"provider":"github-copilot"}'
# Pick a Claude model
bun run dev --settings '{"provider":"github-copilot"}' --model "claude-opus-4.6"

Inside the CLI you can also type:

  • /provider – open an interactive provider selector.
  • /provider info – show which provider is active and any env‑var overrides.
  • /model – list models supported by the current provider.

Authentication is looked up in this order:

  1. Provider‑specific environment variable (e.g., GITHUB_MODELS_TOKEN).
  2. GH_TOKEN
  3. GITHUB_TOKEN
  4. Token stored by the GitHub CLI (gh auth token).

If you prefer the GitHub CLI login flow, run gh auth login first.

⚠️ Known limitations

  • Partial restoration – Some private or native modules could not be recovered from source maps; they are replaced by shims or degraded implementations. Behaviour may differ from the original Claude Code.
  • Tooling gaps – The Copilot‑hosted GPT/Grok style models are not yet wired; only Claude‑family models work through the current chat/message shim.
  • Provider scope – Only GitHub‑based inference endpoints are supported. No direct Anthropic or self‑hosted back‑ends are included.
  • Runtime requirement – The project depends on Bun (a relatively new JavaScript runtime). It will not run under plain Node without modification.

📚 Where to look next

  • README.md – the source of this summary; contains the full command‑line usage.
  • preview.png – a screenshot of the restored CLI.
  • package.json – lists the exact dependencies and Bun scripts.
  • src/ – the reconstructed source tree; files marked as shims are clearly commented.

In short, claude-code-rev gives developers a way to experiment with Claude‑style agents locally, even though the original code is proprietary. It works best for exploring the CLI workflow and testing Claude models via GitHub’s inference services.

Related

  • Project
    wesammustafa/Claude-Code-Everything-You-Need-to-Know

    A thorough, example‑rich tutorial for Anthropic’s Claude Code CLI, covering installation, prompt engineering, and advanced extensions (Skills, Hooks, Sub‑agents, Workflows, MCP) to let developers use Claude for repo‑wide code analysis, editing, testing and automation.

  • Dispatch
    Claude for Apple Foundation Models Integration

    Anthropic has released a Swift package allowing Claude to be used as a server-side language model within Apple's Foundation Models framework, providing a unified API for switching between on-device and cloud-hosted models.

  • Project
    alirezarezvani/claude-skills

    A comprehensive library of 388 modular agent skills and plugins for AI coding tools, providing specialized domain expertise across engineering, business, and research.

  • Project
    Prorise-cool/Claude-Code-Multi-Agent

    Claude Code Multi‑Agent is an open‑source framework that gives Anthropic’s Claude Code project awareness and expert‑agent capabilities. It uses a Python Hook system and a locally‑run Ollama model to automatically detect the codebase type, analyse user intent, load 300+ domain‑specific “Skills” (backend, frontend, testing, security, etc.), and enforce automatic documentation (DEVELOPMENT.md, KNOWLEDGE.md, CHANGELOG.md). After cloning, installing Ollama, a model, and uv, you drop your project into the repo and Claude Code instantly knows its context, suggests appropriate tools, and can be driven with slash commands like `/backend‑specialist` or workflow commands such as `/kiro/spec`. Key points: - Zero‑config start; all configuration lives in `.claude/settings.json` and `prompts.json`. - Hooks run on session start, user prompt, tool use, etc., handling project detection, intent analysis, and doc updates. - Skills are defined by `SKILL.md` files and are auto‑discovered. - Works entirely locally (no external API), integrates with Git, and is MIT‑licensed. Ideal for developers or small teams who want a reproducible, on‑premise AI coding assistant.

  • Dispatch
    Mastering Claude Code: From Prompting to Programmable Engineering

    A deep dive into transforming Claude Code from a simple chatbot into a sophisticated autonomous agent using custom skills, subagents, and the Model Context Protocol.