Source archive of Claude Code and different versions of the clean-room Python rewrite research repository
- 12:20 PM · Apr 02, 2026: Nano Claude Code v3.0: Multi-agent packages, memory package, skill package with built-in skills, argument substitution, fork/inline execution, AI memory search, git worktree isolation, agent type definitions (~5000 Lines)
- 7:40 AM · Apr 02, 2026: Nano Claude Code v2.0: A Minimal Python Reimplementation (~3400 Lines), support open and closed source models, skill and memory
- 8:36 AM · Apr 01, 2026: Nano Claude Code v1.0: A Minimal Python Reimplementation (~1300 Lines)
- 0:20 AM · Apr 01, 2026: Analysis of Claude Code source code (Video: In Chinese)
- 19:45 PM · Mar 31, 2026: Claude Code Research Report (In Chinese)
- 17:13 PM · Mar 31, 2026: Architecture Analysis of Claude Code (In Chinese)
- 16:43 PM · Mar 31, 2026: An Overview of Claude Code Features (In Chinese)
- 15:00 PM · Mar 31, 2026: Hacker News Community Discussion about Claude Code Leak
- 11:48 AM · Mar 31, 2026: How Anthropic Built 7 Layers of Memory and a Dreaming System for Claude Code
- 9:50 AM · Mar 31, 2026: Claude code memory analysis
- 9:48 AM · Mar 31, 2026: Claude Code's source code appears to have leaked: here's what we know
- 9:04 AM · Mar 31, 2026: Deconstructing the Claude Code Architecture (In Chinese)
- 7:07 AM · Mar 31, 2026: A Walkthrough of Claude Code's Source Code (In Chinese)
- 5:41 AM · Mar 31, 2026: Why Claude Code Outperforms Other Tools: An Analysis
- 5:03 AM · Mar 31, 2026: Anthropic's AI Coding Tool Leaks Its Own Source Code
- 3:28 AM · Mar 31, 2026: Claude Code Source Code Unveiled: Prompts, Self-Healing Mechanisms, and Multi-Agent Architecture (In Chinese)
- 3:02 AM · Mar 31, 2026: Community Reaction to the Claude Code Source Exposure
- 2:57 AM · Mar 31, 2026: The code behind Claude Code
- 1:23 AM · Mar 31, 2026: A Viral Post on the Claude Code Source Leak
- 1. original-source-code
- 2. claude-code-source-code
- 3. claw-code
- 4. nano-claude-code
- Comparison of the Projects
- License and Disclaimer
This repository contains subprojects that study Claude Code (Anthropic’s official CLI tool) from multiple angles:
| Subproject | Language | Nature | File Count |
|---|---|---|---|
| original-source-code | TypeScript | Raw leaked source archive | 1,884 files |
| claude-code-source-code | TypeScript | Decompiled source archive (v2.1.88) + docs | 1,940 files |
| claw-code | Python | Clean-room architectural rewrite | 109 files |
| nano-claude-code | Python | Minimal multi-provider reimplementation | ~30 files |
The raw leaked TypeScript source of Claude Code, preserved as-is from the original exposure on March 31, 2026. Contains 1,884 TypeScript/TSX files (packaged as src.zip) spanning the full src/ directory tree — the same files that triggered community discussion and downstream research.
original-source-code/
├── src/ # Full TypeScript source tree (1,884 .ts/.tsx files)
│ ├── main.tsx # CLI entry point
│ ├── query.ts # Core agent loop
│ ├── commands.ts # Slash command definitions
│ ├── tools.ts # Tool registration
│ └── ... # All other source directories (same layout as claude-code-source-code/src)
├── src.zip # Compressed archive (~9.5 MB)
└── readme.md
This directory serves as the unmodified reference snapshot. No annotations, docs, or build tooling have been added — use claude-code-source-code for the researched and annotated version.
A decompiled/unpacked source archive of Claude Code v2.1.88, reconstructed from the npm package @anthropic-ai/claude-code@2.1.88, containing approximately 163,318 lines of TypeScript code.
claude-code-source-code/
├── src/
│ ├── main.tsx # CLI entry and REPL bootstrap (4,683 lines)
│ ├── query.ts # Core main agent loop (largest single file, 785KB)
│ ├── QueryEngine.ts # SDK/Headless query lifecycle engine
│ ├── Tool.ts # Tool interface definitions + buildTool factory
│ ├── commands.ts # Slash command definitions (~25K lines)
│ ├── tools.ts # Tool registration and presets
│ ├── context.ts # User input context handling
│ ├── history.ts # Session history management
│ ├── cost-tracker.ts # API cost tracking
│ ├── setup.ts # First-run initialization
│ │
│ ├── cli/ # CLI infrastructure (stdio, structured transports)
│ ├── commands/ # ~87 slash command implementations
│ ├── components/ # React/Ink terminal UI (33 subdirectories)
│ ├── tools/ # 40+ tool implementations (44 subdirectories)
│ ├── services/ # Business logic layer (22 subdirectories)
│ ├── utils/ # Utility function library
│ ├── state/ # Application state management
│ ├── types/ # TypeScript type definitions
│ ├── hooks/ # React Hooks
│ ├── bridge/ # Claude Desktop remote bridge
│ ├── remote/ # Remote mode
│ ├── coordinator/ # Multi-agent coordination
│ ├── tasks/ # Task management
│ ├── assistant/ # KAIROS assistant mode
│ ├── memdir/ # Long-term memory management
│ ├── plugins/ # Plugin system
│ ├── voice/ # Voice mode
│ └── vim/ # Vim mode
│
├── docs/ # In-depth analysis docs (bilingual: Chinese/English)
│ ├── en/ # English analysis
│ └── zh/ # Chinese analysis
├── vendor/ # Third-party dependencies
├── stubs/ # Module stubs
├── types/ # Global type definitions
├── utils/ # Top-level utility functions
├── scripts/ # Build scripts
└── package.json
User Input
↓
processUserInput() # Parse /slash commands
↓
query() # Main agent loop (query.ts)
├── fetchSystemPromptParts() # Assemble system prompt
├── StreamingToolExecutor # Parallel tool execution
├── autoCompact() # Automatic context compression
└── runTools() # Tool orchestration and scheduling
↓
yield SDKMessage # Stream results back to the consumer
| Component | Technology |
|---|---|
| Language | TypeScript 6.0+ |
| Runtime | Bun (compiled into Node.js >= 18 bundle) |
| Claude API | Anthropic SDK |
| Terminal UI | React + Ink |
| Code Bundling | esbuild |
| Data Validation | Zod |
| Tool Protocol | MCP (Model Context Protocol) |
| Category | Tools |
|---|---|
| File Operations | FileReadTool, FileEditTool, FileWriteTool |
| Code Search | GlobTool, GrepTool |
| System Execution | BashTool |
| Web Access | WebFetchTool, WebSearchTool |
| Task Management | TaskCreateTool, TaskUpdateTool, TaskGetTool, TaskListTool |
| Sub-agents | AgentTool |
| Code Environments | NotebookEditTool, REPLTool, LSPTool |
| Git Workflow | EnterWorktreeTool, ExitWorktreeTool |
| Configuration & Permissions | ConfigTool, AskUserQuestionTool |
| Memory & Planning | TodoWriteTool, EnterPlanModeTool, ExitPlanModeTool |
| Automation | ScheduleCronTool, RemoteTriggerTool, SleepTool |
| MCP Integration | MCPTool |
/commit /commit-push-pr /review /resume /session /memory /config /skills /help /voice /desktop /mcp /permissions /theme /vim /copy and more
- Three modes:
default(ask user) /bypass(auto-allow) /strict(auto-deny) - Tool-level fine-grained control
- ML-based automated permission inference classifier
- Persistent storage of permission rules
- Automatic compression strategies (
autoCompact): reactive compression, micro-compression, trimmed compression - Context collapsing (
CONTEXT_COLLAPSE) - Token counting and estimation
- Session transcript persistence
| Document | Content |
|---|---|
| 01 - Telemetry and Privacy | Dual-layer analysis pipeline (Anthropic + Datadog), with no opt-out switch |
| 02 - Hidden Features and Model Codenames | Internal codenames such as Capybara, Tengu, Fennec, Numbat |
| 03 - Undercover Mode | Anthropic employees automatically entering undercover mode in public repositories |
| 04 - Remote Control and Emergency Switches | Hourly polling, 6+ killswitches, dangerous-change popups |
| 05 - Future Roadmap | KAIROS autonomous agent, voice mode, 17 unreleased tools |
A clean-room Python rewrite of Claude Code (without including original source copies), focused on architectural mirroring and research. Built by @instructkr (Sigrid Jin), and became one of the fastest GitHub repositories in the world to reach 30K stars.
claw-code/
├── src/
│ ├── __init__.py # Package export interface
│ ├── main.py # CLI entry (~200 lines)
│ ├── query_engine.py # Core query engine
│ ├── runtime.py # Runtime session management
│ ├── models.py # Shared data classes
│ ├── commands.py # Command metadata and execution framework
│ ├── tools.py # Tool metadata and execution framework
│ ├── permissions.py # Permission context management
│ ├── context.py # Ported context layer
│ ├── setup.py # Workspace initialization
│ ├── session_store.py # Session persistence
│ ├── transcript.py # Session transcript storage
│ ├── port_manifest.py # Workspace manifest generation
│ ├── execution_registry.py # Execution registry
│ ├── history.py # History logs
│ ├── parity_audit.py # Parity audit against TypeScript source
│ ├── remote_runtime.py # Remote mode simulation
│ ├── bootstrap_graph.py # Bootstrap graph generation
│ ├── command_graph.py # Command graph partitioning
│ ├── tool_pool.py # Tool pool assembly
│ │
│ ├── reference_data/ # JSON snapshot data (drives command/tool metadata)
│ │ ├── commands_snapshot.json
│ │ └── tools_snapshot.json
│ │
│ ├── commands/ # Command implementation subdirectory
│ ├── tools/ # Tool implementation subdirectory
│ ├── services/ # Business logic services
│ ├── components/ # Terminal UI components (Python version)
│ ├── state/ # State management
│ ├── types/ # Type definitions
│ ├── utils/ # Utility functions
│ ├── remote/ # Remote mode
│ ├── bridge/ # Bridge modules
│ ├── hooks/ # Hook system
│ ├── memdir/ # Memory management
│ ├── vim/ # Vim mode
│ ├── voice/ # Voice mode
│ └── plugins/ # Plugin system
│
└── tests/ # Validation tests
| Class / Module | Responsibility |
|---|---|
QueryEnginePort |
Query engine handling message submission, streaming output, and session compression |
PortRuntime |
Runtime manager responsible for routing, session startup, and turn-loop execution |
PortManifest |
Workspace manifest that generates Markdown overviews |
ToolPermissionContext |
Tool permission context (allow / deny / ask) |
WorkspaceSetup |
Environment detection and initialization reporting |
TranscriptStore |
Session transcript storage with append, compaction, and replay support |
python3 -m src.main [COMMAND]
# Overview
summary # Markdown workspace overview
manifest # Print manifest
subsystems # List Python modules
# Routing and indexing
commands # List all commands
tools # List all tools
route [PROMPT] # Route prompt to corresponding command/tool
# Execution
bootstrap [PROMPT] # Start runtime session
turn-loop [PROMPT] # Run turn loop (--max-turns)
exec-command NAME # Execute command
exec-tool NAME # Execute tool
# Session management
flush-transcript # Persist session transcript
load-session ID # Load saved session
# Remote mode
remote-mode TARGET # Simulate remote control
ssh-mode TARGET # Simulate SSH branch
teleport-mode TARGET # Simulate Teleport branch
# Audit and config
parity-audit # Compare consistency with TypeScript source
setup-report # Startup configuration report
bootstrap-graph # Bootstrap phase graph
command-graph # Command graph partition view
tool-pool # Tool pool assembly view- Snapshot-driven: command/tool metadata is loaded through JSON snapshots without requiring full logical implementations
- Clean-room rewrite: does not include original TypeScript code; independently implemented
- Parity audit: built-in
parity_audit.pytracks gaps from the original implementation - Lightweight architecture: core framework implemented in 109 files, suitable for learning and extension
A minimal, fully-runnable Python reimplementation of Claude Code (~5,000 lines). Unlike claw-code (which focuses on architectural mapping), nano-claude-code is a real coding assistant that can be used immediately. It supports 20+ closed-source models and local open-source models, and has grown from a ~900-line prototype to a feature-rich v3.0 with multi-agent orchestration, persistent memory, and a skill system.
| Feature | Details |
|---|---|
| Multi-provider | Anthropic · OpenAI · Gemini · Kimi · Qwen · Zhipu · DeepSeek · Ollama · LM Studio · Custom endpoint |
| Interactive REPL | readline history, Tab-complete slash commands |
| Agent loop | Streaming API + automatic tool-use loop |
| 18 built-in tools | Read · Write · Edit · Bash · Glob · Grep · WebFetch · WebSearch · MemorySave · MemoryDelete · MemorySearch · MemoryList · Agent · SendMessage · CheckAgentResult · ListAgentTasks · ListAgentTypes · Skill · SkillList |
| Diff view | Git-style red/green diff display for Edit and Write |
| Context compression | Auto-compact long conversations to stay within model limits |
| Persistent memory | Dual-scope memory (user + project) with 4 types, AI search, staleness warnings |
| Multi-agent | Spawn typed sub-agents (coder/reviewer/researcher/…), git worktree isolation, background mode |
| Skills | Built-in /commit · /review + custom markdown skills with argument substitution and fork/inline execution |
| Plugin tools | Register custom tools via tool_registry.py |
| Permission system | auto / accept-all / manual modes |
| 17 slash commands | /model · /config · /save · /cost · /memory · /skills · /agents · … |
| Context injection | Auto-loads CLAUDE.md, git status, cwd, persistent memory |
| Session persistence | Save / load conversations to ~/.nano_claude/sessions/ |
| Extended Thinking | Toggle on/off (Claude models only) |
| Cost tracking | Token usage + estimated USD cost |
| Non-interactive mode | --print flag for scripting / CI |
Closed-source (API): Claude (Anthropic), GPT / o-series (OpenAI), Gemini (Google), Kimi (Moonshot AI), Qwen (Alibaba DashScope), GLM (Zhipu), DeepSeek
Open-source (local via Ollama): llama3.3/3.2, qwen2.5-coder, deepseek-r1, phi4, mistral, mixtral, gemma3, codellama, and any model on ollama list
Self-hosted: vLLM, LM Studio, or any OpenAI-compatible endpoint via CUSTOM_BASE_URL
nano-claude-code/
├── nano_claude.py # Entry point: REPL + slash commands + rendering (~748 lines)
├── agent.py # Agent loop: message format + tool dispatch (~174 lines)
├── providers.py # Multi-provider adapters + message conversion (~507 lines)
├── tools.py # Tool dispatch + auto-registration of all packages (~467 lines)
├── tool_registry.py # Central tool registry + plugin entry point (~98 lines)
├── context.py # System prompt builder: CLAUDE.md + git + memory (~135 lines)
├── compaction.py # Context compression (auto-compact) (~196 lines)
├── config.py # Config load/save/defaults (~72 lines)
├── memory.py # Backward-compat shim → memory/
├── skills.py # Backward-compat shim → skill/
├── subagent.py # Backward-compat shim → multi_agent/
│
├── memory/ # Persistent memory package
│ ├── store.py # Save/load/delete/search memory entries
│ ├── scan.py # Index scanning, age/freshness helpers
│ ├── context.py # System-prompt injection + AI-ranked search
│ ├── types.py # MEMORY_TYPES definitions
│ └── tools.py # MemorySave · MemoryDelete · MemorySearch · MemoryList
│
├── skill/ # Skill system package
│ ├── loader.py # SkillDef, file parsing, argument substitution
│ ├── builtin.py # Built-in skills: /commit, /review
│ ├── executor.py # Inline + fork execution modes
│ └── tools.py # Skill · SkillList
│
├── multi_agent/ # Multi-agent orchestration package
│ ├── subagent.py # AgentDefinition, SubAgentTask, SubAgentManager, worktree helpers
│ └── tools.py # Agent · SendMessage · CheckAgentResult · ListAgentTasks · ListAgentTypes
│
├── tests/ # 101 tests (monkeypatched, no real ~/.nano_claude touched)
├── docs/ # Docs and demo assets
└── requirements.txt
Quick start:
pip install anthropic openai httpx rich
export ANTHROPIC_API_KEY=sk-ant-...
python nano_claude.py
# Switch provider at startup
python nano_claude.py --model gpt-4o
python nano_claude.py --model ollama/qwen2.5-coder
# Non-interactive / CI
python nano_claude.py --print "Write a Python fibonacci function" --accept-allMemory — persistent across sessions, dual-scope (user ~/.nano_claude/memory/ and project .nano_claude/memory/):
/memory # list all memories with staleness info
MemorySave(name="...", type="feedback", content="...", scope="user")
MemorySearch(query="...", use_ai=True)
Skills — reusable prompt templates, invoke from REPL:
/commit # built-in: review staged changes and create a git commit
/review 123 # built-in: review PR #123
/skills # list all available skills with triggers and hints
Multi-agent — spawn typed sub-agents with optional git worktree isolation:
Agent(prompt="...", subagent_type="coder", isolation="worktree", wait=False)
SendMessage(agent_name="my-agent", message="...")
/agents # show all active and finished sub-agent tasks
| Dimension | original-source-code | claude-code-source-code | claw-code | nano-claude-code |
|---|---|---|---|---|
| Language | TypeScript | TypeScript | Python | Python |
| Code Size | ~163,000 lines | ~163,000 lines + docs | ~5,000 lines | ~5,000 lines |
| Nature | Raw leaked source archive | Decompiled source archive + analysis | Clean-room architectural rewrite | Minimal functional reimplementation |
| Functional Completeness | Complete (100%) | Complete (100%) | Architectural framework (~20%) | Core loop + memory + multi-agent + skills |
| Core Loop | query.ts (785KB) |
query.ts (785KB) |
QueryEnginePort (~200 lines) |
agent.py (~174 lines) |
| Tool System | 40+ fully implemented tools | 40+ fully implemented tools | Snapshot metadata + execution framework | 18 fully implemented tools + plugin registry |
| Memory System | Yes (7-layer, complex) | Yes (7-layer, complex) | No | Yes (dual-scope, 4 types, AI search) |
| Multi-agent | Yes (full coordinator) | Yes (full coordinator) | No | Yes (typed agents, worktree isolation) |
| Skills | Yes (~87 commands) | Yes (~87 commands) | Snapshot metadata only | Yes (built-in + custom markdown skills) |
| Multi-provider | No (Anthropic only) | No (Anthropic only) | No | Yes (10+ providers) |
| Immediately Runnable | No | No | Limited (CLI metadata only) | Yes |
| Main Use Case | Raw reference snapshot | Deep study of full implementation details | Architectural understanding and porting research | Lightweight full-featured coding assistant |
This repository is for academic research and educational purposes only. All subprojects are built from publicly accessible information. Users are responsible for complying with applicable laws, regulations, and service terms.