A complete starter framework for building software with Claude Code — Anthropic's CLI for AI-assisted development.
Designed for vibe coders: people who build through AI, not by writing code manually.
# 1. Clone this repo
git clone https://github.com/Tokenized2027/Claude-Initilization-V7.git ~/Desktop/Claude
cd ~/Desktop/Claude
# 2. Run the setup script (installs Homebrew, Node.js, Git, Claude Code CLI)
chmod +x setup-mac.sh
./setup-mac.sh
# 3. Set CLAUDE_HOME in your shell profile
echo 'export CLAUDE_HOME="$HOME/Desktop/Claude"' >> ~/.zshrc
source ~/.zshrc
# 4. Copy settings + install hooks (gives Claude memory across sessions)
mkdir -p ~/.claude
cp dot-claude/settings.json ~/.claude/settings.json
chmod +x hooks/install-hooks.sh
./hooks/install-hooks.sh
# 5. Set up persistent memory
mkdir -p ~/.claude/projects/-home-$(whoami)/memory
cp memory/MEMORY.md ~/.claude/projects/-home-$(whoami)/memory/MEMORY.md
# Edit MEMORY.md with your info — this is how Claude remembers you
# 6. Start your first project
mkdir -p ~/projects/my-app && cd ~/projects/my-app && git init
cp $CLAUDE_HOME/examples/CLAUDE.md ./CLAUDE.md
claude├── setup-mac.sh ← One-command Mac setup
├── START_HERE.md ← Decision tree — what to do today
├── YOUR_WORKING_PROFILE.md ← Template: how Claude should work with you
│
├── hooks/ ← ⭐ NEW: Automatic context injection
│ ├── session-recall.py ← Memory across sessions (the killer hook)
│ ├── install-hooks.sh ← One-command hook installer
│ └── README.md ← How hooks work + how to build your own
│
├── memory/ ← ⭐ UPGRADED: Three-tier memory system
│ ├── MEMORY.md ← Template: what Claude remembers (200-line index)
│ ├── guidelines.md ← What to save, what to skip, how to organize
│ ├── patterns.md.template ← Mistake tracking with hit dates
│ ├── pending.md.template ← Transient state (blockers, in-progress)
│ ├── SESSION_MEMORY.md ← Guide: maintaining CLAUDE.md over time
│ └── SESSION_CONTINUITY.md ← Guide: resume mid-task across sessions
│
├── examples/ ← CLAUDE.md examples
│ ├── CLAUDE.md ← Blank template for any project
│ └── todo-app/CLAUDE.md ← Filled-in example
│
├── dot-claude/ ← Claude Code settings
│ └── settings.json ← Permissions + hook configuration
│
├── claude-code-framework/ ← Core framework
│ ├── QUICK_START.md ← 30-minute setup guide
│ ├── DAILY_REFERENCE.md ← Command cheatsheet
│ ├── essential/
│ │ ├── agents/ ← 13 AI agent templates
│ │ ├── guides/ ← 10 how-to guides
│ │ ├── skills/ ← 28 automation skills
│ │ └── toolkit/ ← Setup scripts + templates + safety hooks
│ └── advanced/
│ ├── guides/ ← Architecture, caching, MCP, team mode
│ └── infrastructure/ ← Mini PC / server setup
│
├── multi-agent-system/ ← Autonomous AI agent orchestration
│ ├── orchestrator/ ← FastAPI task router
│ ├── agent-templates/ ← 15 specialized agents
│ ├── mcp-servers/ ← Memory + context servers
│ └── deployment/ ← One-command installer
│
├── telegram-bot/ ← Mobile remote control for Claude
├── whisper-service/ ← Speech-to-text API (Docker)
├── project-contexts/ ← Project documentation templates
└── templates/ ← Docker + boilerplate templates
The hooks/ directory contains the single most impactful upgrade you can make to Claude Code. When you start a session, the session recall hook automatically injects summaries of your last 48 hours of work:
Recent sessions:
- Mar 14 04:37 | fix auth middleware for API routes | last action: PR is open
- Mar 14 02:15 | debug Docker port conflict | last action: Fixed, runs on port 3001
- Mar 13 22:30 | add price chart to dashboard | last action: Committed
Claude instantly knows what you were doing yesterday. No re-explaining. See hooks/README.md for setup and configuration.
Not all memory is equal. This kit organizes it by access frequency:
| Tier | What | Access |
|---|---|---|
| Hot | CLAUDE.md + MEMORY.md |
Every session, automatically |
| Warm | patterns.md + pending.md + topic files |
On demand, when relevant |
| Cold | knowledge-base/ + git history |
Retrieved when needed |
Hot memory is always loaded — keep it dense and under its line limit. When it grows, extract details to warm storage and leave a pointer. See memory/guidelines.md for the full system.
A file in your project root that Claude reads automatically at the start of every session. It tells Claude your project's rules, tech stack, structure, and conventions. This is the single highest-leverage file in Claude Code.
The key to a good CLAUDE.md is maintenance — update it after every meaningful session. See memory/SESSION_MEMORY.md for the maintenance guide and examples/ for templates.
Predefined role prompts (frontend developer, backend developer, security auditor, etc.) that give Claude specialized behavior for different tasks. See claude-code-framework/essential/agents/.
Reusable task automations — git recovery, Docker debugging, build error fixing, API scaffolding, and more. See claude-code-framework/essential/skills/.
| I want to... | Read this |
|---|---|
| Get started from zero | claude-code-framework/essential/guides/DAY_ZERO.md |
| Build my first project | claude-code-framework/essential/guides/FIRST_PROJECT.md |
| Learn daily workflow | claude-code-framework/essential/guides/DAILY_WORKFLOW.md |
| Set up hooks (session memory) | hooks/README.md |
| Understand the memory system | memory/guidelines.md |
| Maintain CLAUDE.md over time | memory/SESSION_MEMORY.md |
| Resume work across sessions | memory/SESSION_CONTINUITY.md |
| Recover from git disasters | claude-code-framework/essential/guides/GIT_FOR_VIBE_CODERS.md |
| Fix something that broke | claude-code-framework/essential/guides/TROUBLESHOOTING.md |
| Save money on API costs | claude-code-framework/advanced/guides/PROMPT_CACHING.md |
| Set up a mini PC / server | claude-code-framework/advanced/infrastructure/README.md |
| Deploy multi-agent system | multi-agent-system/docs/DEPLOYMENT_GUIDE.md |
- macOS (setup script is Mac-specific; framework files work on any OS)
- Python 3.9+ (for session recall hook — stdlib only, no pip install needed)
- Claude subscription (Pro or Max) or Anthropic API key
- Terminal basics (cd, ls, git — see
DAY_ZERO.mdif new)
MIT