vibe-me 2.0.0 → 3.1.0

package/README.md

@@ -1,161 +1,551 @@
-# vibes
+<div align="center">
 
-**Every repository should contain a continuously maintained, AI-readable and human-readable model of itself.**
+# 🌀 vibes
 
-`vibes` creates a `.vibe/` folder in your project — 6 files that let any human or AI understand your codebase without reading the source code.
+### Every project should contain a continuously maintained,<br>AI-readable and human-readable model of itself.
+
+[![npm version](https://img.shields.io/npm/v/vibe-me?style=flat-square&color=7c3aed&label=npm)](https://www.npmjs.com/package/vibe-me)
+[![license](https://img.shields.io/badge/license-MIT-10b981?style=flat-square)](LICENSE)
+[![node](https://img.shields.io/badge/node-%3E%3D16-3b82f6?style=flat-square&logo=node.js&logoColor=white)](https://nodejs.org)
+[![zero deps](https://img.shields.io/badge/dependencies-0-f59e0b?style=flat-square)](package.json)
+
+<br>
+
+**One command. Full project memory. Total clarity.**
+
+`vibes` creates a `.vibe/` semantic layer and a standardized `docs/` folder —<br>
+so any human or AI can understand your project, your product, and your business<br>
+without reading the source code.
+
+<br>
+
+```bash
+npx vibe-me all
+```
+
+<br>
+</div>
+
+---
+
+## 🌊 We're Building in the AI Age — But Our Projects Are Still Dumb
+
+AI coding agents are everywhere. Cursor, Copilot, Windsurf, Claude Code, Gemini — they can all write code. But every one of them hits the same wall:
+
+**They don't understand your project.**
+
+They can read your files. They can parse your syntax. But they can't answer the questions that actually matter:
+
+- *Why does this exist?*
+- *What were the alternatives that were rejected?*
+- *What's the user actually trying to do — and how do they feel?*
+- *What's broken right now? What's the current focus?*
+- *What should I absolutely NOT touch?*
+- *Is this a business? What's the pricing rationale? Who are the competitors?*
+
+So what happens? Your AI agent guesses. It writes plausible-looking code that violates assumptions nobody documented. You spend an hour debugging something that should've taken five minutes. Sound familiar?
+
+### The Research Question
+
+> **What is the minimum set of files that lets a human or AI understand an entire project — not just the code, but the product, the business, and the current momentum — without reading the source?**
+
+We tested this across production repositories — web apps, CLI tools, embedded systems, game engines, blockchain protocols — and found the answer:
+
+**15 files for project memory. 11 files for operational docs.**
+
+Not a wiki. Not a Confluence graveyard. Not a 200-page design doc nobody reads. Just well-structured files that capture everything that matters — from repository architecture to business strategy to user psychology.
+
+### Before and After
+
+<table>
+<tr>
+<td width="50%">
+
+**❌ Without vibes**
+- AI agent reads 50 files to guess the architecture
+- New developer asks "why is it built this way?" — nobody remembers
+- Known bugs live in someone's head, not in a file
+- Every AI session starts from scratch
+- Nobody knows the pricing rationale or competitive landscape
+- User frustrations are invisible — product decisions are guesswork
+- Documentation rots because nobody owns it
+
+</td>
+<td width="50%">
+
+**✅ With vibes**
+- AI agent reads `.vibe/` in 3 seconds, understands everything
+- Decisions recorded with rejected alternatives and tradeoffs
+- Issues tracked, resolved issues archived, context is live
+- Every AI session builds on the last one
+- Business strategy, pricing, and market position are documented
+- User personas include emotional reality — fears, confusion, goals
+- Documentation stays fresh because the AI updates it
+
+</td>
+</tr>
+</table>
+
+> **This isn't documentation. It's a digital twin of your entire venture.**<br>
+> Repository memory. Product memory. Business memory. AI instructions. Living context.<br>
+> It's the difference between a repository and a *project that explains itself.*
+
+---
+
+## ✨ The Solution
+
+`vibes` generates two complementary layers:
+
+<table>
+<tr>
+<td width="50%" valign="top">
+
+### 🧠 `.vibe/` — Project Memory
+*15 files covering 5 layers of understanding*
 
 ```
 .vibe/
-├── purpose.md       — What is this? Who is it for? What does it NOT do?
-├── architecture.md  — Systems and how they connect (not files — systems)
-├── flows.md         — User journeys, step by step
-├── entities.md      — Important nouns and their relationships
-├── decisions.md     — Why things exist the way they do
-└── state.json       — Machine-readable project health snapshot
+│
+│ ── Core (repository memory)
+├── purpose.md          What & why
+├── architecture.md     Systems & connections
+├── flows.md            User journeys
+├── entities.md         Important nouns
+├── decisions.md        Why, with dependency graph
+├── state.json          Machine-readable health
+│
+│ ── Living Context
+├── context.md          What's happening NOW
+│
+│ ── AI Agent Guide
+├── ai.md               Safe zones & rules
+│
+│ ── Product Memory
+├── product.md          Vision & retention
+├── users.md            Personas & emotions
+├── metrics.md          What numbers matter
+├── experiments.md      What we're testing
+│
+│ ── Business Memory
+├── business.md         Strategy & revenue
+├── market.md           Competitors & positioning
+└── risks.md            Threats & mitigations
+```
+
+**Not a business?** The AI writes "N/A" and moves on.<br>
+**Not a product?** Same thing. Only relevant files get filled.
+
+</td>
+<td width="50%" valign="top">
+
+### 📄 `docs/` — Operational Layer
+*11 files for implementation-level documentation*
+
+```
+docs/
+├── README.md           Executive summary
+├── topology.md         File & folder map
+├── architecture.md     Component-level details
+├── api.md              API endpoints & formats
+├── issues.md           Open bugs & blockers
+├── resolved.md         Closed issues archive
+├── roadmap.md          Milestones & priorities
+├── developer_guide.md  Setup, build, test, deploy
+├── troubleshooting.md  Common errors & fixes
+├── glossary.md         Project-specific terms
+└── decisions/          Architecture decision records
 ```
 
-## Why?
+<br>
 
-Most repositories are impossible to understand without reading the source code. Documentation describes *how* things work but never captures:
+**`.vibe/` = intent** (survives a rewrite)<br>
+**`docs/` = implementation** (changes with the code)
 
-- **Why** things exist (what alternatives were rejected?)
-- **What** users actually do (step-by-step journeys, not component descriptions)
-- **What** the important nouns are (domain entities, not class hierarchies)
-- **What** the system *is* (intent, not implementation)
+</td>
+</tr>
+</table>
 
-The `.vibe/` folder answers these questions. It describes **intent rather than implementation** — which means it's language-agnostic, framework-agnostic, and survives complete rewrites.
+---
 
-## Quick Start
+## 🚀 Quick Start
 
-### Option 1: npx (no install)
+### Option 1: npx (no install needed)
 
 ```bash
-npx vibes init
+# Scaffold everything at once
+npx vibe-me all
+
+# Or individually:
+npx vibe-me init    # .vibe/ only (15 files)
+npx vibe-me docs    # docs/ only (11 files)
 ```
 
 ### Option 2: Global install
 
 ```bash
-npm install -g vibes
-vibes init
+npm install -g vibe-me
+vibes all
 ```
 
-### Option 3: Clone and run
+### Option 3: One-click (Windows)
 
-```bash
-git clone https://github.com/unlimitedinfinit/Vibes.git
-cd your-project
-node /path/to/Vibes/bin/vibes.js init
+Download [`Vibe.bat`](Vibe.bat), drop it in any project folder, and double-click it.
+
+---
+
+## 🛠️ Commands
+
+| Command | Description |
+|:---|:---|
+| `vibes init` | Create `.vibe/` with the full suite — 15 files + guide |
+| `vibes docs` | Create `docs/` operational documentation — 11 files |
+| `vibes all` | Create both `.vibe/` and `docs/` at once |
+| `vibes check` | Validate all files — catches unfilled templates, empty sections, missing structure |
+| `vibes status` | Quick health dashboard with filled/N-A/empty counts |
+| `vibes reset` | Delete and recreate everything |
+| | |
+| `vibes hub <path>` | Set up a central hub — git inits, creates `_insights/`, saves config |
+| `vibes export` | Export `.vibe/` + `docs/` from current project to your hub |
+| `vibes help` | Show all commands |
+
+---
+
+## 🤖 The AI Workflow
+
+After scaffolding, point your AI coding agent at the project and say:
+
+> *"Read `.vibe/VIBE_GUIDE.md`, then analyze this codebase and fill out all the skeleton files in `.vibe/` and `docs/`. For files that aren't relevant (like `business.md` for a library), write N/A with a brief explanation. Ask me any questions you can't answer from the code."*
+
+The included `VIBE_GUIDE.md` (33KB) contains everything the agent needs:
+- Detailed instructions for each file
+- Examples across 6+ project types (web apps, CLI tools, games, embedded firmware, libraries, mobile)
+- Anti-patterns and common mistakes
+- Quality checklist
+- The discovery protocol (what to scan, what to ask the user)
+
+### The Update Protocol
+
+Every time an AI agent touches a repository with `.vibe/`, the workflow becomes:
+
+```
+Read .vibe/ to understand the project
+         ↓
+    Make code changes
+         ↓
+Update .vibe/ + docs/ to reflect new understanding
+         ↓
+   Commit everything together
 ```
 
-This creates a `.vibe/` folder with skeleton templates and a `VIBE_GUIDE.md` instruction file.
+> The semantic model is a **first-class citizen** — not an afterthought.
+
+---
+
+## 🎯 The Core Principle
+
+> **Describe intent, not implementation.**
+
+Every line in `.vibe/` should pass this test:
+
+*"Would this still be true if we rewrote the entire codebase in a different language?"*
 
-## Usage
+| | Intent ✅ | Implementation ❌ |
+|:---|:---|:---|
+| **Architecture** | "Frontend talks to API over HTTP" | "`UserController` calls `db.query()`" |
+| **Decisions** | "We chose SQLite because <100 writes/hour" | "The `DatabaseService` class has a `connect()` method" |
+| **Flows** | "User uploads document → AI analyzes → results shown" | "`POST /api/docs` calls `processDocument()` in `services/ai.ts`" |
+| **Entities** | "Invoice — created by billing system, immutable after send" | "`Invoice` model has fields `id`, `amount`, `created_at`" |
 
-### Initialize
+---
+
+## 🔍 Quality Validation
+
+`vibes check` doesn't just count lines — it catches real problems:
 
 ```bash
-cd your-project
-vibes init
+vibes check
 ```
 
-Creates `.vibe/` with 6 skeleton files + the full guide. Auto-detects your project name from `package.json`, `Cargo.toml`, `go.mod`, or `pyproject.toml`.
+**On a freshly scaffolded project (before AI fills it out):**
+```
+  🔍 vibes check
 
-### Fill it out (with your AI agent)
+  ── .vibe/ ──
 
-Point your AI coding agent at the project and tell it:
+  ✖ purpose.md — still contains template instructions (unfilled)
+  ✖ architecture.md — still contains template instructions (unfilled)
+  ✖ flows.md — still contains template instructions (unfilled)
+  ...
 
-> "Read `.vibe/VIBE_GUIDE.md`, then analyze this codebase and fill out all the skeleton files in `.vibe/`. Ask me any questions you can't answer from the code."
+  Result: 24 failed, 1 warnings, 0 passed
+```
 
-The guide contains detailed instructions, examples across 6+ project types, anti-patterns, and a quality checklist.
+**After your AI agent fills everything out:**
+```
+  🔍 vibes check
+
+  ── .vibe/ ──
+
+  ✔ purpose.md — 33 lines
+  ✔ architecture.md — 86 lines
+  ✔ flows.md — 144 lines
+  ✔ entities.md — 135 lines
+  ⚠ decisions.md — 91 lines (no dependency graph fields)
+  ✔ state.json — 82 lines
+  ✔ context.md — 28 lines
+  ✔ ai.md — 45 lines
+  ✔ product.md — 62 lines
+  ✔ users.md — 78 lines
+  ✔ metrics.md — 24 lines
+  ✔ experiments.md — 19 lines
+  ⊘ business.md — N/A (not applicable)
+  ⊘ market.md — N/A (not applicable)
+  ✔ risks.md — 34 lines
+
+  Result: 2 N/A, 1 warnings, 22 passed
+```
+
+What it catches:
+- **Unfilled templates** — files that still have `INSTRUCTIONS FOR AI AGENT` (hard fail)
+- **Missing structure** — `purpose.md` without a "NOT do" section, `decisions.md` without dependency graph
+- **Empty bullets/tables** — placeholder content that wasn't filled in
+- **Stale data** — `state.json` not updated in 30+ days
+- **Missing emotional reality** — `users.md` without frustrations, fears, or goals
+- **N/A files** — recognized and not penalized
+
+---
+
+## 🗂️ Vibes Hub — Cross-Project Sync + Intelligence
 
-### Validate
+When you're running vibes across multiple projects (and multiple computers), the **hub** lets you collect, compare, and analyze them all in one place.
+
+### One-time setup
 
 ```bash
-vibes check
+vibes hub ~/Documents/VibeHub
 ```
 
-Checks that all 6 files exist, are non-empty, and pass quality heuristics:
-- `purpose.md` has a "NOT do" section
-- `decisions.md` has 2+ decisions
-- `entities.md` has "What depends on it?" fields
-- `flows.md` has 2+ user flows
-- `state.json` is valid JSON and not stale (>30 days)
+This creates a folder with git, a README, and an `_insights/` intelligence layer — then walks you through connecting it to a private GitHub repo.
 
-### Reset
+### Export from any project
 
 ```bash
-vibes reset
+cd your-project
+vibes export
 ```
 
-Deletes and recreates `.vibe/` with fresh skeletons.
+Copies `.vibe/` and `docs/` into your hub, organized by project name. Auto-generates an `index.md` dashboard.
 
-## The Core Principle
+### Sync across machines
 
-> **Describe intent, not implementation.**
+Push from one machine, pull on another. Same private repo, all your projects.
+
+### Cross-project intelligence
 
-Every line in `.vibe/` should pass this test: *"Would this still be true if we rewrote the entire codebase in a different language?"*
+After exporting 3+ projects, tell your AI agent:
 
-- ✅ "We chose SQLite because the access pattern is <100 writes/hour"
-- ❌ "The `DatabaseService` class has a `connect()` method"
-- ✅ "User uploads document → AI analyzes → results displayed"
-- ❌ "POST /api/documents calls `processDocument()` in `services/ai.ts`"
+> *"Read `_insights/ANALYZE.md` in my Vibes Hub and follow the instructions."*
 
-## The AI Update Protocol
+The AI reads every project's `.vibe/` files and fills out:
 
-Every time an AI agent touches a repository with `.vibe/`, the workflow is:
+| File | What it finds |
+|:---|:---|
+| `_insights/patterns.md` | Architecture, security, and tech stack patterns across projects |
+| `_insights/standards.md` | Universal rules that should apply everywhere |
+| `_insights/opportunities.md` | "Omni does X well → JustLegal should adopt it" |
 
 ```
-Read .vibe/ to understand the project
-  ↓
-Make code changes
-  ↓
-Update relevant .vibe/ files to reflect new understanding
-  ↓
-Commit code + .vibe changes together
+VibeHub/
+├── index.md              Auto-generated dashboard
+├── _insights/
+│   ├── ANALYZE.md        Prompt for AI analysis
+│   ├── patterns.md       Cross-project patterns
+│   ├── standards.md      Universal standards
+│   └── opportunities.md  Transferable improvements
+├── Omni/
+│   ├── .vibe/
+│   └── docs/
+├── JustLegal/
+│   ├── .vibe/
+│   └── docs/
+└── ...
 ```
 
-The semantic model is a first-class citizen — not an afterthought.
+One project's best practice becomes every project's standard.
 
-## Works With
+---
+
+## 🌍 Works With Everything
 
 `.vibe/` is language-agnostic. It works for:
 
-- Python, TypeScript, Go, Rust, C#, Java, C++, Swift
-- React, Next.js, Django, FastAPI, Spring, Rails
-- Unity, Unreal, Godot
-- Embedded firmware, CLI tools, mobile apps, libraries
-- Monorepos, microservices, serverless
+<table>
+<tr>
+<td>
+
+**Languages**
+- Python
+- TypeScript / JavaScript
+- Go
+- Rust
+- C# / .NET
+- Java
+- C / C++
+- Swift
+
+</td>
+<td>
+
+**Frameworks**
+- React / Next.js / Vue
+- Django / FastAPI / Flask
+- Spring / Rails
+- Unity / Unreal / Godot
+- React Native / Flutter
+- Express / NestJS
+- Tauri / Electron
+
+</td>
+<td>
+
+**Project Types**
+- Web applications
+- CLI tools
+- Mobile apps
+- Libraries & packages
+- Microservices
+- Embedded firmware
+- Game engines
+- Monorepos
+
+</td>
+</tr>
+</table>
 
 Because it describes **what** and **why**, not **how**.
 
-## Repository Structure
+---
+
+## 📁 Repository Structure
 
 ```
 Vibes/
-├── bin/vibes.js            — CLI entry point (zero dependencies)
-├── templates/              — Skeleton files copied into .vibe/
-│   ├── purpose.md
+├── bin/
+│   └── vibes.js                    CLI (zero dependencies)
+├── templates/
+│   ├── purpose.md                  ── Core
 │   ├── architecture.md
 │   ├── flows.md
 │   ├── entities.md
-│   ├── decisions.md
-│   └── state.json
+│   ├── decisions.md                   (with dependency graph)
+│   ├── state.json
+│   ├── context.md                  ── Living Context
+│   ├── ai.md                       ── AI Agent Guide
+│   ├── product.md                  ── Product Memory
+│   ├── users.md
+│   ├── metrics.md
+│   ├── experiments.md
+│   ├── business.md                 ── Business Memory
+│   ├── market.md
+│   ├── risks.md
+│   └── docs/                       ── Operational Docs
+│       ├── README.md
+│       ├── topology.md
+│       ├── architecture.md
+│       ├── api.md
+│       ├── issues.md
+│       ├── resolved.md
+│       ├── roadmap.md
+│       ├── developer_guide.md
+│       ├── troubleshooting.md
+│       ├── glossary.md
+│       └── decisions/
+│           └── 0001-template.md
 ├── spec/
-│   └── VIBE_GUIDE.md       — Full spec with examples and instructions
-├── QUICK_START_PROMPT.md    — Copy-paste prompt for AI agents
-├── VIBE_GUIDE.md            — Full spec (also at root for easy reading)
+│   └── VIBE_GUIDE.md               Full spec (bundled on init)
+├── Vibe.bat                         One-click Windows installer
+├── QUICK_START_PROMPT.md            Copy-paste prompt for AI agents
+├── VIBE_GUIDE.md                    Full spec (readable at root)
 ├── package.json
-├── LICENSE                  — MIT
-└── README.md                — This file
+├── LICENSE
+└── README.md
 ```
 
-## License
+---
+
+## 📜 License
 
 MIT — use it, fork it, build on it.
 
-## Contributing
+---
+
+<div align="center">
+
+## 💡 The Paradigm Shift
+
+<br>
+
+Most AI coding tools today operate in a simple loop:
+
+**Read Code → Write Code → Hope It Works**
+
+The AI has no memory. No context. No understanding of *why* things are the way they are.<br>
+Every session starts from zero. Every agent makes the same mistakes.
+
+With vibes, the loop transforms:
+
+```
+     ┌─────────────────────────────────────┐
+     │                                     │
+     ▼                                     │
+Read .vibe/ → Understand Intent            │
+     │                                     │
+     ▼                                     │
+  Read Code → Make Changes                 │
+     │                                     │
+     ▼                                     │
+Update .vibe/ + docs/ ─────────────────────┘
+```
+
+**The semantic model becomes a first-class citizen.**
+
+It's not generated once and forgotten. It's not a README that rots.<br>
+It's a living, continuously-maintained digital twin that gets smarter every time anyone — human or AI — touches the project.
+
+<br>
+
+---
+
+### 🔮 The Vision
+
+Imagine a world where:
+
+- You open **any** repository and instantly understand what it does, why it exists, and what's broken — before reading a single line of code
+- Your AI agent picks up **exactly** where the last one left off — with full context of every decision, every experiment, every user frustration
+- New team members are productive in **hours**, not weeks
+- Product decisions are grounded in documented user reality, not guesswork
+- Business strategy, pricing rationale, and competitive positioning live next to the code — not in a forgotten Google Doc
+- Documentation **never** goes stale because the AI updates it as part of its normal workflow
+
+That world starts with one command.
+
+<br>
+
+```bash
+npx vibe-me all
+```
+
+<br>
+
+---
+
+*Created by [Joshua M. Abrams](https://github.com/unlimitedinfinit)*
+
+*The `.vibe` spec is version 1.0 — tested across production repositories spanning Go, Rust, TypeScript, Python, and Solidity.*<br>
+*If you create `.vibe/` for your project, [open an issue](https://github.com/unlimitedinfinit/Vibes/issues) and share what you learned.*
+
+<br>
 
-The `.vibe` spec is version 1.0. After testing across 20+ repositories, we'll refine which information is always useful, which is noise, what AI can extract automatically, and what humans must provide.
+**Stop giving your AI agents amnesia. Give them vibes.**
 
-If you create `.vibe/` for your project, open an issue and share what you learned.
+</div>

package/Vibe.bat

@@ -2,19 +2,20 @@
 :: ──────────────────────────────────────────────
 :: Vibe.bat — One-click project setup
 :: Drop this file in any project root and double-click.
-:: Creates .vibe/ semantic layer + docs/ operational docs.
+:: Creates .vibe/ (15 files) + docs/ (11 files)
 :: ──────────────────────────────────────────────
 
 echo.
-echo   ⚡ Vibe.bat — Setting up semantic layer + docs
-echo   ─────────────────────────────────────────────────
+echo   ======================================================
+echo      vibes — The complete semantic layer for any project
+echo   ======================================================
 echo.
 
 :: Check if Node.js is installed
 where node >nul 2>nul
 if %ERRORLEVEL% neq 0 (
-    echo   ✖ Node.js is not installed.
-    echo     Download it from https://nodejs.org
+    echo   ERROR: Node.js is not installed.
+    echo   Download it from https://nodejs.org
     echo.
     pause
     exit /b 1
@@ -26,14 +27,17 @@ echo.
 call npx -y vibe-me@latest all
 
 echo.
-echo   ─────────────────────────────────────────────────
-echo   ✔ Done! Now tell your AI agent:
+echo   ======================================================
+echo   DONE! Your .vibe/ and docs/ folders are ready.
+echo.
+echo   Next step — tell your AI coding agent:
 echo.
 echo     "Read .vibe/VIBE_GUIDE.md, then analyze this
 echo      codebase and fill out all skeleton files in
-echo      .vibe/ and docs/. Ask me any questions you
-echo      can't answer from the code."
+echo      .vibe/ and docs/. For files that aren't relevant
+echo      write N/A with a brief explanation. Ask me any
+echo      questions you can't answer from the code."
 echo.
-echo   ─────────────────────────────────────────────────
+echo   ======================================================
 echo.
 pause