mastery.md
Source Preview

MASTERY.md

A universal, tech-agnostic development process framework. One Markdown file. Any project. Any language. Any stack. Any team size.

v3.3.0 MIT License Open Source

Download mastery.md Download compact GitHub

What is it?

MASTERY.md is a complete development process framework distributed as Markdown files. Copy them into your project's docs/ folder and follow the lifecycle. Every feature starts as a discussion, gets designed, becomes a plan, and ships as clean, tested code.

It works for both humans and AI agents. The built-in AI Agent Protocol gives AI models structured context, clear autonomy boundaries, and a disciplined process to follow.

Two Variants

Variant Tokens Use When
mastery.md (full) ~25k You want the complete framework with all 16 templates — solid, thorough development
mastery-compact.md ~5k You want to save tokens and still get awesome development — all rules, no templates

AI agents load the compact variant by default and pull templates from the full file only when needed. 80% token savings on every context load.

The 7-Stage Lifecycle

Discuss Design Plan Approve Build Ship Reflect

Every feature follows the same disciplined path from idea to production. No shortcuts. No skipped steps.

Quick Start

terminal
git clone https://github.com/raiworks/mastery.md.git
cp MASTERY.md/mastery.md your-project/docs/
cp MASTERY.md/mastery-compact.md your-project/docs/

Then create your project docs:

  1. docs/project-discussion.md - Capture WHY the project exists
  2. docs/project-context.md - Formalize WHAT the project is
  3. docs/project-roadmap.md - Plan WHERE you're going

Start your first feature: docs/features/01-feature-name/discussion.md

Already Have a Project?

MASTERY.md works just as well on existing codebases as on new ones. You don't need to rewrite anything or pause development. The mid-project adoption flow preserves your existing work while bringing structure going forward.

Archive Reconstruct Catalog Build Forward
  1. Archive existing docs to docs/_archive/ — preserved, never deleted
  2. Reconstruct project discussion & context from codebase, git history, and team knowledge
  3. Catalog completed features as lightweight summary.md files — no need to create full lifecycle docs for past work
  4. Build forward — the next feature follows the full lifecycle, and every feature after that

Your project gets structured docs, AI-readable context, and a clear roadmap — without disrupting anything that's already working.

Why MASTERY.md?

Tech-Agnostic

Works with any language, any framework, any stack. From Go to Python to Rust to JavaScript.

AI-Native

Built-in AI Agent Protocol. AI agents load context, follow autonomy boundaries, and hand off sessions cleanly.

Single File

One Markdown file defines the entire process. No tools to install, no SaaS to sign up for, no vendor lock-in.

Any Team Size

Solo developer or 100-person team. The framework scales without changing the process.

Full Lifecycle

From discussion to deployment. Includes templates, git strategy, commit conventions, and hotfix workflows.

Open Source

MIT licensed. Free forever. Use it, modify it, share it. The framework belongs to the community.

MCP Integration

The mastery-mcp-server gives AI agents native access to your Mastery docs through the Model Context Protocol. 14 tools and 11 resources for reading context, tasks, roadmaps, test plans, and more — including the compact variant.

mcp-config.json
{
  "mcpServers": {
    "mastery": {
      "command": "mastery-mcp-server",
      "args": ["--project-dir", "."]
    }
  }
}
  • AI agents load project context, roadmap, and feature state automatically
  • Read discussion, architecture, tasks, test plans, and reviews
  • Works with VS Code (Copilot), Claude Desktop, Cursor, and any MCP client
  • Learn more →

# MASTERY.md
> A universal, tech-agnostic development process framework.
> One Markdown file. Any project. Any language. Any stack. Any team size.
`v3.3.0` `MIT License` `Open Source`
## What is it?
MASTERY.md is a complete development process framework distributed as
Markdown files. Copy them into your project's `docs/` folder
and follow the lifecycle.
### Two Variants
| Variant | Tokens | Use When |
|---|---|---|
| **mastery.md** (full) | ~25k | Complete framework with all 16 templates |
| **mastery-compact.md** | ~5k | All rules, no templates — 80% token savings |
It works for **both humans and AI agents**. The built-in AI Agent
Protocol gives AI models structured context, clear autonomy boundaries,
and a disciplined process to follow.
## The 7-Stage Lifecycle
`Discuss``Design``Plan``Approve``Build``Ship``Reflect`
Every feature follows the same disciplined path from idea to production.
No shortcuts. No skipped steps.
## Quick Start
```bash
git clone https://github.com/raiworks/mastery.md.git
cp MASTERY.md/mastery.md your-project/docs/
cp MASTERY.md/mastery-compact.md your-project/docs/
```
Then create your project docs:
1. `docs/project-discussion.md` - Capture WHY the project exists
2. `docs/project-context.md` - Formalize WHAT the project is
3. `docs/project-roadmap.md` - Plan WHERE you're going
Start your first feature: `docs/features/01-feature-name/discussion.md`
## Already Have a Project?
MASTERY.md works just as well on **existing codebases** as on new ones.
You don't need to rewrite anything or pause development.
`Archive``Reconstruct``Catalog``Build Forward`
1. **Archive** existing docs to `docs/_archive/`
2. **Reconstruct** project discussion & context from codebase + git history
3. **Catalog** completed features as lightweight `summary.md` files
4. **Build forward** — next feature follows the full lifecycle
Your project gets structured docs, AI-readable context, and a clear
roadmap — without disrupting anything that's already working.
## Why MASTERY.md?
- 🌐 **Tech-Agnostic** - Any language, any framework, any stack
- 🤖 **AI-Native** - Built-in AI Agent Protocol with autonomy boundaries
- 📄 **Single File** - One Markdown file. No tools, no SaaS, no lock-in
- 👥 **Any Team Size** - Solo to enterprise, same process
- 🛠️ **Full Lifecycle** - Discussion to deployment, every step covered
- 🔓 **Open Source** - MIT licensed. Free forever
## MCP Integration
The **mastery-mcp-server** gives AI agents native access to your
Mastery docs through the Model Context Protocol. **14 tools** and
**11 resources** for reading context, tasks, roadmaps, and more.
```json
{
"mcpServers": {
"mastery": {
"command": "mastery-mcp-server",
"args": ["--project-dir", "."]
}
}
}
```
- AI agents load project context, roadmap, and feature state automatically
- Read discussion, architecture, tasks, test plans, and reviews
- Works with VS Code (Copilot), Claude Desktop, Cursor, and any MCP client
- [Learn more →](/mcp-server)
---
[View on GitHub](https://github.com/raiworks/mastery.md) · [MCP Server](https://github.com/raiworks/mastery-mcp-server) · [Share Feedback](/feedback)