Micromanaged Driven Development (MMDD)

Yes, we said micromanaged - and this time it makes sense.

Built by and for developers who refuse to lose control of AI-generated code, MMDD is an AI coding workflow packaged as a super prompt.

Unlike tool-specific approaches, it works with any AI tool because it's pure Markdown — giving you a permanent, human-readable audit trail of every decision, inside your own repo.

Quick Start

1. Copy 00_mmdd.md into your project's dev_log/ directory
2. Tell your AI: "Using the workflow in dev_log/00_mmdd.md, generate a 00_main.md file for my project that [describe your project]"
3. Work through units one at a time, approving every step

That's it. 00_mmdd.md is the super prompt — it instructs the AI on structure, workflow, and documentation. This README is for you.

What is MMDD?

Instead of giving broad, vague prompts like "build me a web app," MMDD breaks development into specific, well-defined tasks with systematic documentation of the entire process.

Every unit of work follows a 5-step cycle: create context, plan, implement, validate, commit. The AI suggests; you decide. Nothing gets implemented without your approval, and everything that happens is logged in a chronological audit trail.

Key principles: granular task breakdown, controlled AI direction, documentation as the source of truth. Context is cheap, confusion is expensive.

Repository Structure

This repository demonstrates MMDD by using MMDD to build itself:

├── README.md                # This introduction file (for humans)
└── dev_log/                 # Chronological development units
    ├── 00_mmdd.md           # MMDD workflow, v5 (for your AI — the super prompt)
    ├── 00_kanban.md         # Kanban extension (optional, for unit status tracking)
    ├── 00_main.md           # This project's development plan
    ├── 01_repository.md     # Repository structure and templates unit
    ├── 02_content.md        # Content terminology and documentation unit
    ├── 05_mmdd.md           # Abbreviation change unit
    ├── 06_smallchanges.md   # File renaming and cleanup unit
    ├── 07_workflow.md       # Workflow documentation unit
    ├── 08_types.md          # Version 4 update unit
    ├── 09_optimization.md   # Version 5: AI-optimized workflow unit
    ├── 10_optimization.md   # Unit 10: AI-optimized rewrite of 00_mmdd.md
    └── 11_kanban.md         # Unit 11: Kanban extension creation

How It Works

00_mmdd.md is written for AI consumption: compact, imperative, no persuasion — just the structure, templates, and workflow rules the AI needs in context every session. The explanations, rationale, and adoption guidance live here in the README, where they cost zero tokens.

Want to understand the workflow in depth? Read dev_log/00_mmdd.md — it's short by design. Then check dev_log/00_main.md to see how this very repository was planned and built with it.

Why MMDD?

• Predictable AI Results: Granular control leads to more reliable outputs
• Maintainable Code: Systematic approach produces cleaner, understandable systems
• Knowledge Transfer: Detailed logs help other developers (and future you) understand the codebase
• Quality Assurance: Regular validation points catch issues early

Contributing

These principles evolve through real-world usage. Share your experiences, improve templates, or contribute examples of MMDD in action.

License

MIT License - Use freely in your own projects.

---

Want to see MMDD in practice? This entire repository was built using these principles - explore the dev_log/ directory to see how.