bmad-code-org/BMAD-METHOD

What it solves

BMad Method addresses the tendency of AI coding assistants to implement unstated assumptions into code. It provides a structured, agile framework for AI-Driven Development (AiDD) that ensures product and technical decisions remain explicit, durable, and under human control throughout the software delivery lifecycle.

How it works

BMad implements a delivery loop consisting of Clarify, Plan, Build, and Learn/Adjust. It integrates with AI coding tools (like Claude Code or Codex) via plugins or a CLI to provide specialized agentic perspectives (product, architecture, UX, development, and testing). The process is "right-sized," meaning users can skip directly to implementation for simple changes or use deep planning workflows for complex initiatives, preserving the context of decisions as the project evolves.

Who it’s for

Software developers and teams who use AI coding assistants and want a more disciplined, agile approach to turning ideas or change requests into working software, whether starting a new prototype or managing an existing codebase.

Highlights

• Durable Context: Preserves technical and product decisions so they don't need to be re-explained in every chat session.
• Right-Sized Process: Flexible workflows that scale from small, immediate builds to complex, multi-stage planning.
• Specialized Perspectives: Access to dedicated agent roles for architecture, UX, and testing to guide collaboration.
• Tool Agnostic: Compatible with various AI coding tools through a plugin marketplace and a Skills CLI.