BMad-Method: Universal AI Agent Framework

Agentic planning + context-engineered development workflow for high-quality AI-assisted delivery.

Intermediate<1dayModerateMedium
0
Stars

Video Walkthrough

Video will load when visible

The BMad-Method provides two core innovations for AI-assisted delivery:

  1. Agentic Planning: Analyst, PM, and Architect collaborate to create consistent PRD and Architecture documents using tasks, templates, and knowledge base resources.
  2. Context-Engineered Development: A Scrum Master (SM) converts sharded artifacts into hyper-detailed story files that the Dev agent executes, with QA providing structured review and refactoring.

Core components live under bmad-core/ (agents, agent-teams, workflows, templates, tasks, checklists, data) and are packaged for IDE and Web UI environments. The method emphasizes lean context, explicit dependencies, and gated workflows to reduce planning inconsistency and context loss.

Install (IDE): npx bmad-method install Key roles: Analyst, PM, Architect, PO, SM, Dev, QA, UX Expert.

Flows

Related lifecycles

Maintained by
BMadCode / Brian (BMad) Madison
License
MIT
Last updated
Aug 12, 2025
Versions
4.36.2
Tags
agilescrumdevopsplanningdevelopmenttestingmaintenanceagenticworkflow
Requirements
JavaScriptTypeScriptMarkdown+1 more
Node.js >=20npm or npx+2 more
Team: Any size
View on GitHub