Propose the BMAD process for team of developers; leveraging on the git branches to facilitate parallel features development.
- Setup Next.js
npx create-next-app@latest .
- install bmad
npx bmad-method install
- Setup
.gitignore
# BMAD
/_bmad
/.cursor
/.claude
/.codex
- DO NOT checkin BMAD install and all the correspoding artifacts because it can be
re-installed:
_bmad,.cursor,.claude,.codex, etc.
-
On the
mainbranch, "Analysts" do:- Brainstorming
- Research, and
- Product Brief
-
Next is the Planning phase, where the "Product Manager" creates PRD
-
Still in the Planning phase, the UX Designer creates the following based on the PRD
docs/planning-artifacts/ux-design-specification.mddocs/planning-artifacts/ux-design-directions.html
-
Next is the SOLUTIONING phase, the Architect creates
architecture.mdbased on the PRD and UX Design specs -
With the architure document. PM can create the EPICS & Stories and store in the
docs/planning-artifacts/epics/md -
Finally the Implementation phase, the Scrum Master creates the sprint planning and store it in
docs/implementation-artifacts/sprint-status.yamlthat looks like the following, for example,:
development_status:
epic-1: backlog
1-1-user-registration-email-password: backlog
1-2-user-login-logout: backlog
1-3-onboarding-flow: backlog
1-4-dashboard-landing: backlog
epic-1-retrospective: optional
epic-2: backlog
2-1-microphone-setup-permissions: backlog
2-2-speaker-volume-test: backlog
2-3-quick-start-path-dive-right-in: backlog
2-4-myqueue-path-difficulty-settings-see-the-setup-first: backlog
epic-2-retrospective: optional
epic-3: backlog
3-1-real-time-speech-to-text-assemblyai: backlog
3-2-text-to-speech-output-google-chirp-3-hd: backlog
3-3-claude-conversation-logic-turn-taking: backlog
3-4-session-state-conversation-interface: backlog
epic-3-retrospective: optional
epic-4: backlog
4-1-post-session-feedback-grounded-to-sharpen: backlog
4-2-audio-playback-replay-session-recording: backlog
4-3-session-history-list-feedback-playback: backlog
epic-4-retrospective: optional
- [OPTIONAL REVISION] if an Epic is too big, or a Story is too big, the PM has the option to break it into multiple epics or stories.
-
Once the
docs/implementation-artifacts/sprint-status.yamllooks good, the Developer agent create a Git Branch for development, for example, "epic-3" branch -
Go to the branch, Scrum Master will create the full story
-
Then, Developer agent will start developing the story