Greenfield Workflow — Smart SDLC

This workflow covers net-new products built from a blank canvas — no existing codebase to onboard to, no legacy constraints. Everything is defined and discovered through the SDLC process itself.

Applies to project type: greenfield

Workflow Overview

[Product / BA]    Define the problem, research the market, create product brief
   ↓
[Product / BA]    Write the PRD (requirements, users, features, constraints)
   ↓
[Architect]       Design the system architecture and key technical decisions
   ↓
[Architect / TL]  Break into epics and user stories
   ↓
[Team Lead]       Sprint planning, capacity, prioritisation
   ↓
[Developer]       Implement story by story using TDD

Greenfield is the cleanest path through Smart SDLC — no pre-existing archaeology or legacy constraints. The entire design is produced forward from first principles.

Phase 0 — Project Initialisation

Who runs it: Team Lead or PM (once per project)

npx @supermldev/smart-sdlc init

During init, select:

• Project type: greenfield
• AI tool: GitHub Copilot, Claude, Cursor, or other
• Team names: Customise the name for each persona (e.g. rename Aria to “Jordan”)
• Integrations: Enable JIRA / Confluence / GitHub if applicable
• Artifact paths: Where planning and implementation outputs are saved

Then each team member runs their personal setup:

npx @supermldev/smart-sdlc persona

Phase 1 — Product Discovery

Persona: product Goal: Understand the problem deeply before writing a single requirement.

Step 1 — Activate the BA Agent

@sml-agent-pm

Aria greets you and presents the discovery menu.

Step 2 — Elicitation (optional but valuable)

/sml-elicitation

If the problem is fuzzy or assumptions are unvalidated, run the elicitation skill first. Aria will guide you through:

• Stakeholder mapping (who needs this, who is affected)
• Five-whys analysis (why is this a problem, root causes)
• Assumption surfacing (what are we taking for granted)
• Jobs-to-be-done framing

Output: A set of validated problem statements and assumptions to carry forward into the product brief.

Step 3 — Brainstorming (optional)

/sml-brainstorming

Structured ideation using diverge/converge technique. Useful if the solution space is wide and the team wants to explore before committing to an approach.

Step 4 — Create Product Brief

/sml-product-brief

A focused brief that captures:

• The problem and why it matters
• Target user segments and their goals
• Success metrics (quantitative and qualitative)
• Scope boundaries — what’s in, what’s explicitly out
• Risks and open questions

Output artifact: _superml-output/planning/product-brief.md

Step 5 — Create the PRD

/sml-create-prd

Aria runs the 3-step guided PRD workflow:

Output artifact: _superml-output/planning/prd-v1.md

Mark prd_complete: true in _superml/config.yml when the PRD is reviewed and approved.

Phase 2 — Architecture & Technical Design

Persona: architect Prerequisite: prd_complete: true

Step 1 — Activate the Architect

@sml-agent-architect

Rex loads the PRD and presents the solutioning menu.

Step 2 — Architecture document

/sml-create-architecture

Starting from a blank canvas, Rex designs:

Output artifact: _superml-output/planning/architecture-v1.md

Mark architecture_complete: true in _superml/config.yml when approved.

Step 3 — Generate project context

/sml-generate-context

Produces project-context.md — a compact, AI-optimised project summary. Every agent loads this on activation to maintain consistency across sessions.

Step 4 — Epics and stories

/sml-create-epics-stories

Rex reads the PRD and architecture, then breaks the product into delivery units:

1. Epics — major functional areas (e.g. User Authentication, Payment Processing)
2. Stories — independent, deliverable user stories per epic
3. Acceptance criteria — testable conditions per story
4. Technical notes — implementation guidance referencing architectural decisions
5. Dependencies — which stories must be done before others
6. Suggested sequence — phasing recommendation

Output artifact: _superml-output/planning/epics-stories-v1.md

Mark epics_complete: true in _superml/config.yml when done.

Phase 3 — Sprint Planning

Persona: team_lead Prerequisite: epics_complete: true

Step 1 — Activate the Team Lead

@sml-agent-lead

Step 2 — Plan the first sprint

/sml-sprint-planning

For greenfield projects, Sprint 1 is typically focused on foundation stories:

• Project scaffolding (repo, CI/CD, environments)
• Core infrastructure (auth, database, logging)
• Walking skeleton — the simplest possible end-to-end flow

Team Lead guides you through:

1. Backlog review — which stories are ready (no unresolved dependencies)
2. Capacity — team size, days available, expected interruptions
3. Story selection — focus on highest-value, least-dependent stories
4. Sprint goal — one sentence describing what the sprint achieves

Output artifact: _superml-output/implementation/sprint-1.md

If JIRA is enabled, the sprint is created and stories are assigned automatically.

Phase 4 — Implementation

Persona: developer Prerequisite: Sprint stories are planned

Step 1 — Activate the Developer Agent

@sml-agent-developer

Nova loads project-context.md, the sprint doc, and the epics file, then presents the story queue.

Step 2 — Pick a story and implement

/sml-dev-story

For greenfield projects, Nova applies a test-first discipline from the very first line of code:

Step 3 — Code review

/sml-code-review

At PR time, Nova reviews the diff against:

• Acceptance criteria from the story
• Architectural conventions (ADRs)
• Team coding standards from _superml/reference/developer/
• Security considerations (OWASP surface scan)

Recommended Sprint Sequence for Greenfield

Cross-Persona Collaboration

npx @supermldev/smart-sdlc meeting

Artifact Chain

product-brief.md
    → prd-v1.md                        (prd_complete: true)
        → architecture-v1.md           (architecture_complete: true)
        → project-context.md
            → epics-stories-v1.md      (epics_complete: true)
                → sprint-1.md
                → sprint-2.md
                    → implementation

Greenfield vs General Workflow

Other Workflows

Not the right workflow for your project type? Choose below: