Greenfield to Production

This guide walks through the complete Arness lifecycle — from a raw product idea to a deployed application. It uses all three plugins: Spark for discovery, Code for development, and Infra for deployment.

Phase 1: Spark — From Idea to Validated Concept

Start the Journey

/arn-brainstorming

The brainstorming wizard guides you through the entire greenfield pipeline with decision gates at every step. You can also invoke each skill individually.

Discover Your Product

Spark’s product strategist agent shapes your raw idea into a structured vision through conversation. You describe what you want to build; the agent probes for gaps, challenges assumptions, and structures the concept into:

• Vision and problem statement
• Target personas with detailed profiles
• Competitive landscape
• Product pillars (your non-negotiable principles)

Stress Test the Concept

Before writing code, Spark puts your concept through four adversarial lenses:

• Pre-mortem — assumes your product failed, works backward to identify why
• Competitive gap analysis — maps your feature matrix against real competitors
• Synthetic user interviews — AI personas (Pragmatist, Skeptic, Power User) probe your concept
• PR/FAQ stress test — drafts your launch story, then tears it apart

A concept review consolidates all findings and updates your product vision.

Name and Define

• Brand naming — strategic foundation, creative generation, scoring, and WHOIS domain checks
• Architecture vision — technology evaluation, stack selection, system design
• Use cases — formal Cockburn fully-dressed use cases with expert review

Prototype and Validate

• Scaffold — generate a working project skeleton with your chosen stack
• Visual sketches — multiple design direction proposals as live HTML/CSS
• Clickable prototype — interactive screens with Playwright journey testing
• Lock — freeze the validated prototype as a development reference

Extract Features

/arn-spark-feature-extract

Feature extraction pulls a prioritized backlog from all your Spark artifacts — product concept, architecture, use cases, prototypes. Features are written as structured files and optionally uploaded to GitHub Issues, Jira, or Bitbucket.

Phase 2: Code — From Features to Shipped Code

Hand Off to the Development Pipeline

/arn-planning

Run this in your scaffolded project. Arness automatically analyzes the codebase patterns Spark established and sets up the development pipeline on first use. You can optionally run /arn-code-init first for more control over the configuration.

Build Features from the Backlog

Arness detects the Spark feature backlog and lets you pick a feature to implement. From there, the standard development pipeline takes over: spec → plan → execute → review → ship.

For multiple features, the batch pipeline can plan and implement several features in parallel:

/arn-code-batch-planning
/arn-code-batch-implement
/arn-code-batch-merge

Ship Each Feature

/arn-shipping

Each feature gets its own PR with a structured description referencing the spec and plan.

Phase 3: Infra — From Code to Cloud

Set Up Infrastructure

/arn-infra-wizard

The wizard auto-configures and walks through the infrastructure pipeline:

1. Discover — audits your tools and auth state
2. Containerize — generates Dockerfiles and docker-compose for your application
3. Define — produces IaC (Terraform, Pulumi, CDK, etc.) for your cloud provider
4. Configure — sets up environments (dev/staging/production) and secrets management
5. Pipeline — generates CI/CD workflows
6. Deploy — deploys with cost and security gates
7. Verify — health checks, DNS, SSL validation
8. Monitor — sets up logging, metrics, and alerting

Structured Changes

For subsequent infrastructure changes, use the change management pipeline:

/arn-infra-change-spec

This mirrors Code’s spec → plan → execute → review flow but with infrastructure-specific gates: blast radius assessment, rollback checkpoints, security scanning, and cost estimation.

The Complete Artifact Trail

By the end of this journey, your project contains:

Phase	Artifacts
Spark	Product concept, stress test reports, naming brief, architecture vision, use cases, prototype screenshots, feature backlog
Code	Feature specs, implementation plans, execution reports, review findings, PR descriptions
Infra	Tooling manifest, Dockerfiles, IaC code, environment configs, CI/CD workflows, deployment reports, runbooks

Every artifact is plain Markdown or JSON — readable without Arness, committed to your repo, and serving as both AI context and human documentation.

Next Steps

• Arness Spark — deep dive into the discovery pipeline
• Arness Code — deep dive into the development pipeline
• Arness Infra — deep dive into the infrastructure pipeline