Claude Sonnet AI Development PartnerThis site was built using automated SDLC workflow.
// AI-Augmented Development
From Idea to Deployment: A Practical
AI-Augmented Development Workflow
A platform of tools, experiments, and real-world implementations — focused on building reliable, scalable human – AI collaboration systems
Claude Sonnet AI Development PartnerFramework
How the work gets built.
The AI-fluency methodology behind every build — evolution, the four disciplines, prompt engineering, the SDLC, and the knowledge base.
Effective AI collaboration starts with a clear division of responsibility. Not everything should be delegated, and not everything should be manual. Every task begins with sbx sdlc init — I define the scope, choose which applications and packages are affected, set the action type. The system derives everything else: which steps apply, which domain experts to load, which standards are relevant.
- Defines goals and user stories
- Makes architectural decisions
- Reviews, approves, deploys
- Strategy, UX, business logic
- Implements code from specifications
- Follows governed constraints
- Runs builds, tests, checks
- Step-by-step criteria execution
The quality of AI output is bounded by the quality of context it receives. Manual prompt engineering doesn't scale — you can craft a perfect prompt once, but not hundreds of times across a real project. SBX replaces manual prompting with automated context engineering. Each SDLC step generates a structured prompt with contextual knowledge injection.
AI can produce plausible-looking output that's subtly wrong. Discernment means building evaluation into the process itself, not relying on human vigilance alone. The Description-Discernment loop is embedded in the SDLC state machine — not optional, not skippable. Six gates stand between AI output and production.
When AI helps produce your work, you're still responsible for it. Every decision is immutable and auditable. Every exported function traces to a user story. PR workflow is mandatory — direct push to main is blocked by hooks. This website is the diligence statement: built entirely with AI-augmented development, every commit traceable.
Process
From idea to production, governed end to end.
The end-to-end development process, grounded in a catalogue of industry standards and practices — from goal definition through model, plan, build, and verification.
Packages
Modular building blocks shared across every project.
Components
A reusable interface library composed across every page.
Games
Playable browser games built through the AI-augmented process.
Knowledge
Standards, patterns, and references that guide every build.
A Practical Guide to Feature-Driven Development
The definitive guide to Feature-Driven Development (FDD) — five-process methodology, chief programmer roles, feature lists, design by feature, and build by feature. Primary source for the SBX SDLC governance framework.
Clean Architecture: A Craftsman's Guide to Software Structure and Design
Foundational text on software architecture principles — Dependency Rule, Clean Architecture layers (Entities, Use Cases, Interface Adapters, Frameworks), and component coupling laws. Directly influenced the SBX 4-layer FDD monorepo architecture (PD/MD/SI/UI layers).
Design Patterns: Elements of Reusable Object-Oriented Software
The original pattern catalogue: 23 GoF patterns across Creational, Structural, and Behavioral categories. Observer, Delegate, Factory, Coordinator, Repository — patterns that appear throughout every iOS project built across 20+ years.
Clean Code: A Handbook of Agile Software Craftsmanship
Practical guide to writing readable, maintainable code — meaningful names, small functions, single responsibility, error handling, and test-driven development. Applied daily across every language and platform.
Blog
Notes and write-ups from building the framework, the components, and the projects.