AI-Native Engineering

Flutter AI Engineering @ VGV

The AI-native engineering platform that encodes 8+ years of Flutter expertise into structured, production-proven workflows. From open-source plugins to enterprise automation.

Explore the Platform View on GitHub

claude-code ~ vgv-wingspan

$ /brainstorm

Exploring problem space... gathering requirements

$ /plan

Generating implementation plan with 12 tasks

$ /build

Executing plan... applying Flutter Plugin skills

✓ 13 skills active across 4 categories

$ /review

All checks passed. Ready for PR.

Not Just AI Tools. An AI-Native Platform.

While others bolt AI onto existing workflows, VGV is building a comprehensive engineering platform that encodes production-proven patterns directly into AI-powered tools. From open-source plugins to enterprise automation—every layer is designed to work together.

The Platform

Three Tools. One Unified Vision.

From individual developer productivity to team-wide engineering workflows to enterprise-scale automation—VGV's AI tooling covers the full spectrum.

Open Source

VGV AI Flutter Plugin

Open-source Claude Code plugin with 13 expert skills across architecture, quality, UI, and tooling. Encodes production-proven Flutter conventions into every Claude Code session.

View on GitHub

Open Source

VGV Wingspan

Agentic engineering workflow for Claude Code. Four structured phases—brainstorm, plan, build, review—that replace ad-hoc prompting with repeatable, auditable processes.

View on GitHub

Partners Only

VGV Arcana

Enterprise-scale agentic tooling for native-to-Flutter migrations, comprehensive code assessments, and Figma-to-code patterns at scale.

Get in Touch

flutter-plugin

# VGV AI Flutter Plugin

✓ Loaded 13 expert skills

✓ 4 categories active

→ Architecture & Patterns (4)

→ Quality & Testing (4)

→ UI & Design (3)

→ Tooling & Compliance (2)

✓ All best practices enabled

Ready. Production conventions active for this session.

Open Source

13 Expert Skills. One Plugin.

The VGV AI Flutter Plugin brings production-proven conventions directly into your Claude Code workflow. 13 skills across four categories encode the patterns behind apps used by millions.

View on GitHub

Architecture & Patterns

BLoC State Management

Proper BLoC/Cubit patterns, event-driven architecture, and reactive state handling

Layered Architecture

Clean separation of data, domain, and presentation layers

Navigation

Type-safe routing, deep linking, and navigation state management

Project Scaffolding

Opinionated project structure and folder conventions from Very Good CLI

Quality & Testing

Testing

Widget, unit, and integration test patterns with high coverage targets

Accessibility

WCAG compliance, semantic widget trees, and screen reader support

Static Security

Security linting rules, vulnerability patterns, and safe coding practices

Analysis Upgrades

Automated migration to latest very_good_analysis lint rules

UI & Design

Material Theming

Consistent Material 3 theming, color schemes, and typography scales

UI Packages

Reusable, self-contained UI component packages with golden tests

Internationalization

Localization setup with flutter_localizations and ARB-based l10n

Tooling & Compliance

SDK Upgrades

Safe Dart and Flutter SDK version migration with deprecation fixes

License Compliance

Dependency license auditing and OSS compliance enforcement

Agentic Workflows

From Brainstorm to Review. Structured, Not Ad-Hoc.

VGV Wingspan replaces unstructured prompting with a four-phase engineering workflow for Claude Code. Stack-agnostic by design, it works for Flutter, web, native, and embedded projects—and pairs with the Flutter Plugin for end-to-end coverage.

View on GitHub

vgv-wingspan ~ agentic SDLC

/brainstorm

→ 8 requirements captured, 3 edge cases identified

/plan

→ 12-task plan approved, architecture validated

/build

→ 12/12 tasks complete, all tests passing

/review

→ 0 issues found. Ship it.

Workflow

Four Phases. Zero Guesswork.

/brainstorm

Explore the problem space, gather context, surface requirements, and identify edge cases. The agent asks the right questions before writing any code.

/plan

Generate a detailed implementation plan with file-level changes, dependencies, and sequencing. Review and approve before any code is written.

/build

Execute the plan step by step. The agent writes production-quality code following the approved architecture and project conventions.

/review

Automated review against best practices, test coverage, accessibility, and performance. Catches issues before they reach PR.

Partners Only

Enterprise-Scale AI Engineering

VGV Arcana takes specialized knowledge and agentic engineering beyond individual productivity. Built on the same production-proven patterns as our open-source tools, it’s internal tooling being shaped for enterprise teams facing Flutter’s hardest challenges.

Native-to-Flutter Migration

Automated assessment of existing native iOS/Android codebases, migration pathway generation, and incremental conversion support. Reduces migration timelines from months to weeks by handling the repetitive translation while engineers focus on architecture decisions.

Comprehensive Code Assessments

Deep analysis across architecture, test coverage, accessibility, and performance for entire codebases. Produces actionable reports that prioritize tech debt and identify structural risks before they become production incidents.

Figma-to-Code at Scale

Production-ready Flutter code from Figma designs that respects your design system, component library, and team conventions. Maintains consistency across teams and projects without manual interpretation of design specs.

Let’s Get in Touch

Our Advantage

Why VGV Is Building This

8+ Years of Flutter Expertise, Encoded

Every skill, every pattern, every convention in our AI tools comes from real production apps used by millions. This isn’t generic AI—it’s domain expertise made executable.

Open Source First

The VGV AI Flutter Plugin and VGV Wingspan are fully open source. We believe the best engineering tools should be transparent, extensible, and community-driven.

Structured Workflows, Not Vibe Coding

Our tools enforce disciplined engineering processes—brainstorm, plan, build, review. Every step is auditable, every decision is intentional.

End-to-End Coverage

From individual developer conventions to team-wide workflows to enterprise automation. One platform that scales with your organization.

Better Together

The VGV AI Flutter Plugin enforces conventions at the code level. VGV Wingspan structures the workflow around those conventions. VGV Arcana scales both across the enterprise. Each layer reinforces the others.

Conventions

Flutter Plugin

→

↓

Workflow

VGV Wingspan

→

↓

Enterprise

VGV Arcana

What Engineers Are Saying

Real Results from Top Engineers

“/review cut finding critical issues from hours to minutes. On 3 large PRs alone, that’s 2–3 hours saved. Using /brainstorm, /plan, and /build on Nostr issues delivered solid fixes in hours—work that would have taken triple the time or more without the skills.”

Dominik Simonik

Senior Engineer & Google Developer Expert — VGV & DiVine

Wingspan allows me to focus on the right level of detail at each step. I can explore and clarify the big questions and trade-offs in brainstorm, turn my decisions into a concrete plan, and review the final product for correctness, not analysis errors.

Brian Egan

Senior Engineering Lead — VGV & Google Cloud Next Demo 2026

While teams are struggling to adapt agentic workflows in a predictable way, our AI tooling ecosystem is allowing engineering teams to adopt AI without slop, focusing on the best practices. The combination of VGV AI Flutter Plugin, VGV Wingspan, and VGV Arcana are positively impacting the way we deliver value together with our partners.

Jorge Coca

Head of Engineering — Very Good Ventures

I was shocked the first time I used /brainstorm - it asked incredibly intelligent questions about my codebase. I was implementing a cart page on an ecom flow, and the design had a line for modifications. I hadn't yet considered the modifications and it didn't even exist in the API yet. /brainstorm identified the discrepancy and asked me to pick from three resolutions that were all plausible solutions. It was that day I was a believer in this flow.