flutter-orchestrator
Enables building Flutter and Dart apps through a multi-agent system that orchestrates specialized agents for UI design, backend development, testing, and deployment.
README
Flutter Multi-Agent Orchestrator (MCP Server & Skill)
Enterprise-grade Multi-Agent System & Model Context Protocol (MCP) Server for building production-ready Flutter & Dart apps across Mobile (iOS/Android), Web (WASM), and Desktop (macOS/Windows/Linux) with Clean Architecture, Riverpod/Bloc, and Hackathon Fast Prototyping.
🎯 Target Platform Pre-Check Rule
Before starting execution, if target platforms are not specified in the prompt or arguments, the Prompt Architect Agent pauses and requests platform clarification:
🎯 Target Platforms Required:
Which platform(s) are you targeting for this Flutter app?
1. 📱 Mobile: iOS & Android (Default)
2. 🌐 Web (Responsive WASM / HTML)
3. 💻 Desktop: macOS, Windows, Linux
4. 🚀 All Platforms (Mobile, Web, Desktop)
🌟 Complete Specialized Agent Roster (23 Tools)
🚀 Core Architecture & Engineering Agents
| Specialized Agent | MCP Tool Name | Description & Key Responsibilities | How to Call |
|---|---|---|---|
| 🧩 Prompt Architect | decompose_flutter_prompt |
Validates target platforms (asks if missing), breaks down prompts into Clean Architecture features (P0/P1/P2), and builds sprint plans. | decompose_flutter_prompt({ prompt: "...", targetPlatforms: ["ios", "android"] }) |
| 🌐 Dependency Researcher | research_flutter_dependencies |
Queries pub.dev, ensures Dart 3.6+ / Flutter 3.27+ compatibility, and generates pubspec.yaml & analysis_options.yaml. |
research_flutter_dependencies({ projectName: "my_app", stateManagement: "riverpod" }) |
| 🎨 UI / UX Specialist | design_flutter_ui |
Generates responsive Material 3 ConsumerWidget screens with domain archetype detection (Fintech, E-Commerce, Fitness, AI, Dashboards), 8-pt spatial grid, and fluid micro-animations. |
design_flutter_ui({ screenName: "DashboardScreen", layoutDescription: "..." }) |
| ⚙️ Backend Specialist | scaffold_flutter_backend |
Scaffolds Clean Architecture repositories, Dio REST / Supabase remote data sources, Drift/Hive caching, and Riverpod providers. | scaffold_flutter_backend({ featureName: "Order", backendProvider: "rest-dio" }) |
| 🩺 Compiler Doctor | diagnose_flutter_errors |
Diagnoses RenderFlex overflows, setState during build, pubspec solving conflicts, CocoaPods build errors, and Gradle namespace mismatches. | diagnose_flutter_errors({ errorMessage: "RenderFlex overflowed by 48px..." }) |
| 🧪 QA & Test Specialist | generate_and_run_flutter_tests |
Generates Unit tests with mocktail, Widget tests with testWidgets, and ProviderContainer tests. |
generate_and_run_flutter_tests({ className: "OrderScreen", testType: "widget" }) |
| 🔍 Code Quality Auditor | audit_flutter_codebase |
Audits Flutter code for unclosed controllers, async calls in build(), and performance anti-patterns. |
audit_flutter_codebase({ code: "...", fileType: "widget" }) |
| 🌐 Cross-Bridge Specialist | generate_flutter_api_bridge |
Generates type-safe Dio REST clients, matching Dart DTO models, Riverpod providers, and WebSocket realtime streams. | generate_flutter_api_bridge({ serviceName: "Order", endpoints: [...] }) |
| 📱 Native Platform Specialist | generate_flutter_platform_config |
Generates iOS Info.plist permission strings, Android AndroidManifest.xml permissions, and Web WASM bootstrap. |
generate_flutter_platform_config({ projectName: "MyApp", permissions: ["camera", "location"] }) |
| 🚀 CI/CD & DevOps Engineer | generate_flutter_cicd_pipeline |
Generates multi-platform GitHub Actions workflows with Split-per-ABI signed APKs, automated gh secret set keystore setup, Windows MSIX test-cert signing, and Fastlane. |
generate_flutter_cicd_pipeline({ projectName: "MyApp", enableSplitPerAbiApk: true }) |
| 📸 Visual Golden Specialist | generate_flutter_golden_tests |
Generates multi-device golden snapshot test suites across phone and tablet viewports in Light & Dark modes. | generate_flutter_golden_tests({ screenName: "DashboardScreen" }) |
| 🤖 On-Device AI Specialist | scaffold_flutter_ai_module |
Scaffolds Google Generative AI (Gemini) SDK streaming integration, Riverpod controllers, and interactive chat views. | scaffold_flutter_ai_module({ featureName: "ChatAssistant", provider: "google-generative-ai" }) |
| 🔐 Security & Crypto Auditor | audit_flutter_security |
Audits Flutter code for secret leaks, scaffolds hardware-backed Keychain/Keystore FlutterSecureStorage, Biometric Auth, and SSL Pinning. |
audit_flutter_security({ codeSnippet: "...", enableBiometrics: true }) |
| 💾 Database Architect | scaffold_flutter_database |
Scaffolds type-safe Drift (SQLite) database schemas, reactive DAOs, and schema version migration routines. | scaffold_flutter_database({ databaseName: "Expense", tables: [...] }) |
| 🌍 Localization (i18n) Specialist | generate_flutter_localization |
Generates l10n.yaml, ARB localization files (en, es, ar), dynamic locale switcher, and Right-to-Left (RTL) language handling. |
generate_flutter_localization({ supportedLocales: ["en", "es", "ar"], stringKeys: {...} }) |
| 📊 Observability & Analytics Specialist | scaffold_flutter_observability |
Scaffolds Sentry/Firebase Crashlytics crash reporting, type-safe custom analytics event taxonomy, and HTTP latency interceptors. | scaffold_flutter_observability({ provider: "sentry", customEventNames: ["checkout_done"] }) |
| 🔗 Deep Linking & Push Specialist | configure_flutter_deep_links |
Configures go_router deep links, Android App Links (assetlinks.json), iOS Universal Links (apple-app-site-association), and FCM notification routing. |
configure_flutter_deep_links({ customScheme: "myapp", domainHost: "app.example.com", routes: [...] }) |
| ♿ Accessibility (a11y) Auditor | audit_flutter_accessibility |
Audits Flutter widgets for WCAG 2.1 AA compliance, screen reader Semantics, minimum 48x48dp touch targets, and scaffolds AccessibleTouchTarget. |
audit_flutter_accessibility({ codeSnippet: "IconButton(icon: Icon(Icons.add), ...)" }) |
| 🚀 Master Orchestrator | orchestrate_flutter_project |
Coordinates the entire multi-agent pipeline from platform verification to working project structure. | orchestrate_flutter_project({ prompt: "...", targetPlatforms: ["android", "ios"] }) |
⚡ Hackathon Fast-Prototyping Agents
| Specialized Agent | MCP Tool Name | Description & Key Responsibilities | How to Call |
|---|---|---|---|
| 🎭 Demo Data & Mock Factory | generate_flutter_mock_factory |
Generates deterministic domain mock data factories, fake entities, and in-memory simulated Riverpod repositories for instant UI prototyping. | generate_flutter_mock_factory({ domainName: "Product", itemCount: 10 }) |
| 🔐 Auth & Onboarding Specialist | scaffold_flutter_auth_flow |
Scaffolds an animated Onboarding carousel (PageView with dots), Material 3 Social login UI (Google/Apple/Email), and token-persisting Riverpod Auth controllers. | scaffold_flutter_auth_flow({ appName: "HackathonApp" }) |
| 📈 Charts & Data Storytelling | generate_flutter_charts |
Scaffolds interactive animated Line charts, Bar graphs, and Donut charts via fl_chart with gradient fills and tooltips. |
generate_flutter_charts({ chartTitle: "Revenue Trends", chartType: "line" }) |
| 🛡️ Hackathon Demo Resilience | scaffold_flutter_offline_resilience |
Guarantees zero crashes during live stage pitches by wrapping network repositories in an automatic fallback interceptor that catches Wi-Fi drops and seamlessly serves cached mock data. | scaffold_flutter_offline_resilience({ enableDemoMockFallback: true }) |
📦 Installation & Setup
git clone https://github.com/peter14l/flutter-agent-orchestrator.git
cd flutter-agent-orchestrator
npm install
npm run build
npm test
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。