Mobile App Builder

Mobile App Builder — Full Mobile App Development Pipeline

An agent team collaborates to generate mobile app UX design, code, API integration, and store deployment all at once.

Execution Mode

Agent Team — 5 agents communicate directly via SendMessage and cross-verify.

Agent Composition

Workflow

Phase 1: Preparation (performed directly by the orchestrator)

1. Extract from user input:
   • App Type: What kind of app (social, commerce, utility, game, etc.)
   • Platform: iOS / Android / Cross-platform
   • Framework Preference (optional): Flutter, React Native, SwiftUI, Jetpack Compose
   • Backend API (optional): If an existing API spec exists
   • Existing Files (optional): Design, code, API specs, etc.
2. Create _workspace/ directory at the project root
3. Organize input and save to _workspace/00_input.md
4. If existing files are provided, copy them to _workspace/ and skip the corresponding Phase
5. Determine execution mode based on request scope (see "Scale-Based Modes" below)

Phase 2: Team Assembly and Execution

Assemble the team and assign tasks. Task dependencies are as follows:

Tasks 2a (app code) and 2b (store metadata) run in parallel.

Inter-team Communication Flow:

• ux-designer completes → delivers screen structure/design tokens to app-developer, screenshot scenarios to store-manager, data fields to api-integrator
• app-developer completes → delivers Repository interfaces to api-integrator, permission list to store-manager
• api-integrator completes → delivers API client code to app-developer
• qa-engineer cross-verifies all deliverables. On 🔴 required fix: requests fix from relevant agent → rework → re-verify (max 2 rounds)

Phase 3: Integration and Final Deliverables

Organize final deliverables based on the QA report:

1. Verify all files in _workspace/
2. Confirm all 🔴 required fixes from the QA report have been addressed
3. Report final summary to the user:
   • UX Design — 01_ux_design.md
   • App Architecture — 02_app_architecture.md
   • App Code — 02_app_code/
   • API Integration — 03_api_integration.md
   • Store Deployment — 04_store_listing.md
   • QA Report — 05_qa_report.md

Scale-Based Modes

Leveraging Existing Files: If the user provides design documents, code, or other existing files, skip the corresponding phases.

Data Transfer Protocol

File naming convention: {order}_{agent}_{deliverable}.{extension}

Error Handling

Test Scenarios

Normal Flow

Prompt: "Build me a to-do management app in Flutter. I need add/delete/complete-check, category classification, and notification features." Expected Result:

• UX Design: 5 screens (list/add/detail/category/settings), navigation structure, design system
• App Code: Flutter + Riverpod + go_router, MVVM structure, per-screen widgets/viewmodels
• API Integration: RESTful CRUD endpoints, auth, local cache
• Store: iOS + Android metadata, ASO keywords
• QA: Full cross-verification, consistency matrix

Existing File Flow

Prompt: "Generate iOS app code from this Figma design" + design file Expected Result:

• Existing design copied to _workspace/01_ux_design.md
• Code Mode: app-developer + api-integrator + qa-engineer deployed
• ux-designer, store-manager skipped

Error Flow

Prompt: "Build me an app quickly, anything is fine" Expected Result:

• App type unclear → UX designer proposes 3 trending app types then proceeds
• Full pipeline mode execution
• QA report notes "assumption-based design due to absent user requirements"

Agent Extension Skills

Extension skills that enhance each agent's domain expertise: