@buivietphi/skill-mobile-mt

Master Senior Mobile Engineer — AI skill for Claude Code, Cline, Roo Code, Cursor, Windsurf, Copilot, Codex, Gemini CLI, Kimi, Kilo Code, Kiro, and Antigravity.

Pre-built architecture patterns from 30+ production repos (200k+ GitHub stars) including Ignite, Expensify, Mattermost, Immich, AppFlowy, Now in Android, TCA. Trained from research of top 53k+ star skill repos (everything-claude-code, awesome-cursorrules, planning-with-files, Lovable, Cursor, Manus Agent prompts). Auto-adaptation to your current project.

Install

npx @buivietphi/skill-mobile-mt

Interactive checkbox UI — detects which AI agents are installed, lets you select with arrow keys + space.

Install to specific agents

npx @buivietphi/skill-mobile-mt --claude       # Claude Code
npx @buivietphi/skill-mobile-mt --codex        # Codex
npx @buivietphi/skill-mobile-mt --gemini       # Gemini CLI
npx @buivietphi/skill-mobile-mt --kimi         # Kimi
npx @buivietphi/skill-mobile-mt --antigravity  # Antigravity
npx @buivietphi/skill-mobile-mt --all          # All agents
npx @buivietphi/skill-mobile-mt --path ./dir   # Custom directory

Generate project-level rules

Cline, Roo Code, Cursor, Windsurf, Copilot, Kilo Code, and Kiro read rules from project-level files. Run --init inside your project to generate them:

npx @buivietphi/skill-mobile-mt --init            # Interactive selector
npx @buivietphi/skill-mobile-mt --init cursor     # .cursorrules
npx @buivietphi/skill-mobile-mt --init cline      # .clinerules/mobile-rules.md
npx @buivietphi/skill-mobile-mt --init roocode    # .roo/rules/mobile-rules.md
npx @buivietphi/skill-mobile-mt --init copilot    # .github/copilot-instructions.md
npx @buivietphi/skill-mobile-mt --init windsurf   # .windsurfrules
npx @buivietphi/skill-mobile-mt --init kilocode   # .kilocode/rules/mobile-rules.md
npx @buivietphi/skill-mobile-mt --init kiro       # .kiro/steering/mobile-rules.md
npx @buivietphi/skill-mobile-mt --init all        # All files

What --init does:

1. Auto-detects your project (framework, language, state management, navigation, API client, package manager)
2. Generates rules files pre-filled with your detected stack
3. Includes mobile best practices, security rules, and quality gates
4. Won't overwrite existing files (safe to run multiple times)

Usage

Mode 1: Pre-Built Architecture (default)

@skill-mobile-mt

Uses battle-tested patterns from 30+ production repos:

• React Native — Ignite (19.7k stars), Obytes Template, Expensify/App, Mattermost Mobile, Artsy Eigen
• Flutter — Immich (93.5k stars), AppFlowy (68.2k), Spotube (44.6k), Hiddify, Ente Photos
• iOS — TCA / Point-Free (14.4k), Clean Arch SwiftUI, Modern Clean Arch
• Android — Now in Android (20.7k), Android Showcase, Mihon (18.8k)

Mode 2: Adapt to Your Project (project)

@skill-mobile-mt project

Reads your current project first, then follows your conventions:

• Detects framework, language, package manager, state management, navigation
• Matches your naming, imports, file structure, patterns
• Clones the most similar existing feature when scaffolding new ones
• Never suggests migrations or imposes different architecture

Quick Start Examples

Step 1: Install the skill

# Install globally (one-time setup)
npx @buivietphi/skill-mobile-mt

You'll see an interactive checkbox — use arrow keys to navigate, space to select, Enter to confirm:

  Select agents to install:
  ↑↓ navigate   Space toggle   A select all   Enter confirm   Q cancel

  › ◉ Claude Code   [detected]
    ◯ Cline          [detected]
    ◯ Roo Code       [not found]
    ◯ Cursor         [detected]
    ◯ Windsurf       [not found]
    ◯ Copilot        [not found]
    ◯ Codex          [not found]
    ◉ Gemini CLI     [detected]
    ◯ Kimi           [not found]
    ◯ Kilo Code      [not found]
    ◯ Kiro           [not found]
    ◯ Antigravity    [not found]

Install output:

  ✓ skill-mobile-mt/    → Claude Code  (~/.claude/skills/skill-mobile-mt)
  ✓ humanizer-mobile/   → Claude Code  (~/.claude/skills/humanizer-mobile)

Both are installed automatically. Use @humanizer-mobile for app store copy, release notes, and UX text.

Step 2: Generate project rules

# cd into your mobile project first
cd ~/projects/my-app

# Generate rules for your agents
npx @buivietphi/skill-mobile-mt --init all

Output:

  📁 Project directory: /Users/you/projects/my-app
  🔍 Detected: Expo (React Native) (TypeScript)

  ✓ .cursorrules → Cursor (auto-detected: Expo (React Native))
  ✓ .github/copilot-instructions.md → GitHub Copilot
  ✓ .clinerules/mobile-rules.md → Cline
  ✓ .roo/rules/mobile-rules.md → Roo Code
  ✓ .kilocode/rules/mobile-rules.md → Kilo Code
  ✓ .kiro/steering/mobile-rules.md → Kiro
  ✓ .windsurfrules → Windsurf

  ✅ Done! → 7 file(s) generated

Step 3: Use in your AI agent

Claude Code

# In Claude Code terminal, type:
@skill-mobile-mt

# Then ask anything:
> Create auth feature with login and register screens

# Or with project mode (reads YOUR code first):
@skill-mobile-mt project
> Add cart feature following the same pattern as ProductList

Cline / Roo Code / Kilo Code

After running --init cline (or roocode, kilocode), rules are auto-loaded from the project directory. Just open your project and ask:

> Create a login screen with email/password
> Fix the crash in ProductDetail when images is null
> Review this PR for mobile best practices

Cursor

After running --init cursor, Cursor auto-loads .cursorrules every session. Just open your project and code normally.

Kiro (AWS)

After running --init kiro, Kiro loads .kiro/steering/mobile-rules.md as steering context. The generated file uses inclusion: always frontmatter.

GitHub Copilot

After running --init copilot, Copilot reads .github/copilot-instructions.md as workspace context.

Windsurf

After running --init windsurf, Windsurf auto-loads .windsurfrules. Just code normally.

Gemini CLI

# In Gemini CLI:
@skill-mobile-mt
> Setup a new Flutter project with Riverpod + clean architecture

Codex / Kimi / Antigravity

Same as Claude Code — the skill is loaded from the skills directory.

Example Prompts (all agents)

| What you want | What to type | |--------------|-------------| | New feature | Create auth feature with login, register, forgot password | | New screen | Add settings screen with profile, notifications, theme toggle | | Fix bug | Fix crash when product.images is undefined | | Code review | Review src/features/cart/ for mobile best practices | | Performance | Optimize ProductList — it's janky when scrolling 100+ items | | Architecture | Setup project structure for a new Expo app with Zustand + TanStack Query | | Release | Prepare the app for App Store submission | | Security | Audit the auth flow for security vulnerabilities |

Auto-Detect

The skill automatically detects before any action:

| What | How | |------|-----| | Framework | pubspec.yaml → Flutter, package.json with react-native → RN, *.xcodeproj → iOS, build.gradle → Android | | Language | .dart → Dart, .tsx/.ts → TypeScript, .swift → Swift, .kt → Kotlin | | Package Manager | yarn.lock → yarn, pnpm-lock.yaml → pnpm, package-lock.json → npm, pubspec.lock → flutter pub | | State Management | Redux / MobX / Zustand / Riverpod / BLoC / StateFlow / Combine | | Navigation | React Navigation / Expo Router / GoRouter / NavigationStack |

Smart Loading

Only loads the relevant platform docs — saves ~45% context tokens vs loading everything.

Flutter project?
  → Load: flutter/flutter.md + ios/ + android/ + shared/
  → Skip: react-native/

React Native project?
  → Load: react-native/react-native.md + ios/ + android/ + shared/
  → Skip: flutter/

iOS only?
  → Load: ios/ios-native.md + shared/
  → Skip: flutter/ + react-native/ + android/

Context Cost

| Scenario | Tokens | % of 128K | % of 200K | |----------|-------:|----------:|----------:| | SKILL.md only | ~25,150 | 19.6% | 12.6% | | Smart load (recommended) | ~53,350 – 57,350 | 42% – 45% | 27% – 29% | | Cross-platform (RN+Flutter+iOS+Android) | ~95,350 | 74.5% | 47.7% | | All files loaded | ~166,380 | 130.0% | 83.2% | | | | exceeds 128K | fits 200K |

Per-file token breakdown

| File | Tokens | |------|-------:| | SKILL.md | ~25,150 | | AGENTS.md | ~6,440 | | react-native/react-native.md | ~5,840 | | flutter/flutter.md | ~2,350 | | ios/ios-native.md | ~1,660 | | android/android-native.md | ~4,370 | | shared/code-review.md | ~11,600 | | shared/bug-detection.md | ~7,290 | | shared/debugging-intelligence.md | ~8,730 | | shared/prompt-engineering.md | ~7,440 | | shared/architecture-intelligence.md | ~3,860 | | shared/common-pitfalls.md | ~1,330 | | shared/error-recovery.md | ~2,780 | | shared/document-analysis.md | ~1,520 | | shared/release-checklist.md | ~670 | | shared/anti-patterns.md | ~2,910 | | shared/performance-prediction.md | ~1,310 | | shared/platform-excellence.md | ~1,890 | | shared/version-management.md | ~3,660 | | shared/observability.md | ~5,750 | | shared/offline-first.md | ~2,930 | | shared/testing-strategy.md | ~2,500 | | shared/ci-cd.md | ~2,830 | | shared/ai-dlc-workflow.md | ~1,950 | | shared/ui-ux-mobile.md | ~6,910 | | shared/storage-patterns.md | ~2,760 | | shared/i18n-localization.md | ~3,140 | | shared/intent-analysis.md | ~6,250 | | shared/code-generation-templates.md | ~5,370 | | shared/spec-to-code.md | ~2,640 | | shared/navigation-patterns.md | ~2,900 | | shared/complex-ui-patterns.md | ~3,810 | | shared/data-flow-patterns.md | ~3,250 | | shared/error-handling.md | ~2,960 | | shared/testing-patterns.md | ~3,410 | | shared/claude-md-template.md | ~1,050 | | shared/agent-rules-template.md | ~2,870 | | humanizer/humanizer-mobile.md | ~2,630 | | Total | ~166,380 |

Installed Structure

~/.claude/skills/               (or ~/.cline/skills/, ~/.roo/skills/, ~/.gemini/skills/, etc.)
└── skill-mobile-mt/
    ├── SKILL.md                       Entry point + auto-detect + quality gates
    ├── AGENTS.md                      Multi-agent compatibility
    ├── react-native/
    │   └── react-native.md            React Native + Expo patterns
    ├── flutter/
    │   └── flutter.md                 Flutter + Dart patterns
    ├── ios/
    │   └── ios-native.md              iOS Swift patterns
    ├── android/
    │   └── android-native.md          Android Kotlin + Java patterns
    └── shared/
        ├── code-review.md             Senior review checklist
        ├── bug-detection.md           Auto bug scanner (4 modes + git-aware)
        ├── debugging-intelligence.md  30+ error patterns + root cause tracing
        ├── prompt-engineering.md      Auto-think + prompt templates
        ├── architecture-intelligence.md  Patterns from 30+ production repos
        ├── common-pitfalls.md         Problem → Symptoms → Solution
        ├── error-recovery.md          16 build/runtime error fixes
        ├── document-analysis.md       Parse docs/images → code
        ├── anti-patterns.md           PII, cardinality, payload detection
        ├── performance-prediction.md  Frame budget, FPS prediction
        ├── platform-excellence.md     iOS 18+ vs Android 15+ guidelines + HIG
        ├── version-management.md      SDK compatibility matrix
        ├── observability.md           Sessions as 4th pillar
        ├── release-checklist.md       Pre-release verification
        ├── offline-first.md           Local-first + sync patterns
        ├── storage-patterns.md        Storage selection matrix (AsyncStorage/MMKV/SQLite/etc.)
        ├── i18n-localization.md       Internationalization + RTL + locale patterns
        ├── testing-strategy.md        Detox + Maestro + XCUITest + Espresso E2E
        ├── ci-cd.md                   GitHub Actions CI/CD templates
        ├── ai-dlc-workflow.md         AI-DLC structured dev workflow
        ├── ui-ux-mobile.md            Design system + screen templates + touch + navigation
        ├── claude-md-template.md      CLAUDE.md template for projects
        └── agent-rules-template.md    Rules templates for all agents

Project-level files (generated by --init)

your-project/
├── .cursorrules                       Cursor rules (auto-detected stack)
├── .windsurfrules                     Windsurf rules (auto-detected stack)
├── .clinerules/
│   └── mobile-rules.md               Cline rules (auto-detected stack)
├── .roo/
│   └── rules/
│       └── mobile-rules.md           Roo Code rules (auto-detected stack)
├── .kilocode/
│   └── rules/
│       └── mobile-rules.md           Kilo Code rules (auto-detected stack)
├── .kiro/
│   └── steering/
│       └── mobile-rules.md           Kiro steering (auto-detected stack)
└── .github/
    └── copilot-instructions.md        Copilot rules (auto-detected stack)

What's Included

Clean Architecture & SOLID

• Dependencies flow inward (UI → Domain → Data)
• Single responsibility per file (max 300 lines)
• Dependency injection, no hardcoded singletons

Code Review Protocol (12-category, PR-level)

• PR-Level Review: size check, single responsibility, test accompaniment, commit hygiene — BEFORE code review
• 12 categories: architecture, correctness, boundary conditions, test quality, readability, performance, security (expanded), accessibility (WCAG 2.1), breaking changes, platform, documentation, i18n
• Severity levels: Critical / High / Medium / Low with structured output format
• Auto-fail patterns: 10 code + 4 PR-level + 7 AI-specific grounding checks
• Review output: PR-level summary → findings by severity → verdict (APPROVE / CHANGES REQUESTED)
• Boundary conditions: null, empty, off-by-one, numeric limits, unicode, timezone — dedicated dimension
• Accessibility: touch targets (44pt/48dp), contrast ratios, screen reader, font scaling, focus management
• Breaking change detection: public API, deep links, DB schema, push payload, analytics events
• Security depth: certificate pinning, JWT lifecycle, bridge security, biometric auth, dependency audit
• Grounded review (anti-false-positive): verify every finding before flagging — check actual installed APIs, don't flag from memory, confidence levels (high/medium/low)
• Practical usage review: 25 production crash patterns — 5 cross-platform + 5 React Native + 5 Flutter + 5 iOS Swift + 5 Android Kotlin
• Library-specific traps: React Native, Flutter, iOS Swift, Android Kotlin — real gotchas per platform
• Cross-platform examples: before/after code for all 4 platforms (not just RN)

Bug Detection Scanner

• 4 debug modes: Error Analysis, Fix Bug, Investigate, Diagnostic Scan
• Git-aware debugging: checks recent commits before deep-diving (auto-skips if no git)
• 4-phase fix protocol: Root Cause → Pattern Analysis → Single Hypothesis → Defense in Depth
• 45+ error pattern database with exact search strategies — RN (7) + Flutter (6) + iOS Swift (5) + Android Kotlin (6) + Network (4) + State (3) + Navigation (2) + Build (4) + Platform (3)
• Anti-rationalization table: catches self-deception during debugging

Decision Matrix Protocol

• Structured comparison format when multiple valid approaches exist
• Estimation protocol: classify effort XS/S/M/L/XL with risk assessment
• Migration/upgrade decision protocol: breaking changes analysis + impact scan

Codebase Scan Strategy

• 3 scan levels: Quick (~5 reads), Standard (~15 reads), Deep (~30+ reads)
• Monorepo strategy: focus target package, scan shared deps
• Multi-module strategy: JS/TS vs Native vs Cross-layer task detection

Grounding Protocol (Anti-Hallucination)

• Source hierarchy: project code > skill files > official docs > production repos > AI knowledge
• Mandatory rules: Read before answer, verify APIs exist, cite sources, say "I don't know" when unsure
• No phantom packages: verify package exists in package.json before suggesting
• Version-specific: check actual SDK version, never assume latest
• Grounded bug fix: read file → find exact line → verify types → cite source
• Anti-hallucination checklist: runs before every AI response

Docs-First Protocol (Always Use Latest)

• WebSearch before coding: when setting up any library, ALWAYS search official docs first
• Never rely on AI memory: docs change, APIs change, syntax changes between versions
• Package installation protocol: verify exists → check compatibility → search install guide → test both platforms
• Common AI traps: React Navigation v5 vs v7, Firebase v8 vs v9 modular, Swift async/await availability
• Version matrix + live verification: snapshot matrix in skill + mandatory WebSearch for current data

Security Protocol

• 7-category security scan on every feature: secrets, token storage, input validation, network, data protection, auth flow, platform-specific
• 6 absolute rules: never store tokens in plain storage, never hardcode secrets, never log sensitive data, never trust deep links, never disable SSL, never commit .env
• Platform-specific checks: iOS ATS + privacy manifest, Android cleartext + ProGuard + exported components
• Never skip security: even for prototypes, internal apps, or "quick fixes"

Advanced AI Patterns (from top 53k+ star repos + Lovable, Cursor, Manus prompts)

• 7 Cardinal Rules: Inviolable rules at the top of SKILL.md (read before write, verify before done, clone before create, cite source, 4 states always, platform parity, ask after 3 fails)
• Self-Critique Loop: Generate → Review own code → Refine → Verify (catch bugs before "done")
• Context Staleness Rule: Re-read files after 5+ messages (from Cursor's pattern)
• Parallel Execution: Default to parallel tool calls unless sequential dependency
• Leverage Pyramid: Research → Planning → Implementation (review plans, not just code)
• Session State Tracking: Structured progress tracking for long multi-file tasks
• Verification-First: Include test criteria BEFORE implementation (Anthropic's #1 recommendation)
• Assumption-Driven Progress: State assumptions → proceed → let user correct (don't ask for every decision)
• Negative Space Pattern: Explicit NEVER rules prevent drift better than positive instructions
• Brevity Default: 1-line status updates, detail only when asked (from Lovable, Claude Code)
• Error Recovery Protocol: Max 3 attempts with escalation, then ask user
• Build & Deploy Gates: Platform-specific pre-completion checklists
• Mobile Anti-Patterns: Common mistakes from AI tools (and how to avoid them)

Intelligent Prompt Engineering

• Trained from 7 major AI tools' system prompts (Lovable, Cursor, Manus, Windsurf, Kiro, Replit, Claude Code)
• XML tag structure for clarity (<task>, <context>, <instructions>, <constraints>)
• Enhanced auto-think templates with pre-action validation
• Progressive context loading (5 levels: overview → pattern → files → docs → expert)
• Just-in-time file reading (grep first, read only needed)
• Reference pattern system (clone existing instead of reading docs)
• Multi-AI format support (Claude, Gemini, Kimi, Cursor, Copilot, Windsurf)
• Advanced patterns: verification-first, investigate-before-answer, batched operations, negative space

Document Analysis

• Parse images (mockups, wireframes), PDFs (requirements), DOCX, XML, JSON
• UI Analysis: extract layout, components, styling, states from screenshots
• Requirements Extraction: user stories, screen flows, data models, business rules
• Document → Code pipeline: read → extract → map features → scaffold

Architecture Intelligence (from 30+ production repos)

• Architecture Intelligence (architecture-intelligence.md): Cross-platform patterns from Ignite, Immich, AppFlowy, TCA, Now in Android and more. Includes: dual state management, feature-based modules, bootstrap/startup pattern, functional error handling, architecture validation (Konsist), production folder structure templates, and decision matrices for state/nav/testing.

Production Patterns (from Senaiverse, Mhuxain, VoltAgent, Nexus)

• Anti-Pattern Detection (anti-patterns.md): Detect PII leaks (CRITICAL), high cardinality tags, unbounded payloads, unstructured logs, sync telemetry on main thread — with auto-fix suggestions
• Performance Prediction (performance-prediction.md): Calculate frame budget, FlatList bridge calls, and memory usage BEFORE writing code. Example: 50 items × 3 bridge calls × 0.3ms = 45ms/frame → 22 FPS ❌ JANK
• Platform Excellence (platform-excellence.md): iOS 18+ vs Android 15+ native UX standards — navigation patterns, typography, haptic feedback types, permission timing, ratings prompt flow, Live Activities/Dynamic Island, performance targets (cold start < 1s iOS, < 1.5s Android)
• Version Management (version-management.md): Full SDK compatibility matrix for RN 0.73-0.76, Expo 50-52, Flutter 3.22-3.27, iOS 16-18, Android 13-15. Check SDK compat BEFORE npm install. Release-mode testing protocol.
• Observability (observability.md): Sessions as the 4th pillar (Metrics + Logs + Traces + Sessions). Session lifecycle, enrichment API, unified instrumentation stack, correlation queries. Every event carries session_id for full user journey reconstruction.

UI/UX Mobile Design

• Screen templates with ASCII wireframes: Login, Home/Feed, Detail, Profile/Settings, Onboarding
• Design tokens system (colors, spacing, radius, shadows, typography) for RN and Flutter
• Component patterns: bottom sheet, empty state, loading skeleton, toast/snackbar
• Touch targets (44pt iOS / 48dp Android), dark mode implementation, animation guidelines
• Accessibility checklist (screen reader labels, contrast ratios, font scaling)
• Device sizing reference table (iPhone SE → Pro Max, Android small → large, iPad)

AI-DLC Workflow (AI-Driven Development Lifecycle)

• Based on AWS AI-DLC methodology — structured development for complex features
• Elaboration: decompose task into Intent + Units before coding
• Construction loop: 4 Hats per Unit (Architecture → Builder → Security → Reviewer)
• Backpressure gates: compiler, lint, security, review must pass before next Unit
• Operating modes: HITL (human approves each Unit), OHOTL, AHOTL
• Auto-activates for multi-screen features (3+ units); skips for bug fixes

Error Recovery

• 16 common build/runtime errors with concrete fixes
• Platform-specific solutions (RN CLI, Expo, Flutter, iOS, Android)
• General recovery protocol for systematic debugging

Release & Review

• Pre-release checklist for App Store and Play Store submissions
• Common pitfalls with Problem → Symptoms → Solution format
• Security audit patterns

Security Non-Negotiables

• No hardcoded secrets
• Secure token storage (Keychain / EncryptedSharedPreferences / SecureStore)
• Deep link validation, HTTPS only

Supported Agents

| Agent | How it works | Setup command | |-------|-------------|---------------| | Claude Code | Reads from ~/.claude/skills/ | npx skill-mobile-mt --claude | | Cline | Reads .clinerules/ from project root | npx skill-mobile-mt --init cline | | Roo Code | Reads .roo/rules/ from project root | npx skill-mobile-mt --init roocode | | Cursor | Reads .cursorrules from project root | npx skill-mobile-mt --init cursor | | Windsurf | Reads .windsurfrules from project root | npx skill-mobile-mt --init windsurf | | Copilot | Reads .github/copilot-instructions.md | npx skill-mobile-mt --init copilot | | Codex | Reads from ~/.codex/skills/ | npx skill-mobile-mt --codex | | Gemini CLI | Reads from ~/.gemini/skills/ | npx skill-mobile-mt --gemini | | Kimi | Reads from ~/.kimi/skills/ | npx skill-mobile-mt --kimi | | Kilo Code | Reads .kilocode/rules/ from project root | npx skill-mobile-mt --init kilocode | | Kiro | Reads .kiro/steering/ from project root | npx skill-mobile-mt --init kiro | | Antigravity | Reads from ~/.agents/skills/ | npx skill-mobile-mt --antigravity |

Companion Skill

humanizer-mobile is auto-installed alongside the main skill. It humanizes AI-generated mobile copy — app store descriptions, release notes, error messages, push notifications, paywall text.

@humanizer-mobile
> "Humanize this app store description"

License

MIT — by buivietphi