mcpField
T
tauri-app-dev

tauri-app-dev

github.com/xiaolaiDevelopment87876 installsUpdated today
Repository

About tauri-app-dev

A development skill for building Tauri 2.0 desktop apps with web frontends and Rust backends. It covers IPC, security permissions, plugins, packaging, and testing practices.

Key features

  1. 01Structure web frontends and Rust backends
  2. 02Handle IPC with commands and events
  3. 03Configure capabilities and API permissions
  4. 04Use plugins such as filesystem access
  5. 05Build, bundle, and test apps

Use cases

  1. 01Call Rust commands from a web frontend
  2. 02Set file-access permissions for windows
  3. 03Implement file dialogs and read/write operations
  4. 04Test a running Tauri app

MCP support

It describes using Tauri MCP tools for end-to-end testing of a running app, and advises against using chrome-devtools MCP for this purpose.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/xiaolai/vmark --skill tauri-app-dev

After installing, type/tauri-app-devin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

references/
bundling.md
commands-and-ipc.md
patterns.md
plugins.md
security.md

Details

From the project README

Tauri 2.0 App Development

Tauri is a framework for building small, fast, secure desktop apps using web frontends and Rust backends.

Architecture Overview

┌─────────────────────────────────────────┐
│           Frontend (Webview)            │
│     HTML/CSS/JS • React/Vue/Svelte      │
└────────────────┬────────────────────────┘
                 │ IPC (invoke/events)
┌────────────────▼────────────────────────┐
│           Tauri Core (Rust)             │
│  Commands • State • Plugins • Events    │
└────────────────┬────────────────────────┘
                 │ TAO (windows) + WRY (webview)
┌────────────────▼────────────────────────┐
│          Operating System               │
│   macOS • Windows • Linux • Mobile      │
└─────────────────────────────────────────┘

Project Structure

my-app/
├── src/                    # Frontend source
├── src-tauri/
│   ├── Cargo.toml          # Rust dependencies
│   ├── tauri.conf.json     # Tauri configuration
│   ├── capabilities/       # Security permissions (v2)
│   │   └── default.json
│   ├── src/
│   │   ├── main.rs         # Desktop entry point
│   │   └── lib.rs          # Main app logic + mobile entry
│   └── icons/
└── package.json

Commands (Frontend → Rust)

Define commands in Rust with #[tauri::command]:

// src-tauri/src/lib.rs
#[tauri::command]
fn greet(name: String) -> String {
    format!("Hello, {}!", name)
}

#[tauri::command]
async fn read_file(path: String) -> Result<String, String> {
    std::fs::read_to_string(&path).map_err(|e| e.to_string())
}

pub fn run() {
    tauri::Builder::default()
        .invoke_handler(tauri::generate_handler![greet, read_file])
        .run(tauri::generate_context!())
        .expect("error running app");
}

Call from frontend (direct):

import { invoke } from '@tauri-apps/api/core';

const greeting = await invoke<string>('greet', { name: 'World' });
const content = await invoke<string>('read_file', { path: '/tmp/test.txt' });

Project convention: Wrap invoke() with TanStack Query for caching and state management:

import { useQuery, useMutation } from '@tanstack/react-query';
import { invoke } from '@tauri-apps/api/core';

// Query (read operations)
const { data: content } = useQuery({
  queryKey: ['file', path],
  queryFn: () => invoke<string>('read_file', { path }),
});

// Mutation (write operations)
const { mutate: saveFile } = useMutation({
  mutationFn: (content: string) => invoke('write_file', { path, content }),
});

Key rules:

  • Arguments must implement serde::Deserialize
  • Return types must implement serde::Serialize
  • Use Result<T, E> for fallible operations
  • Async commands run on thread pool (non-blocking)
  • Snake_case in Rust → camelCase in JS arguments

State Management

Share state across commands:

use std::sync::Mutex;
use tauri::State;

struct AppState {
    counter: Mutex<i32>,
    db: Mutex<Option<Database>>,
}

#[tauri::command]
fn increment(state: State<'_, AppState>) -> i32 {
    let mut counter = state.counter.lock().unwrap();
    *counter += 1;
    *counter
}

pub fn run() {
    tauri::Builder::default()
        .manage(AppState {
            counter: Mutex::new(0),
            db: Mutex::new(None),
        })
        .invoke_handler(tauri::generate_handler![increment])
        .run(tauri::generate_context!())
        .expect("error running app");
}

Access via AppHandle (for background threads):

use tauri::Manager;

#[tauri::command]
async fn background_task(app: tauri::AppHandle) {
    let state = app.state::<AppState>();
    // use state...
}

Events (Rust → Frontend)

Emit events from Rust:

use tauri::Emitter;

#[tauri::command]
fn start_process(app: tauri::AppHandle) {
    std::thread::spawn(move || {
        for i in 0..100 {
            app.emit("progress", i).unwrap();
            std::thread::sleep(std::time::Duration::from_millis(50));
        }
        app.emit("complete", "Done!").unwrap();
    });
}

Listen in frontend:

import { listen } from '@tauri-apps/api/event';

const unlisten = await listen<number>('progress', (event) => {
    console.log(`Progress: ${event.payload}%`);
});

// Clean up when done
unlisten();

Essential Plugins

Install plugins: cargo add <plugin> in src-tauri, pnpm add <package> in frontend.

PluginCargo CrateNPM PackagePurpose
File Systemtauri-plugin-fs@tauri-apps/plugin-fsRead/write files
Dialogtauri-plugin-dialog@tauri-apps/plugin-dialogOpen/save dialogs
Clipboardtauri-plugin-clipboard-manager@tauri-apps/plugin-clipboard-managerCopy/paste
Shelltauri-plugin-shell@tauri-apps/plugin-shellRun external commands
Storetauri-plugin-store@tauri-apps/plugin-storeKey-value persistence
Updatertauri-plugin-updater@tauri-apps/plugin-updaterAuto-updates

Register in Rust:

pub fn run() {
    tauri::Builder::default()
        .plugin(tauri_plugin_fs::init())
        .plugin(tauri_plugin_dialog::init())
        .plugin(tauri_plugin_clipboard_manager::init())
        .run(tauri::generate_context!())
        .expect("error running app");
}

Security: Capabilities & Permissions

Tauri 2.0 uses capabilities (in src-tauri/capabilities/) to control what APIs each window can access.

src-tauri/capabilities/default.json:

{
  "$schema": "../gen/schemas/desktop-schema.json",
  "identifier": "main-capability",
  "windows": ["main"],
  "permissions": [
    "core:default",
    "fs:default",
    "fs:allow-read-text-file",
    "dialog:default",
    {
      "identifier": "fs:scope",
      "allow": [{ "path": "$APPDATA/**" }, { "path": "$DOCUMENT/**" }]
    }
  ]
}

Scope variables: $APPDATA, $APPCONFIG, $DOCUMENT, $DOWNLOAD, $HOME, $TEMP, etc.

File Operations (Editor Pattern)

import { open, save } from '@tauri-apps/plugin-dialog';
import { readTextFile, writeTextFile } from '@tauri-apps/plugin-fs';

// Open file dialog
const path = await open({
    filters: [{ name: 'Markdown', extensions: ['md'] }],
    multiple: false,
});

if (path) {
    const content = await readTextFile(path);
    // Edit content...
    await writeTextFile(path, modifiedContent);
}

// Save as dialog
const savePath = await save({
    filters: [{ name: 'Markdown', extensions: ['md'] }],
    defaultPath: 'untitled.md',
});

if (savePath) {
    await writeTextFile(savePath, content);
}

Window Management

Create windows at runtime:

use tauri::{WebviewUrl, WebviewWindowBuilder};

#[tauri::command]
async fn open_settings(app: tauri::AppHandle) -> Result<(), String> {
    WebviewWindowBuilder::new(&app, "settings", WebviewUrl::App("settings.html".into()))
        .title("Settings")
        .inner_size(600.0, 400.0)
        .build()
        .map_err(|e| e.to_string())?;
    Ok(())
}

Configure in tauri.conf.json:

{
  "app": {
    "windows": [
      {
        "label": "main",
        "title": "My App",
        "width": 1200,
        "height": 800,
        "decorations": true,
        "resizable": true
      }
    ]
  }
}

Custom Titlebar

Set decorations: false in config, then:

<div data-tauri-drag-region class="titlebar">
    <span>My App</span>
    <button id="minimize">−</button>
    <button id="maximize">□</button>
    <button id="close">×</button>
</div>
import { getCurrentWindow } from '@tauri-apps/api/window';

const appWindow = getCurrentWindow();
document.getElementById('minimize')?.addEventListener('click', () => appWindow.minimize());
document.getElementById('maximize')?.addEventListener('click', () => appWindow.toggleMaximize());
document.getElementById('close')?.addEventListener('click', () => appWindow.close());

Building & Distribution

# Development
pnpm tauri dev

# Production build
pnpm tauri build

# Build specific targets
pnpm tauri build --target universal-apple-darwin  # macOS universal
pnpm tauri build --bundles deb,appimage           # Linux only
pnpm tauri build --bundles nsis                   # Windows NSIS

Output locations:

  • macOS: target/release/bundle/macos/*.app, *.dmg
  • Windows: target/release/bundle/nsis/*-setup.exe, msi/*.msi
  • Linux: target/release/bundle/deb/*.deb, appimage/*.AppImage

Quick Reference

TaskResource
Commands, IPC, channelsSee references/commands-and-ipc.md
Plugin usage & developmentSee references/plugins.md
Security configurationSee references/security.md
Bundling & distributionSee references/bundling.md
Common app patternsSee references/patterns.md

Test-Driven Development (TDD)

CRITICAL: Always follow TDD - write tests BEFORE implementation.

TDD Workflow
1. RED    → Write failing test first
2. GREEN  → Write minimal code to pass
3. REFACTOR → Clean up, keep tests green
Testing Stack
LayerToolPurpose
Rust Unitcargo testTest commands, business logic
React UnitVitestTest components, hooks, stores
IntegrationVitest + MSWTest frontend with mocked IPC
E2ETauri MCPTest running app (NOT Chrome DevTools)
E2E Testing with Tauri MCP

IMPORTANT: Always use tauri_* MCP tools for testing the running app. Do NOT use chrome-devtools MCP - it's for browser pages only.

// Tauri MCP workflow for E2E tests:

// 1. Start session (connect to running Tauri app)
tauri_driver_session({ action: 'start', port: 9223 })

// 2. Take snapshot (get DOM state)
tauri_webview_screenshot()
tauri_webview_find_element({ selector: '.editor-content' })

// 3. Interact with app
tauri_webview_interact({ action: 'click', selector: '#save-button' })
tauri_webview_keyboard({ action: 'type', selector: 'input', text: 'hello' })

// 4. Wait for results
tauri_webview_wait_for({ type: 'selector', value: '.success-toast' })

// 5. Verify IPC calls
tauri_ipc_monitor({ action: 'start' })
tauri_ipc_get_captured({ filter: 'save_file' })

// 6. Check backend state
tauri_ipc_execute_command({ command: 'get_app_state' })

// 7. Read logs for debugging
tauri_read_logs({ source: 'console', lines: 50 })
Rust Unit Tests
// src-tauri/src/lib.rs

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_greet() {
        let result = greet("World".to_string());
        assert_eq!(result, "Hello, World!");
    }

    #[test]
    fn test_parse_markdown() {
        let input = "# Hello";
        let result = parse_markdown(input);
        assert!(result.is_ok());
        assert_eq!(result.unwrap().title, "Hello");
    }

    #[tokio::test]
    async fn test_async_command() {
        let result = read_file("/tmp/test.txt".to_string()).await;
        // Test with temp files or mocks
    }
}

Run: cd src-tauri && cargo test

React Component Tests (Vitest)
// src/components/Editor.test.tsx
import { describe, it, expect, vi } from 'vitest'
import { render, screen } from '@testing-library/react'
import userEvent from '@testing-library/user-event'
import { Editor } from './Editor'

// Mock Tauri invoke
vi.mock('@tauri-apps/api/core', () => ({
  invoke: vi.fn()
}))

describe('Editor', () => {
  it('should render editor content', () => {
    render(<Editor initialValue="# Hello" />)
    expect(screen.getByText('Hello')).toBeInTheDocument()
  })

  it('should call save on Ctrl+S', async () => {
    const { invoke } = await import('@tauri-apps/api/core')
    render(<Editor initialValue="test" />)

    await userEvent.keyboard('{Control>}s{/Control}')

    expect(invoke).toHaveBeenCalledWith('save_file', expect.any(Object))
  })
})
Zustand Store Tests
// src/stores/editorStore.test.ts
import { describe, it, expect, beforeEach } from 'vitest'
import { useEditorStore } from './editorStore'

describe('editorStore', () => {
  beforeEach(() => {
    // Reset store before each test
    useEditorStore.setState({
      content: '',
      isDirty: false,
      filePath: null
    })
  })

  it('should update content and mark dirty', () => {
    const { setContent } = useEditorStore.getState()

    setContent('new content')

    const state = useEditorStore.getState()
    expect(state.content).toBe('new content')
    expect(state.isDirty).toBe(true)
  })

  it('should clear dirty flag after save', () => {
    useEditorStore.setState({ isDirty: true })
    const { markSaved } = useEditorStore.getState()

    markSaved()

    expect(useEditorStore.getState().isDirty).toBe(false)
  })
})
Integration Tests with Mocked IPC
// src/features/file/useFileOperations.test.ts
import { describe, it, expect, vi, beforeEach } from 'vitest'
import { renderHook, waitFor } from '@testing-library/react'
import { QueryClient, QueryClientProvider } from '@tanstack/react-query'
import { useFileOperations } from './useFileOperations'

vi.mock('@tauri-apps/api/core', () => ({
  invoke: vi.fn()
}))

vi.mock('@tauri-apps/plugin-dialog', () => ({
  open: vi.fn(),
  save: vi.fn()
}))

describe('useFileOperations', () => {
  let queryClient: QueryClient

  beforeEach(() => {
    queryClient = new QueryClient({
      defaultOptions: { queries: { retry: false } }
    })
    vi.clearAllMocks()
  })

  it('should open file and load content', async () => {
    const { invoke } = await import('@tauri-apps/api/core')
    const { open } = await import('@tauri-apps/plugin-dialog')

    vi.mocked(open).mockResolvedValue('/path/to/file.md')
    vi.mocked(invoke).mockResolvedValue('# File Content')

    const { result } = renderHook(() => useFileOperations(), {
      wrapper: ({ children }) => (
        <QueryClientProvider client={queryClient}>
          {children}
        </QueryClientProvider>
      )
    })

    await result.current.openFile()

    await waitFor(() => {
      expect(invoke).toHaveBeenCalledWith('read_file', { path: '/path/to/file.md' })
    })
  })
})
TDD Example: Adding a New Feature
// Step 1: RED - Write failing test first
// src/features/wordcount/useWordCount.test.ts
describe('useWordCount', () => {
  it('should count words in content', () => {
    const { result } = renderHook(() => useWordCount('hello world'))
    expect(result.current.words).toBe(2)
  })

  it('should handle empty content', () => {
    const { result } = renderHook(() => useWordCount(''))
    expect(result.current.words).toBe(0)
  })

  it('should count characters', () => {
    const { result } = renderHook(() => useWordCount('hello'))
    expect(result.current.characters).toBe(5)
  })
})

// Step 2: GREEN - Minimal implementation
// src/features/wordcount/useWordCount.ts
export function useWordCount(content: string) {
  return {
    words: content.trim() ? content.trim().split(/\s+/).length : 0,
    characters: content.length
  }
}

// Step 3: REFACTOR - Add memoization, types, etc.
export function useWordCount(content: string): WordCountResult {
  return useMemo(() => ({
    words: content.trim() ? content.trim().split(/\s+/).length : 0,
    characters: content.length,
    charactersNoSpaces: content.replace(/\s/g, '').length
  }), [content])
}
Running Tests
# All tests
pnpm test

# Watch mode
pnpm test:watch

# Coverage
pnpm test:coverage

# Rust tests only
cd src-tauri && cargo test

# Type check + lint + test
pnpm check:all

Debugging Tips

  • DevTools: Right-click → Inspect, or Cmd+Option+I (macOS) / Ctrl+Shift+I (Windows/Linux)
  • Rust logs: Use log crate + tauri-plugin-log or println! (visible in terminal)
  • Check capabilities: "Not allowed" errors mean missing permissions in capabilities
  • IPC errors: Ensure argument names match (snake_case Rust → camelCase JS)
  • E2E debugging: Use tauri_read_logs({ source: 'console' }) to see webview console

Related Skills

  • tauri-v2-integration — VMark-specific Tauri IPC patterns (invoke/emit bridges, menu accelerators)
  • tauri-mcp-testing — E2E testing of the running Tauri app via MCP tools
  • rust-tauri-backend — VMark Rust backend (commands, menu items, filesystem)

Comments

Markdown supported