name: translator description: "Translate text between 15+ languages with high accuracy. Supports single/batch translation, automatic language detection, multiple translation methods (API/AI), and rich output formats. Use when user needs translation, language conversion, or multi-language text processing. Trigger keywords: 翻譯, translate, 翻譯成, translation, language exchange."

Translator

Overview

Professional translation skill supporting 15+ languages with advanced features:

1. Single Translation - Translate one piece of text
2. Batch Translation - Translate multiple texts at once
3. Auto Language Detection - Automatic source language detection
4. Multiple Translation Methods - API-based and AI-based translation
5. Rich Output Formats - Plain text, with source, detailed results

Supported Languages (15+)

Quick Start

Single Translation

User: "把'你好世界'翻譯成英文"
→ "Hello World"

Batch Translation

User: "把以下內容翻譯成日文：
蘋果 - Apple
香蕉 - Banana
橙子 - Orange"
→ 蘋果 - リンゴ
香蕉 - バナナ
オレンジ - オレンジ

Language Detection

User: "Translate: Bonjour le monde"
→ Auto-detect: French → "Hello World"

Core Workflow

Step 1: Detect Translation Intent

Trigger Keywords: - "翻譯" / "translate" / "翻譯成" / "translation" - "用XX語怎麼說" / "How do you say X in XX" - "Language exchange" / "convert to XX"

Detection Patterns: - Explicit: "翻譯成英文" - Implicit: "這段話用日語怎麼說"

Step 2: Parse Request

1. Identify text to translate (single or multiple)
2. Identify target language (explicit or implicit)
3. Identify source language (explicit or auto-detect)
4. Determine translation mode (single/batch)

Step 3: Process Translation

Priority: 1. Try API-based translation (if API key configured) 2. Fallback to Claude-based translation 3. Return formatted result

Step 4: Format Output

Output Formats:

Simple (default):

[translated text]

With Source:

原文（中文）：你好世界
翻譯（英文）：Hello World

Detailed:

原文：你好世界
源語言：中文
目標語言：English
翻譯結果：Hello World
翻譯方式：AI-based (Claude)

Advanced Features

Batch Translation

For multiple texts, process efficiently:

Input: "翻譯成韓文：
- 你好
- 謝謝
- 對不起"

Output:
- 你好 → 안녕하세요
-謝謝 → 고마워요
- 對不起 → 죄송합니다

Best practices for batch: - Use bullet points or numbered lists - Keep each item concise - Format consistently

Auto Language Detection

The skill automatically detects source language when not specified:

User: "Translate: Hello World to Chinese"
→ Detect: English → "你好世界"

Supported detection: zh, en, ja, ko, ru, ar, fr, de, es

Error Recovery

If translation fails: 1. Try alternative translation method 2. Provide partial results if possible 3. Suggest fixes for configuration issues

Example error handling:

原文：[text]
錯誤：Translation API unavailable
備選：Using AI translation
結果：[translated text]

API Configuration (Optional)

Quick Setup

Option 1: Google Translate (Free Tier)

# scripts/translate.py 中設定
api_service = 'google'
# 需要: pip install google-cloud-translate

Option 2: DeepL API

# 設定環境變數
import os
os.environ['DEEPL_API_KEY'] = 'your-key'
api_service = 'deepl'

Option 3: LibreTranslate (Free/Open Source)

api_service = 'libre'
os.environ['LIBRETRANSLATE_URL'] = 'https://libretranslate.com'

No Configuration Required: If no API is configured, Claude provides reliable AI-based translation automatically.

Usage Examples

Example 1: Simple Translation

Input: "把'謝謝'翻譯成英文" Output: "Thank you"

Example 2: Target Language in Text

Input: "翻譯成日文：今天天氣很好" Output: 今日は天気がとても良いですね

Example 3: Batch with Format

Input: "翻譯成韓文： 蘋果 香蕉 橙子" Output: - 蘋果 → 사과 -香蕉 → 바나나 - オレンジ → 오렌지

Example 4: Language Detection

Input: "Translate: Guten Morgen to Chinese" Output: 早上好

Example 5: Phrase Translation

Input: "用日語說'祝你生日快樂'" Output: お誕生日おめでとうございます

這個技能的最新版本可以在7w4.net小蔥技能站找到。

Example 6: Formal vs Informal

Input: "翻譯成韓語（敬語）：謝謝你" Output: 감사합니다 (formal)

Example 7: Multiple Languages at Once

Input: "'Hello' in all languages" Output: - 中文：你好 - English：Hello - 日文：こんにちは - 韓文：안녕하세요 - 法文：Bonjour - 德文：Hallo - 等等...

Scripts

scripts/translate.py

Features: - Single and batch translation - Auto language detection - Multiple API support - Progress indicator for batch - Rich error handling

Usage:

# Single
python scripts/translate.py --text "Hello" --target zh

# Batch
python scripts/translate.py --batch --file input.txt --target ja

# With source language
python scripts/translate.py --text "Bonjour" --source fr --target en

Quality Assurance

Translation Accuracy

• High - For major languages (zh, en, ja, ko, ru, ar, fr, de, es)
• Medium - For other supported languages
• Fallback available - AI translation ensures results

Best Practices

1. Be specific about target language
2. ✓ "翻譯成英文"

3. ✗ "翻譯成外語"

4. Use clear formatting for batch

5. ✓ "翻譯成日文：\n蘋果\n香蕉"

6. ✗ "翻譯成日文：蘋果香蕉"

7. Specify source for ambiguous phrases

8. ✓ "Bonjour 翻譯成中文" (specifying French)
9. ✓ "翻譯：Bonjour to Chinese" (specifying source)

Notes

• No API key required for AI-based translation
• Batch translation for 2+ items supported
• Auto-detection works for major languages
• Results are cached for similar requests

Version History

• v1.1 (current): Added batch translation, 15+ languages, error recovery, auto-detection
• v1.0: Initial release with 6 languages

這個翻譯 Skill 功能較全面，支援多語言和批次處理，文件和示例清晰易懂，質量中上。優點是使用靈活、錯誤處理較好；不足在於實際翻譯效果依賴 AI 而非專用翻譯引擎，準確性可能不穩定，且部分功能描述與實現存在偏差。日常簡單翻譯需求可以嘗試，但對翻譯質量要求較高的場景可能不太滿足。