Translator AI Skill

Professional multilingual translation guidance system designed for meaning-focused, high-quality translations across multiple language pairs.

When to Use This Skill

Activate this skill when:

• User asks “translate X to [language]”
• User needs professional translation guidance
• User wants to understand translation decisions
• User seeks idiom or cultural adaptation help
• User needs quality-checked translations
• User is building multilingual content

Quick Start Workflow

1. UNDERSTAND THE TEXT
   └── Read fully, identify tone, purpose, audience

2. CLASSIFY
   └── Determine category: Technical / Marketing / Casual / Formal / Creative

3. IDENTIFY LANGUAGE PAIR & RULES
   └── Load appropriate rules from references/translation-rules.md

4. TRANSLATE WITH MEANING
   └── Focus on meaning, tone, cultural fit—not word-for-word

5. QUALITY CHECK
   └── Apply checklist from references/quality-checklist.md

6. DELIVER WITH NOTES
   └── Provide translation + decisions (if detailed output requested)

Core Principles (All Language Pairs)

1. Meaning Over Words – Translate meaning and intent, not words
2. Natural Output – Should read like native speaker wrote it originally
3. Tone Preservation – Match original tone (formal/casual/humorous/technical)
4. Format Integrity – Preserve structure (headers, lists, paragraphs, code)
5. No Additions – Never add information not in original
6. No Omissions – Never remove information from original

Language Pair Quality Assurance

English-Turkish (Primary)

• Status: Production-ready v2.0
• Coverage: 6 primary rules + 4 secondary rules + 5 special case categories
• Examples: 4 real-world examples with full analysis
• Quality Gates: 11-point comprehensive checklist
• Testing: Validated against real use cases
• Confidence Level: ⭐⭐⭐⭐⭐ Highest

English-Spanish / English-French / English-German (Framework Ready)

• Status: Framework ready, culture-specific rules needed
• How to Maintain Quality:
  1. Use base rules from references/translation-rules.md
  2. Apply culture-specific adaptations (in progress)
  3. Add language-pair examples (in progress)
  4. Validate against native speakers
  5. Update references/ as quality improves

How We Ensure Consistent High Quality Across All Language Pairs

1. Shared Foundation

All language pairs use the same 6 core principles and workflow, ensuring consistency in approach.

2. Culture-Specific Layers

Each language pair gets specific rules for:

• Grammar particularities
• Idiomatic expressions
• Cultural references
• Formality levels
• Common pitfalls

3. Progressive Quality Validation

Level 1: Framework Rules (all pairs)
  ↓
Level 2: Language-Specific Guidance (being added)
  ↓
Level 3: Real Examples (being collected)
  ↓
Level 4: Quality Checklist (pair-specific)
  ↓
Level 5: Native Speaker Validation (ongoing)

4. Quality Escalation for New Pairs

When adding a new language pair:

1. Document base rules from references/
2. Identify top 10 edge cases for that pair
3. Create 3-5 real examples
4. Build pair-specific quality checklist
5. Test with native speakers
6. Update skill with findings
7. Document lessons learned

5. Continuous Improvement

• Collect feedback from each use
• Document unexpected edge cases
• Update references/ with new insights
• Version skill as improvements accumulate
• Share updates across all agents

Using This Skill

Basic Usage (English-Turkish)

Simply request translation:

Translate to Turkish: [Your English text]

Detailed Output Mode

Get translation + reasoning:

Translate to Turkish (detailed): [Your text]

Specify Language Pair

Translate to Spanish: [Your English text]

Override Parameters

Translate to Turkish (formal, detailed, no cultural adaptation): [Text]

What You Get

Standard Output

• Translated text in target language
• Original formatting preserved
• Clean, ready-to-use output

Detailed Output (On Request)

• Translation category classification
• Addressing/formality decision
• Special translation decisions + reasoning
• Alternative options (if applicable)
• Cultural adaptation notes

Inside This Skill

• references/translation-rules.md – Complete rule system for all pairs
• references/special-cases.md – Idioms, humor, cultural refs, numbers, addressing
• references/quality-checklist.md – QC framework for all language pairs
• references/examples.md – Real-world examples with detailed analysis
• references/examples-100-complete.md – 100 comprehensive real-world scenarios (English + Turkish), professional grade, skills.sh-ready
• references/language-pair-quality-matrix.md – Current status of each pair
• references/grammar-comparison.md – Grammar structures across all 7 languages
• references/false-friends.md – Words with misleading similarities across languages
• references/common-mistakes.md – Critical translation errors to avoid per language
• assets/workflow-diagram.txt – Visual workflow

Maintenance & Updates

This skill is actively maintained. As translations improve:

• New language pairs will be added with full quality assurance
• Examples will be updated from real usage
• Rules will be refined based on feedback
• Quality metrics will be tracked

Current Version: 2.0 (English-Turkish production-ready) Last Updated: February 2026 Next Update: When new language pair reaches production quality

Ethics & Boundaries

• Never translate sensitive/private information without explicit approval
• Flag potentially offensive or culturally inappropriate content
• Refuse translations for harmful purposes
• Maintain confidentiality of translated content

Want more details? Each reference file contains deep guidance:

• Need translation rules? See references/translation-rules.md
• Handling idioms/humor? See references/special-cases.md
• Quality checking? See references/quality-checklist.md
• Real examples? See references/examples.md
• Language pair status? See references/language-pair-quality-matrix.md