Context-aware translation that preserves tone, style, and natural word order. Use when translating UI strings, documentation, marketing copy, or any multilingual content. Infers register, domain, and style from the source text and surrounding codebase context.
10k tokens
context cost
the whole folder, loaded on every use
3
files
instructions only
0
copies elsewhere
how many repositories repackaged it
1201
stars on the repo
on the repository, not the skill itself
Install
one command, takes just this skill from the repository
Translate, review, or adapt multilingual content while preserving meaning, register, placeholders, structure, domain terminology, and natural target-language word order.
Intent signature
User asks to translate, localize, review translation quality, create a glossary, or adapt UI/docs/marketing copy.
User needs context-aware translation rather than mechanical word substitution.
When to use
Translating UI strings, error messages, or microcopy
Translating documentation, README, or guides
Translating marketing copy or landing pages
Reviewing existing translations for naturalness
Creating glossaries or translation style guides
Any task involving multilingual content
When NOT to use
i18n infrastructure setup (key extraction, routing, build) -> use dev-workflow
Adding new locale to framework config -> use dev-workflow
Domain terms: Words that need consistent translation (check existing translations first)
Cultural references: Idioms, metaphors, humor that won't transfer directly
Sentence rhythm: Short/punchy vs. long/flowing — note parallel structures, intentional repetition, and emphasis patterns
Comprehension challenges: Terms or references target readers may struggle with — domain jargon lacking standard translations, cultural references (pop culture, history, social norms), implicit knowledge the author assumes, wordplay or puns, named concepts (e.g., "Dunning-Kruger effect"). For each, note: the original term, why it may confuse, and a concise plain-language explanation for a potential translator's note
Figurative language mapping: For each metaphor, simile, idiom, or figurative expression, classify the handling approach:
Interpret: Discard source image entirely, express the intended meaning directly in natural target language
Substitute: Replace with a target-language idiom or image that conveys the same idea and emotional effect
Retain: Keep the original image if it works equally well in the target language
Emotional connotations: Words carrying subjective feeling beyond dictionary meaning (e.g., "alarming" = urgency, "haunting" = lingering unease) — note the emotional effect to preserve in translation
Stage 2: Extract Meaning
Strip away source language structure. Ask yourself:
What is the author actually trying to say?
What emotion or tone should the reader feel?
What action should the reader take?
Do NOT start forming target sentences yet.
Stage 2.5: Persona Assignment
Persona resolution has two layers: content-type (what kind of text) and voice (how punchy or formal the rhythm). Both are needed.
Layer 1: Read translation_voice from .agents/oma-config.yaml
The translation_voice field controls global rhythm/formality. Three values:
| Voice | Style override applied on top of content-type |
|---|---|
| formal | complete sentences only, no fragments, strict 합니다체/です・ます, no padding cuts |
| balanced (default) | content-type defaults — fragments allowed only in label/cell positions |
| interpreter | interpreter mindset across all content types: punchy, audience-first, spoken cadence, fragments allowed when natural in target, drops formal padding ("을 받았습니다" → "받음" / "을 모두" → drop) |
If the field is missing, default to balanced. If oma-config.yaml is unreadable, also balanced.
Content-type = technical reporter + voice = balanced → complete sentences in body, fragments allowed in table cells (current default).
Content-type = technical reporter + voice = interpreter → punchier rhythm, list-item fragments allowed (e.g., "39턴 / 8m 13s / $1.28 (파일당 $0.14)" instead of "39턴, 8m 13s, 총 $1.28을 썼습니다(파일당 약 $0.14)"), drops "을 모두 받았습니다" padding.
The persona is then localized to the target language at execution time — translating into Korean as a "technical reporter" with interpreter voice means thinking as a Korean technical reporter who values rhythm and audience scan-speed over formal completeness.
Stage 3: Reconstruct in Target Language
Rebuild from meaning as the assigned persona, following target language norms:
Word order: Follow target language's natural structure.
EN → KO: SVO → SOV, move verb to end, particles replace prepositions
EN → JA: Similar SOV restructuring, honorific system alignment
EN → ZH: Maintain SVO but restructure modifiers (pre-nominal in ZH)
Register matching:
Infer from existing translations in the project, or from source text tone
English compound sentences often split into shorter Korean/Japanese sentences
English bullet points may merge into flowing paragraphs in some languages
Omission of the obvious:
Many languages (Korean, Japanese, Chinese, etc.) allow subject or pronoun omission when contextually clear
Don't force subjects or pronouns that feel unnatural in the target language
Stage 4: Verification Gate (blocking — do not emit output until every item is confirmed)
This stage is mandatory. Skipping any item is a bug, not a shortcut. Before producing the final translation, run the mechanical checks first, then the rubric.
A. Mechanical checks (run before rubric, must all pass):
CJK em dash scan: For Korean, Japanese, or Chinese targets, search the draft output for —. Every occurrence must be replaced with a comma, colon, parenthesis, or restructured sentence. Zero em dashes in the emitted output.
Placeholder integrity: Every {name}, {{count}}, %s, <tag>, and code from the source appears unchanged in the target.
Structure parity: Headings, list bullets, table rows, code blocks, and links match the source count and nesting.
Register consistency: One sentence-ending style throughout (don't mix -ㅂ니다 with -다, formal with casual).
If any mechanical check fails, revise and re-run. Do not proceed to the rubric until all pass.
B. Translation rubric (see resources/translation-rubric.md):
Does it read like it was originally written in the target language?
Are domain terms consistent with existing translations in the project?
Is the register consistent throughout?
Is the meaning preserved (not just words)?
Are cultural references adapted appropriately?
Are emotional connotations preserved (not flattened into neutral descriptions)?
C. Anti-AI patterns (see resources/anti-ai-patterns.md):
No AI vocabulary clustering or inflated significance
No promotional tone upgrade beyond the source
No synonym cycling — consistent terminology
10. No source-language word order leaking through
11. No unnecessary bold or formatting artifacts (em dashes already covered in mechanical check A)
13. Were all metaphors/idioms handled per the classify decision (interpret/substitute/retain)?
14. Do figurative expressions read naturally in the target language, not as literal calques?
Translator's Notes Guidelines
When adding explanatory notes for terms, cultural references, or concepts that target readers may struggle with:
Format: 번역어（원어, 쉬운 설명） or 번역어(원어) for well-known terms that just need the original
Calibration by audience:
Technical readers: Skip annotation on common tech terms (API, deploy, refactor). Only annotate domain-specific or coined terms
General readers: More generous annotation. Explain jargon, cultural references, and domain concepts in plain language
Short texts (< 5 sentences): Minimize — only annotate terms the target audience is unlikely to know
Rules:
Annotate on first occurrence only — don't repeat the note
Keep notes concise (aim for under 10 words)
Explain *what it means*, not just provide the English original
Don't annotate self-explanatory terms or widely recognized loanwords
If a comprehension challenge was identified in Stage 1, use the pre-planned explanation
Reflection Mode (default for non-trivial content)
Reflection passes (Stage 5–7) are the default — not optional — for any content that is more than a short snippet. Empirical evidence (Slator 2024, Self-Refine paper) shows a single polish pass cuts translationese rates roughly in half. Skipping reflection on non-trivial content is the most common cause of translationese complaints.
When to run Stage 5–7
Default ON for:
Documentation (README, guides, API reference)
Reports, benchmarks, changelogs, blog posts
Marketing copy and landing pages
Any prose longer than ~3 sentences
Anything containing tables, bullet lists, or code blocks mixed with prose
Translation review mode
Default OFF (Stage 4 verification only) for:
Single short UI string (< 10 words) with established glossary
Batch UI key translations where each value is independent and < 1 sentence
User explicitly requests "fast translation", "skip reflection", or "직역"
When in doubt, run reflection. The cost is roughly 1.5–2× tokens; the quality gain on body-text fragments and Europeanized patterns is large.
Extended workflow
After completing Stage 1–4, continue with:
Stage 5: Critical Review
Re-read the translation against the source with fresh eyes. Produce a diagnostic review (no rewriting yet):
Accuracy: Compare paragraph by paragraph — any facts, numbers, or qualifiers altered?
Europeanized language: Scan for unnecessary connectives, passive voice, noun pile-up, over-nominalization, forced pronouns (see resources/anti-ai-patterns.md)
Figurative language fidelity: Cross-check metaphor mapping from Stage 1 — were all handled per the classify decision? Any literal calques that sound unnatural?
Emotional fidelity: Were subjective/emotional word choices flattened into neutral descriptions?
Tone drift: Does the register stay consistent from start to finish, or does it shift mid-document (e.g., formal intro drifting into casual explanation)?
Expression & flow: Flag sentences that still read like "translation-ese" — stiff phrasing, unnatural word order, awkward transitions
Translator's notes quality: Too many? Too few? Accurate and concise?
Stage 6: Revision
Apply all findings from Stage 5 to produce a revised translation:
Fix accuracy issues
Rewrite Europeanized expressions into native patterns
Re-interpret literally translated metaphors per the mapping
Restore flattened emotional connotations
Restructure stiff sentences for fluency
Adjust translator's notes per review recommendations
Stage 7: Polish
Final pass for publication quality:
Read as a standalone piece — does it flow as native content?
Smooth remaining rough transitions between paragraphs
Ensure narrative voice is consistent throughout
Final scan for surviving literal metaphors or translation-ese
File editing tools only when the user requests file changes
Canonical workflow path
1. Analyze source register, intent, domain terms, placeholders, and structure.
2. Reconstruct meaning in the target language, not word-for-word.
3. Run mechanical checks and `resources/translation-rubric.md` before emitting output.
| USER_DATA | User-provided text and target-language requirements |
Preconditions
Source text and target language are known.
Placeholder and structure constraints are identifiable.
Ambiguities are resolved or explicitly flagged.
Effects and side effects
Produces translated text or translation review.
May modify locale/docs files only when requested.
Preserves source structure and placeholders.
Guardrails
Scan existing locale files before translating to align with project conventions
Preserve placeholders and interpolation syntax
Translate meaning, not words
Preserve emotional connotations — translate the feeling, not just the dictionary meaning (e.g., "alarming" carries urgency/concern, not merely "surprising")
Match register consistently throughout a single piece
Split, merge, or restructure sentences for target language naturalness
Flag ambiguous source text rather than guessing
Preserve domain terminology — if a term has established meaning in the field (e.g., harness, scaffold, shim, polyfill, middleware), keep it even if a "simpler" native word exists
Never produce literal word-for-word translations
10. Never mix registers within a single piece (formal + casual)
11. Never replace domain-specific terms with generic equivalents (e.g., "harness" → "framework", "shim" → "wrapper")
12. Never translate proper nouns unless existing translations do so
13. Never change the meaning to "sound better"
14. Never skip verification stage for batches > 10 strings
15. Never modify source file structure (keys, nesting, comments)
16. Never preserve source-language formatting artifacts that are unnatural in the target language. For CJK targets (Korean, Japanese, Chinese), em dashes (—), title case in headings, and trailing "-ing" participle clauses must be restructured — even when the source uses them. See resources/anti-ai-patterns.md rules 13–16.