About turkish-native
turkish-native is a writing skill for producing and reviewing natural Turkish copy. It is intended for writers and agents handling translation, localization, or Turkish-language content, with guidance spanning sentence structure, domain fit, register, and factual fidelity.
Key features
- 01Rebuilds sentences around intended meaning
- 02Adapts language to audience and medium
- 03Checks domain-specific terminology and content
- 04Preserves factual scope and product details
- 05Includes 57 Turkish-specific pattern references
Use cases
- 01Localizing product and marketing copy into Turkish
- 02Reviewing Turkish pages for translation-like phrasing
- 03Editing technical or regulated-domain text
- 04Guiding an agent to revise Turkish prose
Install
Copy the command and run it in a terminal.
npx skills add https:/ / github.com/ oguzhankayan/ turkish-native --skill turkish-nativeAfter installing, type/turkish-nativein Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
.gitignoreLICENSEAGENTS.mdROADMAP.mdCHANGELOG.mdRELEASE_CHECKLIST.mdCITATION.cffREADME.tr.mdCODE_OF_CONDUCT.mdCONTRIBUTING.mdREADME.mdSECURITY.md.claude-plugin/plugin.jsonmarketplace.json.github/CODEOWNERSpull_request_template.md.github/ISSUE_TEMPLATE/bug_report.ymlpattern_proposal.ymlconfig.yml.github/workflows/validate.ymlrelease.ymlagents/openai.yamldocs/RELEASE_v0.3.0.mdRELEASE_v0.5.0.mdPHILOSOPHY.mdBENCHMARKING.mdevals/README.mdevals/benchmark/README.mdcases.yamlevals/cases/01-sentence-architecture.yamlDetails
From the project READMETurkish Native
Write Turkish as Turkish. Do not preserve the sentence skeleton of English or another source language and merely replace its words with Turkish ones.
A sentence can be grammatically valid and still sound translated. Work from meaning, not source-language shape.
Core test
Before keeping a sentence, ask:
If I had received only this meaning, with no source-language wording, would I naturally build the Turkish sentence this way?
If not, rebuild it from the intended meaning.
How to work
- Identify the job. Determine audience, medium, register, sector, and intended action. A landing-page hero, button label, legal notice, clinic page, and API document need different Turkish.
- Extract the meaning. Separate facts and required terminology from wording. Preserve names, numbers, claims, dates, links, commands, legal terms, identifiers, and product behavior.
- Fix structure before vocabulary. Check clause structure, word order, case, references, verb choice, and collocations first.
- Check domain fit. Make sure fields, CTAs, process steps, labels, and examples belong to the actual sector and task rather than a neighboring template.
- Rebuild instead of patching. If the sentence is structurally foreign or domain-wrong, rewrite it around its main meaning. Do not repair translationese one word at a time.
- Read it as Turkish. A native reader should not need to reconstruct another language mentally.
- Check fidelity. Do not invent, remove, strengthen, weaken, or generalize factual claims, operational promises, source labels, or regulated-domain outcomes.
- Check restraint. Do not rewrite natural Turkish merely because another phrasing is possible.
Voice and register
Match a supplied writing sample or existing product voice. Preserve deliberate short sentences, colloquial expressions, terminology, humor, or formality when they fit the audience.
Do not turn natural Turkish into corporate Turkish. Sen satışa bak. can be better than a longer formal alternative. Do not force every English loanword into Turkish: link, demo, dashboard, API, MCP, and commit may be natural in technical or product contexts.
What to return
- Write mode: return the finished Turkish text. Explain rules only when asked.
- Review mode: briefly identify important problems and give a revised version. For page reviews,
Bölüm | Mevcut | Önerilenis often useful. - File mode: change prose only unless instructed otherwise. Preserve code, commands, URLs, route names, placeholders, translation keys, variable names, frontmatter, and machine-readable structure.
- Embedded mode: when another agent invokes the skill, return only the requested final copy unless analysis is requested.
Translation and localization
Preserve meaning, scope, and product terminology, but do not preserve source-language word order, clause boundaries, or part of speech unless Turkish naturally uses the same structure. A good translation may split or merge sentences, drop an explicit pronoun, turn a noun into a verb, or replace a source-language metaphor with a direct statement.
If the user asks for a literal, legal, or line-by-line translation, honor that fidelity while still using grammatical Turkish.
Pattern references
This skill contains 57 Turkish-specific patterns. Read the relevant bundled reference before rewriting; for broad copy review or translation, read all six.
- 1–10 Sentence architecture:
references/01-sentence-architecture.md - 11–20 Verb choice and lexical fit:
references/02-verb-and-lexical-fit.md - 21–30 Product, UI, and localization:
references/03-product-ui-localization.md - 31–40 Marketing and model-written copy:
references/04-marketing-model-copy.md - 41–51 Register, fidelity, and final checks:
references/05-register-fidelity.md - 52–57 Domain and page integrity:
references/06-domain-page-integrity.md
These are diagnostic rules, not search-and-replace rules. Context always wins.
Non-negotiable checks
Before returning the text, verify that:
- meaning and factual scope are preserved,
- clause structure is Turkish rather than copied from another language,
- verbs fit their nouns and real actions,
- case suffixes and references are clear,
- product copy uses user language rather than database language,
- every field, CTA, process step, and example belongs to the actual sector and task,
- marketing copy does not repeat or inflate ordinary claims,
- operational promises, medical or other high-stakes claims, privacy statements, ratings, and source labels are supported,
- synthetic examples are not presented as sourced customer evidence,
- process sections describe a real process rather than narrating obvious buttons or navigation,
- legal, security, and technical text preserves precision,
- natural colloquial Turkish has not been “improved” into corporate prose,
- one repaired shape has not been copied onto every sentence, and no line repeats what another line on the same page already said,
- the result sounds native when read without the source.