
Translation Expertise
Give your agent Traditional Chinese (繁體) grammar, measure words, and register rules so docs, UI strings, and support copy localize accurately for Taiwan/Hong Kong audiences.
Overview
Translation Expertise is an agent skill most often used in Build (also Launch distribution, Grow content) that supplies Traditional Chinese linguistic rules so agents translate docs and UI copy with correct grammar, 量詞,
Install
npx skills add https://github.com/shino369/claude-code-personal-workspace --skill translation-expertiseWhat is this skill?
- Reference corpus for Traditional Chinese structural and grammatical conventions (SVO, aspect markers, 量詞)
- Distinguishes 繁體字 usage for Taiwan, Hong Kong, and Macau versus Simplified norms
- Covers omission, topic-prominent phrasing, and 成語 idioms agents often mishandle in EN→ZH-TW work
- Explains tone and aspect particles (了, 過, 著) for natural tense and completion in translation
- No executable automation—pure linguistic expertise for agent-mediated translation tasks
Adoption & trust: 526 installs on skills.sh; 2 GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your agent outputs stiff or Simplified-flavored Chinese because it lacks explicit Traditional grammar, measure-word, and register guidance.
Who is it for?
Indie makers localizing product UI and docs to 繁體中文 who want consistent agent behavior across README, strings.xml, and help center drafts.
Skip if: Certified legal, medical, or financial translation requiring licensed human review, or projects targeting Simplified Chinese (简体) markets exclusively.
When should I use this skill?
Agent-mediated translation or copy editing into Traditional Chinese for Taiwan, Hong Kong, or Macau when linguistic accuracy matters.
What do I get? / Deliverables
Translations and localized strings follow documented ZH-TW structural norms so you can paste into docs, l10n files, or support templates with fewer manual rewrites.
- Traditional Chinese draft text aligned to documented grammar and register rules
- Localization notes on idioms or measure-word choices
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Build/docs because most invocations start when writing or translating in-repo documentation and string catalogs. Docs subphase is primary for reference-driven translation of READMEs, comments, and product copy—not ASO keyword research or paid ad creative alone.
Where it fits
Translate API README and error messages to 繁體 while keeping technical terms consistent.
Refine strings.xml or React i18n JSON entries for Taiwan users with correct 量詞 and polite register.
Polish landing hero and feature bullets for HK/Macau audiences without Simplified character drift.
Localize changelog posts and help-center articles with natural topic-comment phrasing.
Draft support macros that respect omission and aspect markers in customer-facing ZH-TW replies.
How it compares
Use as a language reference skill—not a CAT tool integration, TMS connector, or automatic i18n pipeline.
Common Questions / FAQ
Who is translation-expertise for?
Solo builders and tiny teams using Claude Code (or similar) who need Traditional Chinese translation quality for documentation, UI strings, and user-facing copy.
When should I use translation-expertise?
While building docs and string catalogs; before launch when polishing landing and store copy for TW/HK; when growing support articles and lifecycle emails in 繁體中文.
Is translation-expertise safe to install?
It is static reference text in a personal workspace with no declared tool permissions; still review Security Audits on this Prism page and avoid pasting secrets into translation prompts.
SKILL.md
READMESKILL.md - Translation Expertise
# Chinese Traditional Language Reference for Translation This document contains comprehensive information about Traditional Chinese language characteristics, conventions, and considerations relevant to translation work. ## Core Characteristics ### Structural Features - **Word Order**: Subject-Verb-Object (SVO) like English - **No Verb Conjugation**: Tense indicated by context or time words - **Subject Omission**: Context-dependent (like Japanese) - **Measure Words**: Classifiers (量詞) required for counting nouns - **Tone Language**: 4 tones in Mandarin, 6-9 in Cantonese (not visible in writing) - **Idiomatic Expressions**: Rich tradition of four-character idioms (成語 - chéngyǔ) - **Topic-Prominent**: Topic-comment structure similar to Japanese ### Writing System **Traditional Characters (繁體字)**: - More complex forms than Simplified Chinese - Used in Taiwan, Hong Kong, Macau - Preserves historical character forms - Often more visually descriptive **Character Structure**: - Radicals (部首): Component parts that often indicate meaning - Stroke order: Specific order for writing characters - Complex characters can have 20+ strokes ### Grammatical Features **No Grammatical Tense**: - Time expressed through context or time words - 了 (le): Completion or change of state - 過 (guo): Past experience - 著 (zhe): Continuous action **Aspect Markers**: - 了 (le): Completed action (我吃了飯 - I ate) - 過 (guo): Experience (我去過日本 - I've been to Japan) - 著 (zhe): Continuous state (門開著 - The door is open) - 在 (zài): Progressive action (我在吃飯 - I'm eating) **No Grammatical Number**: - Singular/plural not typically marked on nouns - Context or measure words indicate quantity - 們 (men) can pluralize pronouns: 我們 (women - we) **Sentence-Final Particles**: - 嗎 (ma): Question marker - 呢 (ne): Softens questions, indicates continuation - 吧 (ba): Suggestion, supposition - 啊 (a): Exclamation, emphasis - 了 (le): Change of state, new situation ## Measure Words (量詞 - Liàngcí) Measure words (classifiers) are required between numbers/demonstratives and nouns. ### Common Measure Words **General**: - 個 (gè): General classifier, people, abstract things - 一個人 (yī gè rén): One person - 三個蘋果 (sān gè píngguǒ): Three apples **People**: - 位 (wèi): People (polite/formal) - 兩位客人 (liǎng wèi kèrén): Two guests - 個 (gè): People (casual) **Objects**: - 本 (běn): Books, volumes - 一本書 (yī běn shū): One book - 張 (zhāng): Flat objects (paper, tables, beds, tickets) - 兩張紙 (liǎng zhāng zhǐ): Two sheets of paper - 件 (jiàn): Items, clothing, matters - 三件衣服 (sān jiàn yīfú): Three pieces of clothing - 隻/只 (zhī): Animals, one of a pair - 一隻貓 (yī zhī māo): One cat - 條 (tiáo): Long narrow objects (rivers, roads, fish, pants) - 一條河 (yī tiáo hé): One river - 輛 (liàng): Vehicles - 兩輛車 (liǎng liàng chē): Two cars - 棟 (dòng): Buildings - 一棟房子 (yī dòng fángzi): One building **Quantities**: - 杯 (bēi): Cups/glasses - 一杯咖啡 (yī bēi kāfēi): One cup of coffee - 碗 (wǎn): Bowls - 一碗飯 (yī wǎn fàn): One bowl of rice - 瓶 (píng): Bottles - 兩瓶水 (liǎng píng shuǐ): Two bottles of water ### Translation Consideration - English/Japanese don't always require equivalents - Choose appropriate measure word for the noun - Incorrect measure word sounds unnatural to native speakers - Some nouns have specific measure words that must be learned ## Regional Variations ### Taiwan Mandarin (台灣華語 / 國語 - Guóyǔ) **Characteristics**: - More formal and literary influenced by traditional education - Retains classical expressions and vocabulary - Uses Traditional Chinese characters - More use of 您 (nín - polite "you") in formal contexts **Unique Vocabulary**: - 軟體 (ruǎntǐ): Software (Taiwan) - 駭客 (hàikè): Hacker (Taiwan) - 網路 (wǎnglù): Internet (Taiwan) - 資訊 (zīxùn): Information/IT (Taiwan) - 程式 (chéngshì): Program (Taiwan) **Pronunciation Differences**: - 和 (hé): Often pronounced as hàn in Taiwan - Some tonal variations from mainland Mandarin - Taiwanese influence on certain wor