Token导航 LogoToken导航

Agent Skill · 研究检索

translation(translation 搜索)

translation 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

translation 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,674 个 Skill 里排第 71,089。

总安装 196周安装 8GitHub Stars 67下载 63核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:translation(translation 搜索)
来源仓库:https://github.com/seb1n/awesome-ai-agent-skills
仓库路径:skills/translation
安装命令:
npx skills add https://github.com/seb1n/awesome-ai-agent-skills --skill translation
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/seb1n/awesome-ai-agent-skills --skill translation

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

translation 用于查找、检索和筛选相关信息,适合跨语言内容处理。适用于需要根据关键词或任务场景从来源中获取信息的场景。通过 npx skills add 命令安装指定 GitHub 仓库中的技能模块。安装前需确认权限范围、维护状态,以及是否触发联网或文件读写操作。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/translation

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/seb1n/awesome-ai-agent-skills --skill translat…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

translation 怎么安装?

在终端执行 npx skills add https://github.com/seb1n/awesome-ai-agent-skills --skill translation。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

translation 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

translation 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

translation 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

translation 用的人多吗?

公开数据是累计安装 196 次,近一周安装 8 次,下载 63 次,GitHub 67 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

translation 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
translation

translation 搜索 · 本页

19667接触敏感数据未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Translation

This skill enables an AI agent to translate text between languages while preserving meaning, tone, and context. It goes beyond word-for-word substitution by handling idiomatic expressions, cultural nuance, domain-specific terminology, and formality registers. The agent can translate technical documentation, marketing copy, UI strings, legal text, and general prose, adapting output to the conventions of the target locale.

Workflow

  1. Analyze Source Context Examine the source text to determine its domain (technical, marketing, legal, casual), register (formal, informal, neutral), and intended audience. Identify any idiomatic expressions, culturally specific references, or humor that will require adaptation rather than literal translation. Note the document format (plain text, Markdown, JSON, HTML) so structure is preserved in the output.
  2. Research Terminology For specialized content, build or consult a terminology glossary to ensure consistency. Technical documents require exact translations of domain terms (e.g., "endpoint" stays "endpoint" in many languages or maps to an accepted equivalent). Marketing copy may need creative transcreation rather than direct translation. If the user provides a glossary or translation memory, integrate it before proceeding.
  3. Translate the Content Produce the translation following the determined tone, formality level, and terminology. Preserve all formatting markers such as Markdown headings, code blocks, HTML tags, and JSON keys. For UI strings, keep placeholder tokens (e.g., {username}, %d) intact and in the correct grammatical position for the target language. Handle pluralization rules specific to the target language.
  4. Localize Cultural References Adapt units of measurement, date formats, currency symbols, and culturally bound metaphors to the target locale. For example, convert miles to kilometers when translating for a European audience, or replace an American sports analogy with one that resonates in the target culture. For marketing content, ensure slogans and calls-to-action carry the same persuasive impact in the target language.
  5. Review for Accuracy and Fluency Re-read the translated text to verify grammatical correctness, natural phrasing, and faithful representation of the source meaning. Check that no sentences were accidentally omitted or duplicated. Confirm terminology consistency across the entire document by cross-referencing the glossary established in step 2.
  6. Deliver with Annotations Return the translated text in the same format as the source. Where translation choices were ambiguous, include brief translator notes explaining the reasoning. If the user requested multiple formality levels or regional variants, provide each variant clearly labeled.

Usage

Provide the text to translate along with the target language. Optionally specify:

  • Source language (if not auto-detectable)
  • Formality level — formal, informal, or neutral
  • Domain — technical, marketing, legal, medical, general
  • Glossary — key terms and their required translations
  • Output format — plain text, Markdown, JSON locale file, etc.

Example prompt: "Translate the following API error messages to Japanese (formal). Keep the JSON structure and all placeholder tokens intact."

Examples

Example 1: Translating a Technical Paragraph (English to German, Formal)

Input:

The authentication module uses OAuth 2.0 with PKCE to secure API requests. When a user initiates a login, the client generates a code verifier and code challenge. The authorization server validates the challenge before issuing an access token. Tokens expire after 3600 seconds and must be refreshed using the refresh token endpoint.

Output:

Das Authentifizierungsmodul verwendet OAuth 2.0 mit PKCE, um API-Anfragen abzusichern. Wenn ein Benutzer eine Anmeldung einleitet, erzeugt der Client einen Code-Verifier und eine Code-Challenge. Der Autorisierungsserver validiert die Challenge, bevor er ein Zugriffstoken ausstellt. Token laufen nach 3600 Sekunden ab und müssen über den Refresh-Token-Endpunkt erneuert werden.

Translator note: "Code Verifier," "Code Challenge," and "Refresh-Token-Endpunkt" are kept close to the English originals because these are established terms in the German-speaking developer community.

Example 2: Localizing UI Strings (English to Spanish, JSON Locale File)

Input:

{
  "greeting": "Hello, {username}!",
  "items_in_cart": "You have {count} item(s) in your cart.",
  "checkout_button": "Proceed to Checkout",
  "empty_cart": "Your cart is empty. Start shopping!",
  "order_confirmation": "Your order #{orderId} has been placed on {date}."
}

Output:

{
  "greeting": "Hola, {username}!",
  "items_in_cart": "Tienes {count} artículo(s) en tu carrito.",
  "checkout_button": "Proceder al pago",
  "empty_cart": "Tu carrito está vacío. ¡Empieza a comprar!",
  "order_confirmation": "Tu pedido n.º {orderId} se ha realizado el {date}."
}

Translator note: Placeholder tokens {username}, {count}, {orderId}, and {date} are preserved exactly. The informal "tú" form is used; switch to "usted" for formal contexts.

Best Practices

  • Preserve formatting and structure. Never strip Markdown, HTML tags, or JSON keys during translation. The translated output must be drop-in compatible with the source format.
  • Maintain terminology consistency. Use the same translated term for a given source term throughout the entire document. Build a glossary at the start and reference it continuously.
  • Respect formality registers. Languages like Japanese, Korean, German, and Spanish have distinct formal and informal modes. Always confirm the desired register before translating.
  • Handle pluralization rules. Many languages have complex plural forms (e.g., Arabic has six). When translating UI strings with count-dependent text, account for the target language's pluralization system.
  • Avoid literal translation of idioms. Phrases like "hit the ground running" or "low-hanging fruit" should be replaced with equivalent expressions in the target language, not translated word-for-word.
  • Annotate ambiguous choices. When a source term has multiple valid translations, include a brief note explaining the chosen translation and alternatives.

Edge Cases

  • Untranslatable terms: Some technical terms (e.g., "API," "OAuth," "webhook") have no standard translation in certain languages. Keep them in the original language and note this decision.
  • Right-to-left languages: When translating to Arabic, Hebrew, or Persian, ensure that mixed-direction text (e.g., English brand names within Arabic sentences) is handled with proper bidirectional markers.
  • Character encoding: CJK languages may cause string length issues in fixed-width UI fields. Flag any translations that significantly exceed the source string's character count.
  • Gender-neutral language: Some target languages require grammatical gender. When the source is gender-neutral, ask the user for guidance or provide both gendered variants.
  • Empty or placeholder source text: If the source contains only placeholders or empty strings, return them unchanged and note that no translation was needed.

平台分布

Codex

35.45%
按下载量换算22

Claude

25.59%
按下载量换算16

Cursor

19.49%
按下载量换算12

Gemini CLI

9.69%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills