Translation Diff Export
Devolutions/UniGetUI
Compares UniGetUI JSON locale files against English, identifies untranslated or source-changed keys, and generates patch, reference, and handoff files for a target language.
Translate new/changed English UI strings into all supported languages using a translate-review-refine pipeline.
$ npx skills add shapeshift/web --skill translate -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install shapeshift/web translate --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/translate .claude/skills/translate && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .claude/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/shapeshift/web/tree/develop/.claude/skills/translateType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add shapeshift/web --skill translate -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install shapeshift/web translate --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/translate .agents/skills/translate && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .agents/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add shapeshift/web --skill translate -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install shapeshift/web translate --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/translate .cursor/skills/translate && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .cursor/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/shapeshift/web.git --path .claude/skills/translate--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add shapeshift/web --skill translate -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install shapeshift/web translate --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/translate .gemini/skills/translate && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .gemini/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install shapeshift/web translateInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add shapeshift/web --skill translate -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/translate .github/skills/translate && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .github/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add shapeshift/web --skill translate -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install shapeshift/web translate --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/shapeshift/web.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/translate .opencode/skills/translate && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "translate" agent skill from https://github.com/shapeshift/web/tree/develop/.claude/skills/translate into .opencode/skills/translate/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "translate", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
translateTranslate new/changed English UI strings into all supported languages using a translate-review-refine pipeline.
Translate is an agent skill from shapeshift/web. Translate new/changed English UI strings into all supported languages using a translate-review-refine pipeline. Invoke with /translate to detect untranslated strings and produce high-quality translations for de, es, fr, ja, pt, ru, tr, uk, zh.
Its SKILL.md is about 4.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 19 other files, including scripts (for example `locales/de.md`, `locales/es.md` and `locales/fr.md`).
It sits in Writing & Content, covering Translation. The licence is MIT.
6 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 45096d2. It shows what the files ask for, not the result of running them.
Pre-approves these tools, so the agent can use them without asking each time:
ReadWriteEditGrepGlobBash(node *)Bash(git diff*)Bash(git log*)Bash(git rev-parse*)Bash(wc *)…and 3 more on the same allowed-tools line.
From allowed-tools in the SKILL.md frontmatter.
Ships 8 files in scripts/ (JavaScript), which the agent can run.
Shell commands in SKILL.md call:
nodegitFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Translate loads about 4.5k tokens when it runs. Until then it costs about 63 tokens; SKILL.md has 1,282 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
The full file from shapeshift/web at commit 45096d2, republished under its MIT licence (© shapeshift). 1,282 words, ~4,504 tokens.
.claude/skills/translate/SKILL.md (or your agent's skills folder). This skill also uses 17 other files; get the full folder from GitHub.Translates new or changed English UI strings into all 9 supported non-English languages using a translate-review-refine pipeline. Detects what changed, translates in batches with terminology glossary enforcement, validates output programmatically, and merges results into the existing translation files.
| Locale | Language | Register |
|---|---|---|
de | German | Formal (Sie) |
es | Spanish | Informal (tú) |
fr | French | Formal (vous) |
ja | Japanese | Polite (です/ます) |
pt | Portuguese (BR) | Informal (você) |
ru | Russian | Formal (вы) |
tr | Turkish | Formal (siz) |
uk | Ukrainian | Formal (ви) |
zh | Chinese (Simplified) | Neutral/formal |
See .claude/skills/translate/locales/{locale}.md for detailed locale-specific rules.
en — English is the source language.id or ko — these locales exist on disk but are not imported in index.ts or declared in constants.ts.%{variableName} placeholders must survive translation intact, including exact variable names.saveJSONFile in scripts/translations/utils.ts).Determine which English strings need translation.
Read src/assets/translations/.last-translation-shaThe marker contains a git SHA from the last translation run. Use it to find new/modified English strings since then.
node .claude/skills/translate/scripts/diff.jsThis follows the same recursive comparison pattern as scripts/translations/utils.ts findStringsToTranslate().
Compare each non-English language against English to find missing keys:
node .claude/skills/translate/scripts/missing-keys.jsIf no changes or missing keys are found, print:
Translations are up to date. No new or modified English strings detected.
and stop.
Mark these as needing re-translation in ALL languages, not just those missing the key.
Group the strings to translate by their top-level JSON namespace (the first segment of the dotted path, e.g., agenticChat, common, trade).
{ dottedPath, englishValue }Read src/assets/translations/glossary.json. This file contains:
null = never translate, keep in English (e.g., "Bitcoin", "DeFi", "MetaMask")The glossary is passed to every translator and reviewer sub-agent. If the glossary file is missing or contains invalid JSON, log an error and exit immediately with a clear message.
Before translation begins, extract significant terms from the new English strings and search all namespaces in the target locale for existing translations containing those terms. This ensures that domain terms like "pool", "vault", "bridge", "fee", etc. are translated consistently with how they've already been translated elsewhere in the app — even across different namespaces.
Run the term-context script once per locale at pipeline start:
node .claude/skills/translate/scripts/term-context.js LOCALE NEW_STRINGS_JSON_OR_FILELOCALE — the target locale code (e.g., fr)NEW_STRINGS_JSON_OR_FILE — path to a temp file containing the new strings as [{ path, value }] array or { "dotted.path": "english value" } objectThe script:
{locale}/main.json for existing translated strings whose English source contains those termsOutput format:
{
"liquidity pool": [
{ "key": "defi.liquidityPools.title", "en": "Liquidity Pool", "fr": "Pool de liquidité" }
],
"fee": [
{ "key": "common.gasFee", "en": "Gas Fee", "fr": "Frais de gas" },
{ "key": "trade.networkFee", "en": "Network Fee", "fr": "Frais de réseau" }
]
}Pass this as {TERM_CONTEXT} in the translator prompt (see Step 5a). If the script returns an empty object (no matching terms found in existing translations), omit the {TERM_CONTEXT} section from the prompt.
Caching: Run term-context.js and load few-shot context once per locale at pipeline start. Reuse this context for all batches of that locale — do not re-run scripts per batch.
Before translating each batch for a given locale, sample ~10 existing translations from the same namespace(s) being translated. These serve as tone/style reference for the translator.
src/assets/translations/{locale}/main.json{ "dotted.path": { "en": "English source", "{locale}": "Existing translation" } }{EXISTING_TRANSLATIONS_SAMPLE} in the translator prompt (see Step 5a)This grounds the translator in the existing voice and terminology of the project for that locale.
Caching: Load the locale file once at pipeline start and reuse for all batches of that locale.
After loading few-shot context (Step 4), run prepare-locale.js once per locale to bundle all translation context into a single file. This prevents sub-agents from needing to read any codebase files.
node .claude/skills/translate/scripts/prepare-locale.js LOCALE --batches=BATCHES_FILE --term-context=TERM_CONTEXT_FILE --few-shot=FEW_SHOT_FILEBATCHES_FILE — path to a JSON file containing an array of batch objects (each batch is { "dotted.path": "english value" })TERM_CONTEXT_FILE — path to term-context output from Step 3b (omit if empty)FEW_SHOT_FILE — path to few-shot context from Step 4 (omit if empty)Output: /tmp/translate-{locale}.json containing locale rules, glossary, term context, few-shot examples, and all batches in a single file.
Launch 9 Task sub-agents in parallel (one per language) using the Task tool with model: "sonnet". Each language agent owns its entire lifecycle: translate → validate → retry → review → refine → merge → verify.
The orchestrator's only job after spawning is to read status files and compile the report (Step 6).
For each locale, spawn a Task with the following prompt (substituting {LOCALE_CODE} and {LANGUAGE_NAME}):
You are a self-contained translation agent for {LANGUAGE_NAME} ({LOCALE_CODE}) in a cryptocurrency/DeFi application.
You own the full translation lifecycle for your locale. Do NOT read any codebase source files — all context is in the locale bundle.
## Your Locale Bundle
Read `/tmp/translate-{LOCALE_CODE}.json`. It contains:
- `locale`, `language`, `register` — your target locale metadata
- `localeRules` — locale-specific translation rules (follow these precisely)
- `neverTranslate` — terms that must remain in English
- `approvedTerms` — terms with mandatory translations for your locale
- `termContext` — how key terms have been translated elsewhere in this project
- `fewShot` — reference translations for tone/style
- `batches` — array of batch objects, each containing:
- `strings` — the key-value pairs to translate (`{ "dotted.path": "english value" }`)
- `relevantNeverTranslate` — never-translate terms that appear in this batch's strings
- `relevantApprovedTerms` — approved translations relevant to this batch
## Per-Batch Pipeline (process batches sequentially, 0-indexed)
For each batch in the `batches` array:
### 1. Translate
Translate all strings in `batch.strings` from English to {LANGUAGE_NAME}.
GLOSSARY REMINDER for this batch:
- Never translate these terms (keep in English): {batch.relevantNeverTranslate}
- Use these approved translations: {batch.relevantApprovedTerms}
RULES:
1. INTERPOLATION: Preserve all %{variableName} placeholders exactly as-is. Do not translate variable names inside %{}.
2. TERMINOLOGY:
- NEVER TRANSLATE terms in `relevantNeverTranslate` for this batch (keep in English)
- USE APPROVED TRANSLATIONS from `relevantApprovedTerms` for this batch
- Also reference the full `neverTranslate` and `approvedTerms` in the bundle as fallback
- When a term in `termContext` has an established translation, use it unless the context clearly demands a different meaning
3. Keep translations concise — UI space is limited. Match the approximate length of the English source.
4. FORMAT: Preserve HTML entities and markdown. If a string is a single word that's also a UI label (like "Done", "Cancel"), translate it as a UI action.
5. SOURCE FAITHFULNESS: Do not add information not present in the English source.
6. CONCISENESS: Prefer shorter synonyms or abbreviations common in {LANGUAGE_NAME} UI conventions.
7. KEY INTEGRITY: Output keys must EXACTLY match input keys. No additions, removals, or modifications to key names.
8. TAG KEYS: If the key path contains `.tags.`, the value is likely a short label or abbreviation. Preserve abbreviations as-is without expanding them. Check the `tagKeys` array in the bundle to identify these keys.
Use the `fewShot` examples from the bundle as tone/style reference.
### 2. Validate
Write the source batch (`batch.strings`) to `/tmp/batch-{LOCALE_CODE}-{BATCH_IDX}-source.json` and your translation to `/tmp/batch-{LOCALE_CODE}-{BATCH_IDX}-target.json`, then run:
```bash
node .claude/skills/translate/scripts/validate.js {LOCALE_CODE} /tmp/batch-{LOCALE_CODE}-{BATCH_IDX}-source.json /tmp/batch-{LOCALE_CODE}-{BATCH_IDX}-target.jsonThis outputs { rejected, flagged, passed }.
For any strings in rejected: re-translate them incorporating the rejection reason as feedback. Run validation again. Retry up to 2 times total. Strings that still fail after 2 retries become "manual review" items.
Collect flagged strings plus a 10% random sample of passed strings. If there are any strings to review, spawn a separate reviewer sub-agent using the Task tool (model: sonnet) with this prompt:
You are a senior localization reviewer for a cryptocurrency/DeFi application ({LANGUAGE_NAME}).
Do NOT read any other files from the codebase. All context you need is provided below.
Review these translations for quality. For each string, respond with either "approved" or a specific issue description.
LOCALE RULES:
{LOCALE_RULES_FROM_BUNDLE}
FOCUS ON:
1. Naturalness - does it sound natural to a native {LANGUAGE_NAME} speaker?
2. Semantic accuracy - does the translation accurately convey the English meaning?
3. Cultural appropriateness - are there any culturally awkward or inappropriate phrasings?
4. UI appropriateness - translations should be concise enough for UI elements
5. Source faithfulness - verify translation doesn't add information not in the English source
6. Register consistency — verify ALL address forms, verb conjugations, imperatives, and possessives match the declared register. Check beyond just pronouns: verb forms, possessives, and sentence endings must all be consistent.
TERM CONTEXT:
{TERM_CONTEXT_FROM_BUNDLE}
VALIDATION FLAGS:
{FLAGS_FOR_FLAGGED_STRINGS_OR_NONE}
STRINGS TO REVIEW (JSON: { "path": { "en": "source", "translation": "target" } }):
{STRINGS_TO_REVIEW}
OUTPUT: JSON object with dotted paths as keys. Value is either "approved" or "Issue: [one-sentence description]".If the reviewer flagged any strings, spawn a separate refiner sub-agent using the Task tool (model: sonnet):
You are a professional UI translator for a cryptocurrency/DeFi application.
Do NOT read any other files from the codebase. All context you need is provided below.
Fix the following {LANGUAGE_NAME} translations based on reviewer feedback.
LOCALE RULES:
{LOCALE_RULES_FROM_BUNDLE}
RULES: Preserve %{placeholders}, use approved terminology, be concise, be faithful to source.
INPUT: { "dotted.path": { "en": "source", "translation": "current", "feedback": "Issue: ..." } }
STRINGS TO FIX:
{STRINGS_WITH_FEEDBACK}
OUTPUT: JSON object with dotted paths as keys and corrected translations as values.Re-validate refined output. If it still fails after 1 retry, mark as "manual review".
After processing all batches, combine all passing translations into a single object.
After all batches are complete:
Write accumulated translations to /tmp/translations-{LOCALE_CODE}-final.json
Run merge (which creates a pre-merge backup automatically). By default, merge only adds new keys — existing translations are never overwritten. Pass --force only when re-translating changed English strings:
node .claude/skills/translate/scripts/merge.js {LOCALE_CODE} /tmp/translations-{LOCALE_CODE}-final.jsonRun post-merge validation:
node .claude/skills/translate/scripts/validate-file.js {LOCALE_CODE} --pre-merge=/tmp/pre-merge-{LOCALE_CODE}.jsonIf validate-file reports valid: false:
/tmp/pre-merge-{LOCALE_CODE}.json back to src/assets/translations/{LOCALE_CODE}/main.jsonWrite status to /tmp/translate-status-{LOCALE_CODE}.json:
{
"locale": "{LOCALE_CODE}",
"status": "success" | "failed",
"translated": <count>,
"manualReview": [{ "path": "...", "reason": "..." }],
"errors": ["..."]
}manualReview array in status file
### Temp File Conventions
All files are namespaced by locale — zero overlap between parallel agents:
| File | Writer | Reader |
|------|--------|--------|
| `/tmp/translate-{locale}.json` | orchestrator | language agent |
| `/tmp/batch-{locale}-{idx}-source.json` | language agent | validate.js |
| `/tmp/batch-{locale}-{idx}-target.json` | language agent | validate.js |
| `/tmp/translations-{locale}-final.json` | language agent | merge.js |
| `/tmp/pre-merge-{locale}.json` | merge.js | language agent (rollback), validate-file.js |
| `/tmp/translate-status-{locale}.json` | language agent | orchestrator |
## Step 6: Update Marker & Report
After all 9 language agents complete, read their status files and compile results.
1. **Read status files**: Read `/tmp/translate-status-{locale}.json` for each locale. If a status file is missing, report that locale as "no response" (agent may have crashed — locale file is unchanged).
2. **Write marker file** (only if at least one locale succeeded):
```bash
git rev-parse HEAD > src/assets/translations/.last-translation-shaUpdate glossary timestamp (if glossary was modified during this run):
Update _meta.lastUpdated in src/assets/translations/glossary.json to today's date.
Print summary report:
=== Translation Summary ===
SHA marker: <sha>
Strings translated: <count> across <locale_count> languages
Strings skipped (manual review needed): <count>
Locales failed (rolled back): <count>
Per-language breakdown:
de: <count> translated, <count> skipped [success|failed|no response]
es: <count> translated, <count> skipped [success|failed|no response]
...
Skipped strings (need manual review):
- <dottedPath> (<locale>): <reason>After all translations complete, scan for English terms that appear untranslated (kept as-is) in 7 or more locales. These are candidates for the glossary never-translate list.
src/assets/translations/glossary.json with value null© shapeshift, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 17 other files (scripts) in .claude/skills/translate of shapeshift/web.
Open the folder on GitHubat commit 45096d2
Translate next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Translate this skillshapeshift/web | 206 | — | ~4.5k | Automated safety check: Pass | MIT | |
| Translation Diff ExportDevolutions/UniGetUI | 26k | — | ~1.1k | Automated safety check: Pass | MIT | |
| Sync Translationssymfony/symfony | 31k | — | ~1.9k | Automated safety check: Pass | MIT | |
| Translation Diff ImportDevolutions/UniGetUI | 26k | — | ~750 | Automated safety check: Pass | MIT | |
| Translation Diff TranslateDevolutions/UniGetUI | 26k | — | ~934 | Automated safety check: Pass | MIT | |
| Generate Translationspayloadcms/payload | 45k | — | ~1.1k | Automated safety check: Pass | MIT |
Devolutions/UniGetUI
Compares UniGetUI JSON locale files against English, identifies untranslated or source-changed keys, and generates patch, reference, and handoff files for a target language.
symfony/symfony
Synchronize translation catalogs across maintained Symfony branches: find messages that newer branches added to the English catalogs but that are still missing from the oldest maintained branch…
Devolutions/UniGetUI
Merges translated key-value pairs from a UniGetUI JSON localization patch back into the full language file and validates the merged result.
Devolutions/UniGetUI
Translates a sparse UniGetUI JSON language patch, writes completed entries into the working copy, preserves placeholders and terminology, and prepares the patch for merge-back.
payloadcms/payload
A skill your agent uses when new translation keys are added to packages to generate new translations strings
Narcooo/inkos
Drives long-form fiction, scripts, storyboards, interactive films and long-document translation through InkOS, with every change made by a typed action.
shapeshift/web
Comprehensive React and Next.js performance optimization guide with 40+ rules for eliminating waterfalls, optimizing bundles, and improving rendering.
shapeshift/web
Run a quality benchmark of the /translate skill by selecting stratified test keys, capturing ground truth, translating, judging with sub-agents, and compiling a regression report.
shapeshift/web
Create a new qabot E2E test fixture interactively. An agent skill from shapeshift/web.
shapeshift/web
Integrate a new blockchain as a second-class citizen in ShapeShift Web.
shapeshift/web
Run QA tests using agent-browser and post results to the qabot dashboard.
shapeshift/web
Integrate new DEX aggregators, swappers, or bridge protocols (like Bebop, Portals, Jupiter, 0x, 1inch, etc.) into ShapeShift Web.
Categories
Translate new/changed English UI strings into all supported languages using a translate-review-refine pipeline. Translate is an agent skill from shapeshift/web. Translate new/changed English UI strings into all supported languages using a translate-review-refine pipeline.
Translate fits situations like: tasks that involve Translation.
Run `npx skills add shapeshift/web --skill translate -a claude-code`. Or copy the skill folder (.claude/skills/translate in shapeshift/web) into .claude/skills/translate in your project. Claude Code loads it when a task matches its description.
Run `npx skills add shapeshift/web --skill translate -a codex`. Or copy the skill folder (.claude/skills/translate in shapeshift/web) into .agents/skills/translate in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add shapeshift/web --skill translate -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/translate, .gemini/skills/translate, .github/skills/translate and .opencode/skills/translate in your project.
Going by SKILL.md and its folder, Translate needs JavaScript for the scripts in its folder and the command-line tools its instructions call (node and git). Our summary lists: Node.js. Its frontmatter pre-approves these tools: Read, Write, Edit, Grep, Glob, Bash(node *), Bash(git diff*), Bash(git log*), Bash(git rev-parse*), Bash(wc *), Bash(mkdir *), Task, AskUserQuestion.
SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
Translate is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 4.5k tokens (SKILL.md is roughly 18k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Translate: Translation Diff Export (Devolutions/UniGetUI, 26k stars), Sync Translations (symfony/symfony, 31k stars), Translation Diff Import (Devolutions/UniGetUI, 26k stars) and Translation Diff Translate (Devolutions/UniGetUI, 26k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
shapeshift (a GitHub organization) maintains it in shapeshift/web, which has 206 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on October 7, 2026.
Source: shapeshift/web on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.