1---
2name: baoyu-translate
3description: >-
4  This skill should be used when the user asks to "translate", "翻译", "精翻", "translate article",
5  "translate to Chinese", "translate to English", "改成中文", "改成英文", "convert to Chinese",
6  "localize", "本地化", "refined translation", "精细翻译", "proofread translation", "快速翻译", "快翻",
7  "这篇文章翻译一下", or provides a URL/file with translation intent. Supports three modes
8  (quick/normal/refined) with custom glossary support.
9version: 1.117.3
10metadata:
11  openclaw:
12    homepage: https://github.com/JimLiu/baoyu-skills#baoyu-translate
13    requires:
14      anyBins:
15        - bun
16        - npx
17---
18 
19# Translator
20 
21Three-mode translation skill: **quick** for direct translation, **normal** for analysis-informed translation, **refined** for full publication-quality workflow with review and polish.
22 
23## User Input Tools
24 
25When this skill prompts the user, follow this tool-selection rule (priority order):
26 
271. **Prefer built-in user-input tools** exposed by the current agent runtime — e.g., `AskUserQuestion`, `request_user_input`, `clarify`, `ask_user`, or any equivalent.
282. **Fallback**: if no such tool exists, emit a numbered plain-text message and ask the user to reply with the chosen number/answer for each question.
293. **Batching**: if the tool supports multiple questions per call, combine all applicable questions into a single call; if only single-question, ask them one at a time in priority order.
30 
31Concrete `AskUserQuestion` references below are examples — substitute the local equivalent in other runtimes.
32 
33## Script Directory
34 
35Scripts in `scripts/` subdirectory. `{baseDir}` = this SKILL.md's directory path. Resolve `${BUN_X}` runtime: if `bun` installed → `bun`; if `npx` available → `npx -y bun`; else suggest installing bun. Replace `{baseDir}` and `${BUN_X}` with actual values.
36 
37| Script | Purpose |
38|--------|---------|
39| `scripts/main.ts` | CLI entry point. Default action splits markdown into chunks; also supports explicit `chunk` subcommand |
40| `scripts/chunk.ts` | Markdown chunking implementation used by `main.ts` and kept compatible for direct invocation |
41 
42## Preferences (EXTEND.md)
43 
44Check EXTEND.md in priority order — the first one found wins:
45 
46| Priority | Path | Scope |
47|----------|------|-------|
48| 1 | `.baoyu-skills/baoyu-translate/EXTEND.md` | Project |
49| 2 | `${XDG_CONFIG_HOME:-$HOME/.config}/baoyu-skills/baoyu-translate/EXTEND.md` | XDG |
50| 3 | `$HOME/.baoyu-skills/baoyu-translate/EXTEND.md` | User home |
51 
52| Result | Action |
53|--------|--------|
54| Found | Read, parse, apply. On first use in session, briefly remind: "Using preferences from [path]. You can edit EXTEND.md to customize glossary, audience, etc." |
55| Not found | **MUST** run first-time setup (see below) — do NOT silently use defaults |
56 
57**EXTEND.md supports**: default target language, default mode, target audience, custom glossaries (inline or file path), translation style, chunk settings.
58 
59Schema: [references/config/extend-schema.md](references/config/extend-schema.md).
60 
61### First-Time Setup (BLOCKING)
62 
63**CRITICAL**: When EXTEND.md is not found, you **MUST** run the first-time setup before ANY translation. This is a **BLOCKING** operation.
64 
65Full reference: [references/config/first-time-setup.md](references/config/first-time-setup.md)
66 
67Use `AskUserQuestion` with all questions (target language, mode, audience, style, save location) in ONE call. After user answers, create EXTEND.md at the chosen location, confirm "Preferences saved to [path]", then continue.
68 
69## Defaults
70 
71All configurable values in one place. EXTEND.md overrides these; CLI flags override EXTEND.md.
72 
73| Setting | Default | EXTEND.md key | CLI flag | Description |
74|---------|---------|---------------|----------|-------------|
75| Target language | `zh-CN` | `target_language` | `--to` | Translation target language |
76| Mode | `normal` | `default_mode` | `--mode` | Translation mode |
77| Audience | `general` | `audience` | `--audience` | Target reader profile |
78| Style | `storytelling` | `style` | `--style` | Translation style preference |
79| Chunk threshold | `4000` | `chunk_threshold` | — | Word count to trigger chunked translation |
80| Chunk max words | `5000` | `chunk_max_words` | — | Max words per chunk |
81 
82## Modes
83 
84| Mode | Flag | Steps | When to Use |
85|------|------|-------|-------------|
86| Quick | `--mode quick` | Translate | Short texts, informal content, quick tasks |
87| Normal | `--mode normal` (default) | Analyze → Translate | Articles, blog posts, general content |
88| Refined | `--mode refined` | Analyze → Translate → Review → Polish | Publication-quality, important documents |
89 
90**Default mode**: Normal (can be overridden in EXTEND.md `default_mode` setting).
91 
92**Style presets** — control the voice and tone of the translation (independent of audience):
93 
94| Value | Description | Effect |
95|-------|-------------|--------|
96| `storytelling` | Engaging narrative flow (default) | Draws readers in, smooth transitions, vivid phrasing |
97| `formal` | Professional, structured | Neutral tone, clear organization, no colloquialisms |
98| `technical` | Precise, documentation-style | Concise, terminology-heavy, minimal embellishment |
99| `literal` | Close to original structure | Minimal restructuring, preserves source sentence patterns |
100| `academic` | Scholarly, rigorous | Formal register, complex clauses OK, citation-aware |
101| `business` | Concise, results-focused | Action-oriented, executive-friendly, bullet-point mindset |
102| `humorous` | Preserves and adapts humor | Witty, playful, recreates comedic effect in target language |
103| `conversational` | Casual, spoken-like | Friendly, approachable, as if explaining to a friend |
104| `elegant` | Literary, polished prose | Aesthetically refined, rhythmic, carefully crafted word choices |
105 
106Custom style descriptions are also accepted, e.g., `--style "poetic and lyrical"`.
107 
108**Auto-detection**:
109- "快翻", "quick", "直接翻译" → quick mode
110- "精翻", "refined", "publication quality", "proofread" → refined mode
111- Otherwise → default mode (normal)
112 
113**Upgrade prompt**: After normal mode completes, display:
114> Translation saved. To further review and polish, reply "继续润色" or "refine".
115 
116If user responds, continue with review → polish steps (same as refined mode Steps 4-6 in refined-workflow.md) on the existing output.
117 
118**Audience presets**:
119 
120| Value | Description | Effect |
121|-------|-------------|--------|
122| `general` | General readers (default) | Plain language, more translator's notes for jargon |
123| `technical` | Developers / engineers | Less annotation on common tech terms |
124| `academic` | Researchers / scholars | Formal register, precise terminology |
125| `business` | Business professionals | Business-friendly tone, explain tech concepts |
126 
127Custom audience descriptions are also accepted, e.g., `--audience "AI感兴趣的普通读者"`.
128 
129## Workflow
130 
131### Step 1: Load Preferences
132 
1331.1 Check EXTEND.md (see Preferences section above)
134 
1351.2 Load built-in glossary for the language pair if available:
136- EN→ZH: [references/glossary-en-zh.md](references/glossary-en-zh.md)
137 
1381.3 Merge glossaries: EXTEND.md `glossary` (inline) + EXTEND.md `glossary_files` (external files, paths relative to EXTEND.md location) + built-in glossary + `--glossary` file (CLI overrides all)
139 
140### Step 2: Materialize Source & Create Output Directory
141 
142Materialize source (file as-is, inline text/URL → save to `translate/{slug}.md`), then create output directory: `{source-dir}/{source-basename}-{target-lang}/`. Detect source language if `--from` not specified.
143 
144Full details: [references/workflow-mechanics.md](references/workflow-mechanics.md)
145 
146**Output directory contents** (all intermediate and final files go here):
147 
148| File | Mode | Description |
149|------|------|-------------|
150| `translation.md` | All | Final translation (always this name) |
151| `01-analysis.md` | Normal, Refined | Content analysis (domain, tone, terminology) |
152| `02-prompt.md` | Normal, Refined | Assembled translation prompt |
153| `03-draft.md` | Refined | Initial draft before review |
154| `04-critique.md` | Refined | Critical review findings (diagnosis only) |
155| `05-revision.md` | Refined | Revised translation based on critique |
156| `chunks/` | Chunked | Source chunks + translated chunks |
157 
158### Step 3: Assess Content Length
159 
160Quick mode does not chunk — translate directly regardless of length. Before translating, estimate word count. If content exceeds chunk threshold (default 4000 words), proactively warn: "This article is ~{N} words. Quick mode translates in one pass without chunking — for long content, `--mode normal` produces better results with terminology consistency." Then proceed if user doesn't switch.
161 
162For normal and refined modes:
163 
164| Content | Action |
165|---------|--------|
166| < chunk threshold | Translate as single unit |
167| >= chunk threshold | Chunk translation (see Step 3.1) |
168 
169**3.1 Long Content Preparation** (normal/refined modes, >= chunk threshold only)
170 
171Before translating chunks:
172 
1731. **Extract terminology**: Scan entire document for proper nouns, technical terms, recurring phrases
1742. **Build session glossary**: Merge extracted terms with loaded glossaries, establish consistent translations
1753. **Split into chunks**: Use `${BUN_X} {baseDir}/scripts/main.ts <file> [--max-words <chunk_max_words>] [--output-dir <output-dir>]`
176   - Parses markdown blocks (headings, paragraphs, lists, code blocks, tables, etc.)
177   - Splits at markdown block boundaries to preserve structure
178   - If a single block exceeds the threshold, falls back to line splitting, then word splitting
1794. **Assemble translation prompt**:
180   - Main agent reads `01-analysis.md` (if exists) and assembles shared context using Part 1 of [references/subagent-prompt-template.md](references/subagent-prompt-template.md) — inlining: target style, content background, merged glossary, and translation challenges
181   - Save as `02-prompt.md` in the output directory (shared context only, no task instructions)
1825. **Draft translation via subagents** (if Agent tool available):
183   - Spawn one subagent **per chunk**, all in parallel (Part 2 of the template)
184   - Each subagent reads `02-prompt.md` for shared context, receives chunk position info (chunk N of M + brief context of where it sits in the argument), translates its chunk, saves to `chunks/chunk-NN-draft.md`
185   - Consistency is guaranteed by the shared `02-prompt.md` (glossary, figurative language mapping, comprehension challenges, source voice, and translation challenges from analysis)
186   - If no chunks (content under threshold): spawn one subagent for the entire source file
187   - If Agent tool is unavailable, translate chunks sequentially inline using `02-prompt.md`
1886. **Merge**: Once all subagents complete, combine translated chunks in order. If `chunks/frontmatter.md` exists, prepend it. Save as `03-draft.md` (refined) or `translation.md` (normal)
1897. All intermediate files (source chunks + translated chunks) are preserved in `chunks/`
190 
191**After chunked draft is merged**, return control to main agent for critical review, revision, and polish (Step 4).
192 
193### Step 4: Translate & Refine
194 
195**Translation principles** (apply to all modes):
196 
197- **Rewrite, not translate**: Rewrite content into natural, engaging target language as if a skilled native writer composed it from scratch. Quality test: "Does this read like it was originally written in the target language?"
198- **Accuracy first**: Facts, data, and logic must match the original exactly
199- **Natural flow**: Use idiomatic target language word order. Break long source sentences into shorter, natural ones. Interpret metaphors and idioms by intended meaning, not word-for-word
200- **Terminology**: Use standard translations consistently. First occurrence of specialized terms: annotate with original in parentheses
201- **Preserve format**: Keep all markdown formatting (headings, bold, italic, images, links, code blocks)
202- **Proactive interpretation**: For jargon or concepts the target audience may lack context for, add concise explanations in **bold parentheses** `（**解释**）`. Keep annotations few — only where genuinely needed for comprehension
203- **Frontmatter**: If source has YAML frontmatter, rename source-metadata fields with `source` prefix (camelCase: `url`→`sourceUrl`, `title`→`sourceTitle`, etc.), add translated values as new top-level fields (skip `title` if body has H1), keep other fields as-is
204 
205#### Quick Mode
206 
207Translate directly → save to `translation.md`. Apply all translation principles above.
208 
209#### Normal Mode
210 
2111. **Analyze** → `01-analysis.md` (domain, tone, terminology, translation challenges)
2122. **Assemble prompt** → `02-prompt.md` (translation instructions with context, glossary, challenges)
2133. **Translate** (following `02-prompt.md`) → `translation.md`
214 
215After completion, prompt user: "Translation saved. To further review and polish, reply **继续润色** or **refine**."
216 
217If user continues, proceed with critical review → revision → polish (same as refined mode Steps 4-6 below), saving `03-draft.md` (rename current `translation.md`), `04-critique.md`, `05-revision.md`, and updated `translation.md`.
218 
219#### Refined Mode
220 
221Full workflow for publication quality. See [references/refined-workflow.md](references/refined-workflow.md) for detailed guidelines per step.
222 
223The subagent (if used in Step 3.1) only handles the initial draft. All subsequent steps (critical review, revision, polish) are handled by the main agent, which may delegate to subagents at its discretion.
224 
225Steps and saved files (all in output directory):
2261. **Analyze** → `01-analysis.md` (domain, tone, terminology, translation challenges)
2272. **Assemble prompt** → `02-prompt.md` (translation instructions with inlined context)
2283. **Draft** → `03-draft.md` (initial translation with translator's notes; from subagent if chunked)
2294. **Critical review** → `04-critique.md` (diagnosis only: accuracy, Europeanized language, strategy execution, expression issues)
2305. **Revision** → `05-revision.md` (apply all critique findings to produce revised translation)
2316. **Polish** → `translation.md` (final publication-quality translation)
232 
233Each step reads the previous step's file and builds on it.
234 
235### Step 5: Output
236 
237Final translation is always at `translation.md` in the output directory.
238 
239After the final translation is written, do a lightweight image-language pass:
240 
2411. Collect image references from the translated article
2422. Identify likely text-heavy images such as covers, screenshots, diagrams, charts, frameworks, and infographics
2433. If any image likely contains a main text language that does not match the translated article language, proactively remind the user
2444. The reminder must be a list only. Do not automatically localize those images unless the user asks
245 
246Reminder format (use whatever image syntax the article already uses — standard markdown or wikilink):
247```text
248Possible image localization needed:
249- ![example cover](attachments/example-cover.png): likely still contains source-language text while the article is now in target language
250- ![example diagram](attachments/example-diagram.png): likely text-heavy framework graphic, check whether labels need translation
251```
252 
253Display summary:
254```
255**Translation complete** ({mode} mode)
256 
257Source: {source-path}
258Languages: {from} → {to}
259Output dir: {output-dir}/
260Final: {output-dir}/translation.md
261Glossary terms applied: {count}
262```
263 
264If mismatched image-language candidates were found, append a short note after the summary telling the user that some embedded images may still need image-text localization, followed by the candidate list.
265 
266## Extension Support
267 
268Custom configurations via EXTEND.md. See **Preferences** section for paths and supported options.
269