chore: release v1.87.2

Condense 15+ verbose translation principles to 7 concise ones.
Consolidate analysis from 8 sections to 4, review from 8 to 3.
Simplify subagent prompt template accordingly.

.claude-plugin/marketplace.json

@@ -6,7 +6,7 @@
   },
   "metadata": {
     "description": "Skills shared by Baoyu for improving daily work efficiency",
-    "version": "1.85.0"
+    "version": "1.87.2"
   },
   "plugins": [
     {
@@ -22,7 +22,6 @@
         "./skills/baoyu-danger-gemini-web",
         "./skills/baoyu-danger-x-to-markdown",
         "./skills/baoyu-format-markdown",
-        "./skills/baoyu-image-gen",
         "./skills/baoyu-imagine",
         "./skills/baoyu-infographic",
         "./skills/baoyu-markdown-to-html",

CHANGELOG.md

@@ -2,6 +2,27 @@
 
 English | [中文](./CHANGELOG.zh.md)
 
+## 1.87.2 - 2026-03-26
+
+### Refactor
+- `baoyu-translate`: simplify translation prompts from 15+ verbose principles to 7 concise ones, consolidate analysis and review steps in workflow references
+
+## 1.87.1 - 2026-03-26
+
+### Maintenance
+- Add deprecation notice to `baoyu-image-gen` SKILL.md redirecting users to `baoyu-imagine`
+- Document deprecated skills policy in CLAUDE.md
+
+## 1.87.0 - 2026-03-26
+
+### Maintenance
+- Remove deprecated `baoyu-image-gen` redirect skill and plugin manifest entry — migration to `baoyu-imagine` is complete
+
+## 1.86.0 - 2026-03-25
+
+### Features
+- `baoyu-translate`: enrich translation prompt with full analysis context — source voice assessment, structured figurative language mapping, comprehension challenge reasoning, structural/creative challenges, and chunk position context for subagents
+
 ## 1.85.0 - 2026-03-25
 
 ### Features

CHANGELOG.zh.md

@@ -2,6 +2,27 @@
 
 [English](./CHANGELOG.md) | 中文
 
+## 1.87.2 - 2026-03-26
+
+### 重构
+- `baoyu-translate`：精简翻译提示词，将 15+ 条冗长原则压缩为 7 条，合并分析和审校步骤
+
+## 1.87.1 - 2026-03-26
+
+### 维护
+- 在 `baoyu-image-gen` SKILL.md 中添加废弃提示，引导用户使用 `baoyu-imagine`
+- 在 CLAUDE.md 中记录废弃技能策略
+
+## 1.87.0 - 2026-03-26
+
+### 维护
+- 移除已废弃的 `baoyu-image-gen` 重定向技能及插件清单条目 — 向 `baoyu-imagine` 的迁移已完成
+
+## 1.86.0 - 2026-03-25
+
+### 新功能
+- `baoyu-translate`：丰富翻译提示词的分析上下文 — 加入原文语气评估、结构化比喻映射表、理解难点推理、结构性/创造性翻译挑战，以及分块翻译的位置上下文
+
 ## 1.85.0 - 2026-03-25
 
 ### 新功能

CLAUDE.md

@@ -1,6 +1,6 @@
 # CLAUDE.md
 
-Claude Code marketplace plugin providing AI-powered content generation skills. Version: **1.84.0**.
+Claude Code marketplace plugin providing AI-powered content generation skills. Version: **1.87.2**.
 
 ## Architecture
 
@@ -50,6 +50,12 @@ Execute: `${BUN_X} skills/<skill>/scripts/main.ts [options]`
 
 Priority: project `skills/` → `$HOME/.baoyu-skills/` → system-level.
 
+## Deprecated Skills
+
+| Skill | Note |
+|-------|------|
+| `baoyu-image-gen` | Migrated to `baoyu-imagine`. Do NOT add to `.claude-plugin/marketplace.json`. Do NOT update README for this skill. |
+
 ## Release Process
 
 Use `/release-skills` workflow. Never skip:

skills/baoyu-image-gen/SKILL.md

@@ -1,19 +1,408 @@
 ---
 name: baoyu-image-gen
-description: "[DEPRECATED] This skill has been renamed to baoyu-imagine. Triggers on any reference to 'baoyu-image-gen' so users can be guided to upgrade. Use this whenever you detect the old skill name in user prompts, EXTEND.md, or installed skill lists."
+description: AI image generation with OpenAI, Azure OpenAI, Google, OpenRouter, DashScope, MiniMax, Jimeng, Seedream and Replicate APIs. Supports text-to-image, reference images, aspect ratios, and batch generation from saved prompt files. Sequential by default; use batch parallel generation when the user already has multiple prompts or wants stable multi-image throughput. Use when user asks to generate, create, or draw images.
+version: 1.56.4
+metadata:
+  openclaw:
+    homepage: https://github.com/JimLiu/baoyu-skills#baoyu-image-gen
+    requires:
+      anyBins:
+        - bun
+        - npx
 ---
 
-# baoyu-image-gen is now baoyu-imagine
+> **⚠️ DEPRECATED**: This skill has been migrated to `baoyu-imagine`. Print a bold red warning to the user **in the user's language** (use ANSI escape `\033[1;31m`) stating that `baoyu-image-gen` has been migrated to `baoyu-imagine` and they should use `baoyu-imagine` instead. Then STOP — do NOT continue executing this skill.
 
-This skill has been renamed to **baoyu-imagine** with new features and providers.
+# Image Generation (AI SDK)
 
-## What to do
+Official API-based image generation. Supports OpenAI, Azure OpenAI, Google, OpenRouter, DashScope (阿里通义万象), MiniMax, Jimeng (即梦), Seedream (豆包) and Replicate providers.
 
-When this skill is triggered, inform the user that `baoyu-image-gen` has been renamed to `baoyu-imagine` with new features and providers, and ask for their confirmation to:
+## Script Directory
 
-1. **Install the new skill** from: `https://github.com/JimLiu/baoyu-skills/blob/main/skills/baoyu-imagine/SKILL.md`
-2. **Remove the old `baoyu-image-gen` skill**
+**Agent Execution**:
+1. `{baseDir}` = this SKILL.md file's directory
+2. Script path = `{baseDir}/scripts/main.ts`
+3. Resolve `${BUN_X}` runtime: if `bun` installed → `bun`; if `npx` available → `npx -y bun`; else suggest installing bun
 
-No config changes are needed — `baoyu-imagine` reads the same EXTEND.md settings (API keys, default provider, etc.).
+## Step 0: Load Preferences ⛔ BLOCKING
 
-Once the user confirms, help them complete the installation and removal using whatever mechanism the current environment supports. If the user also has an image generation request, proceed with `baoyu-imagine` after migration.
+**CRITICAL**: This step MUST complete BEFORE any image generation. Do NOT skip or defer.
+
+Check EXTEND.md existence (priority: project → user):
+
+```bash
+# macOS, Linux, WSL, Git Bash
+test -f .baoyu-skills/baoyu-image-gen/EXTEND.md && echo "project"
+test -f "${XDG_CONFIG_HOME:-$HOME/.config}/baoyu-skills/baoyu-image-gen/EXTEND.md" && echo "xdg"
+test -f "$HOME/.baoyu-skills/baoyu-image-gen/EXTEND.md" && echo "user"
+```
+
+```powershell
+# PowerShell (Windows)
+if (Test-Path .baoyu-skills/baoyu-image-gen/EXTEND.md) { "project" }
+$xdg = if ($env:XDG_CONFIG_HOME) { $env:XDG_CONFIG_HOME } else { "$HOME/.config" }
+if (Test-Path "$xdg/baoyu-skills/baoyu-image-gen/EXTEND.md") { "xdg" }
+if (Test-Path "$HOME/.baoyu-skills/baoyu-image-gen/EXTEND.md") { "user" }
+```
+
+| Result | Action |
+|--------|--------|
+| Found | Load, parse, apply settings. If `default_model.[provider]` is null → ask model only (Flow 2) |
+| Not found | ⛔ Run first-time setup ([references/config/first-time-setup.md](references/config/first-time-setup.md)) → Save EXTEND.md → Then continue |
+
+**CRITICAL**: If not found, complete the full setup (provider + model + quality + save location) using AskUserQuestion BEFORE generating any images. Generation is BLOCKED until EXTEND.md is created.
+
+| Path | Location |
+|------|----------|
+| `.baoyu-skills/baoyu-image-gen/EXTEND.md` | Project directory |
+| `$HOME/.baoyu-skills/baoyu-image-gen/EXTEND.md` | User home |
+
+**EXTEND.md Supports**: Default provider | Default quality | Default aspect ratio | Default image size | Default models | Batch worker cap | Provider-specific batch limits
+
+Schema: `references/config/preferences-schema.md`
+
+## Usage
+
+```bash
+# Basic
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image cat.png
+
+# With aspect ratio
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A landscape" --image out.png --ar 16:9
+
+# High quality
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --quality 2k
+
+# From prompt files
+${BUN_X} {baseDir}/scripts/main.ts --promptfiles system.md content.md --image out.png
+
+# With reference images (Google, OpenAI, Azure OpenAI, OpenRouter, Replicate, MiniMax, or Seedream 4.0/4.5/5.0)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "Make blue" --image out.png --ref source.png
+
+# With reference images (explicit provider/model)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "Make blue" --image out.png --provider google --model gemini-3-pro-image-preview --ref source.png
+
+# Azure OpenAI (model means deployment name)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider azure --model gpt-image-1.5
+
+# OpenRouter (recommended default model)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider openrouter
+
+# OpenRouter with reference images
+${BUN_X} {baseDir}/scripts/main.ts --prompt "Make blue" --image out.png --provider openrouter --model google/gemini-3.1-flash-image-preview --ref source.png
+
+# Specific provider
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider openai
+
+# DashScope (阿里通义万象)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "一只可爱的猫" --image out.png --provider dashscope
+
+# DashScope Qwen-Image 2.0 Pro (recommended for custom sizes and text rendering)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "为咖啡品牌设计一张 21:9 横幅海报，包含清晰中文标题" --image out.png --provider dashscope --model qwen-image-2.0-pro --size 2048x872
+
+# DashScope legacy Qwen fixed-size model
+${BUN_X} {baseDir}/scripts/main.ts --prompt "一张电影感海报" --image out.png --provider dashscope --model qwen-image-max --size 1664x928
+
+# MiniMax
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A fashion editorial portrait by a bright studio window" --image out.jpg --provider minimax
+
+# MiniMax with subject reference (best for character/portrait consistency)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A girl stands by the library window, cinematic lighting" --image out.jpg --provider minimax --model image-01 --ref portrait.png --ar 16:9
+
+# MiniMax with custom size (documented for image-01)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cinematic poster" --image out.jpg --provider minimax --model image-01 --size 1536x1024
+
+# Replicate (google/nano-banana-pro)
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider replicate
+
+# Replicate with specific model
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider replicate --model google/nano-banana
+
+# Batch mode with saved prompt files
+${BUN_X} {baseDir}/scripts/main.ts --batchfile batch.json
+
+# Batch mode with explicit worker count
+${BUN_X} {baseDir}/scripts/main.ts --batchfile batch.json --jobs 4 --json
+```
+
+### Batch File Format
+
+```json
+{
+  "jobs": 4,
+  "tasks": [
+    {
+      "id": "hero",
+      "promptFiles": ["prompts/hero.md"],
+      "image": "out/hero.png",
+      "provider": "replicate",
+      "model": "google/nano-banana-pro",
+      "ar": "16:9",
+      "quality": "2k"
+    },
+    {
+      "id": "diagram",
+      "promptFiles": ["prompts/diagram.md"],
+      "image": "out/diagram.png",
+      "ref": ["references/original.png"]
+    }
+  ]
+}
+```
+
+Paths in `promptFiles`, `image`, and `ref` are resolved relative to the batch file's directory. `jobs` is optional (overridden by CLI `--jobs`). Top-level array format (without `jobs` wrapper) is also accepted.
+
+## Options
+
+| Option | Description |
+|--------|-------------|
+| `--prompt <text>`, `-p` | Prompt text |
+| `--promptfiles <files...>` | Read prompt from files (concatenated) |
+| `--image <path>` | Output image path (required in single-image mode) |
+| `--batchfile <path>` | JSON batch file for multi-image generation |
+| `--jobs <count>` | Worker count for batch mode (default: auto, max from config, built-in default 10) |
+| `--provider google\|openai\|azure\|openrouter\|dashscope\|minimax\|jimeng\|seedream\|replicate` | Force provider (default: auto-detect) |
+| `--model <id>`, `-m` | Model ID (Google: `gemini-3-pro-image-preview`; OpenAI: `gpt-image-1.5`; Azure: deployment name such as `gpt-image-1.5` or `image-prod`; OpenRouter: `google/gemini-3.1-flash-image-preview`; DashScope: `qwen-image-2.0-pro`; MiniMax: `image-01`) |
+| `--ar <ratio>` | Aspect ratio (e.g., `16:9`, `1:1`, `4:3`) |
+| `--size <WxH>` | Size (e.g., `1024x1024`) |
+| `--quality normal\|2k` | Quality preset (default: `2k`) |
+| `--imageSize 1K\|2K\|4K` | Image size for Google/OpenRouter (default: from quality) |
+| `--ref <files...>` | Reference images. Supported by Google multimodal, OpenAI GPT Image edits, Azure OpenAI edits (PNG/JPG only), OpenRouter multimodal models, Replicate, MiniMax subject-reference, and Seedream 5.0/4.5/4.0. Not supported by Jimeng, Seedream 3.0, or removed SeedEdit 3.0 |
+| `--n <count>` | Number of images |
+| `--json` | JSON output |
+
+## Environment Variables
+
+| Variable | Description |
+|----------|-------------|
+| `OPENAI_API_KEY` | OpenAI API key |
+| `AZURE_OPENAI_API_KEY` | Azure OpenAI API key |
+| `OPENROUTER_API_KEY` | OpenRouter API key |
+| `GOOGLE_API_KEY` | Google API key |
+| `DASHSCOPE_API_KEY` | DashScope API key (阿里云) |
+| `MINIMAX_API_KEY` | MiniMax API key |
+| `REPLICATE_API_TOKEN` | Replicate API token |
+| `JIMENG_ACCESS_KEY_ID` | Jimeng (即梦) Volcengine access key |
+| `JIMENG_SECRET_ACCESS_KEY` | Jimeng (即梦) Volcengine secret key |
+| `ARK_API_KEY` | Seedream (豆包) Volcengine ARK API key |
+| `OPENAI_IMAGE_MODEL` | OpenAI model override |
+| `AZURE_OPENAI_DEPLOYMENT` | Azure default deployment name |
+| `AZURE_OPENAI_IMAGE_MODEL` | Backward-compatible alias for Azure default deployment/model name |
+| `OPENROUTER_IMAGE_MODEL` | OpenRouter model override (default: `google/gemini-3.1-flash-image-preview`) |
+| `GOOGLE_IMAGE_MODEL` | Google model override |
+| `DASHSCOPE_IMAGE_MODEL` | DashScope model override (default: `qwen-image-2.0-pro`) |
+| `MINIMAX_IMAGE_MODEL` | MiniMax model override (default: `image-01`) |
+| `REPLICATE_IMAGE_MODEL` | Replicate model override (default: google/nano-banana-pro) |
+| `JIMENG_IMAGE_MODEL` | Jimeng model override (default: jimeng_t2i_v40) |
+| `SEEDREAM_IMAGE_MODEL` | Seedream model override (default: doubao-seedream-5-0-260128) |
+| `OPENAI_BASE_URL` | Custom OpenAI endpoint |
+| `AZURE_OPENAI_BASE_URL` | Azure resource endpoint or deployment endpoint |
+| `AZURE_API_VERSION` | Azure image API version (default: `2025-04-01-preview`) |
+| `OPENROUTER_BASE_URL` | Custom OpenRouter endpoint (default: `https://openrouter.ai/api/v1`) |
+| `OPENROUTER_HTTP_REFERER` | Optional app/site URL for OpenRouter attribution |
+| `OPENROUTER_TITLE` | Optional app name for OpenRouter attribution |
+| `GOOGLE_BASE_URL` | Custom Google endpoint |
+| `DASHSCOPE_BASE_URL` | Custom DashScope endpoint |
+| `MINIMAX_BASE_URL` | Custom MiniMax endpoint (default: `https://api.minimax.io`) |
+| `REPLICATE_BASE_URL` | Custom Replicate endpoint |
+| `JIMENG_BASE_URL` | Custom Jimeng endpoint (default: `https://visual.volcengineapi.com`) |
+| `JIMENG_REGION` | Jimeng region (default: `cn-north-1`) |
+| `SEEDREAM_BASE_URL` | Custom Seedream endpoint (default: `https://ark.cn-beijing.volces.com/api/v3`) |
+| `BAOYU_IMAGE_GEN_MAX_WORKERS` | Override batch worker cap |
+| `BAOYU_IMAGE_GEN_<PROVIDER>_CONCURRENCY` | Override provider concurrency, e.g. `BAOYU_IMAGE_GEN_REPLICATE_CONCURRENCY` |
+| `BAOYU_IMAGE_GEN_<PROVIDER>_START_INTERVAL_MS` | Override provider start gap, e.g. `BAOYU_IMAGE_GEN_REPLICATE_START_INTERVAL_MS` |
+
+**Load Priority**: CLI args > EXTEND.md > env vars > `<cwd>/.baoyu-skills/.env` > `~/.baoyu-skills/.env`
+
+## Model Resolution
+
+Model priority (highest → lowest), applies to all providers:
+
+1. CLI flag: `--model <id>`
+2. EXTEND.md: `default_model.[provider]`
+3. Env var: `<PROVIDER>_IMAGE_MODEL` (e.g., `GOOGLE_IMAGE_MODEL`)
+4. Built-in default
+
+For Azure, `--model` / `default_model.azure` should be the Azure deployment name. `AZURE_OPENAI_DEPLOYMENT` is the preferred env var, and `AZURE_OPENAI_IMAGE_MODEL` remains as a backward-compatible alias.
+
+**EXTEND.md overrides env vars**. If both EXTEND.md `default_model.google: "gemini-3-pro-image-preview"` and env var `GOOGLE_IMAGE_MODEL=gemini-3.1-flash-image-preview` exist, EXTEND.md wins.
+
+**Agent MUST display model info** before each generation:
+- Show: `Using [provider] / [model]`
+- Show switch hint: `Switch model: --model <id> | EXTEND.md default_model.[provider] | env <PROVIDER>_IMAGE_MODEL`
+
+### DashScope Models
+
+Use `--model qwen-image-2.0-pro` or set `default_model.dashscope` / `DASHSCOPE_IMAGE_MODEL` when the user wants official Qwen-Image behavior.
+
+Official DashScope model families:
+
+- `qwen-image-2.0-pro`, `qwen-image-2.0-pro-2026-03-03`, `qwen-image-2.0`, `qwen-image-2.0-2026-03-03`
+  - Free-form `size` in `宽*高` format
+  - Total pixels must stay between `512*512` and `2048*2048`
+  - Default size is approximately `1024*1024`
+  - Best choice for custom ratios such as `21:9` and text-heavy Chinese/English layouts
+- `qwen-image-max`, `qwen-image-max-2025-12-30`, `qwen-image-plus`, `qwen-image-plus-2026-01-09`, `qwen-image`
+  - Fixed sizes only: `1664*928`, `1472*1104`, `1328*1328`, `1104*1472`, `928*1664`
+  - Default size is `1664*928`
+  - `qwen-image` currently has the same capability as `qwen-image-plus`
+- Legacy DashScope models such as `z-image-turbo`, `z-image-ultra`, `wanx-v1`
+  - Keep using them only when the user explicitly asks for legacy behavior or compatibility
+
+When translating CLI args into DashScope behavior:
+
+- `--size` wins over `--ar`
+- For `qwen-image-2.0*`, prefer explicit `--size`; otherwise infer from `--ar` and use the official recommended resolutions below
+- For `qwen-image-max/plus/image`, only use the five official fixed sizes; if the requested ratio is not covered, switch to `qwen-image-2.0-pro`
+- `--quality` is a baoyu-image-gen compatibility preset, not a native DashScope API field. Mapping `normal` / `2k` onto the `qwen-image-2.0*` table below is an implementation inference, not an official API guarantee
+
+Recommended `qwen-image-2.0*` sizes for common aspect ratios:
+
+| Ratio | `normal` | `2k` |
+|-------|----------|------|
+| `1:1` | `1024*1024` | `1536*1536` |
+| `2:3` | `768*1152` | `1024*1536` |
+| `3:2` | `1152*768` | `1536*1024` |
+| `3:4` | `960*1280` | `1080*1440` |
+| `4:3` | `1280*960` | `1440*1080` |
+| `9:16` | `720*1280` | `1080*1920` |
+| `16:9` | `1280*720` | `1920*1080` |
+| `21:9` | `1344*576` | `2048*872` |
+
+DashScope official APIs also expose `negative_prompt`, `prompt_extend`, and `watermark`, but `baoyu-image-gen` does not expose them as dedicated CLI flags today.
+
+Official references:
+
+- [Qwen-Image API](https://help.aliyun.com/zh/model-studio/qwen-image-api)
+- [Text-to-image guide](https://help.aliyun.com/zh/model-studio/text-to-image)
+- [Qwen-Image Edit API](https://help.aliyun.com/zh/model-studio/qwen-image-edit-api)
+
+### MiniMax Models
+
+Use `--model image-01` or set `default_model.minimax` / `MINIMAX_IMAGE_MODEL` when the user wants MiniMax image generation.
+
+Official MiniMax image model options currently documented in the API reference:
+
+- `image-01` (recommended default)
+  - Supports text-to-image and subject-reference image generation
+  - Supports official `aspect_ratio` values: `1:1`, `16:9`, `4:3`, `3:2`, `2:3`, `3:4`, `9:16`, `21:9`
+  - Supports documented custom `width` / `height` output sizes when using `--size <WxH>`
+  - `width` and `height` must both be between `512` and `2048`, and both must be divisible by `8`
+- `image-01-live`
+  - Lower-latency variant
+  - Use `--ar` for sizing; MiniMax documents custom `width` / `height` as only effective for `image-01`
+
+MiniMax subject reference notes:
+
+- `--ref` files are sent as MiniMax `subject_reference`
+- MiniMax docs currently describe `subject_reference[].type` as `character`
+- Official docs say `image_file` supports public URLs or Base64 Data URLs; `baoyu-image-gen` sends local refs as Data URLs
+- Official docs recommend front-facing portrait references in JPG/JPEG/PNG under 10MB
+
+Official references:
+
+- [MiniMax Image Generation Guide](https://platform.minimax.io/docs/guides/image-generation)
+- [MiniMax Text-to-Image API](https://platform.minimax.io/docs/api-reference/image-generation-t2i)
+- [MiniMax Image-to-Image API](https://platform.minimax.io/docs/api-reference/image-generation-i2i)
+
+### OpenRouter Models
+
+Use full OpenRouter model IDs, e.g.:
+
+- `google/gemini-3.1-flash-image-preview` (recommended, supports image output and reference-image workflows)
+- `google/gemini-2.5-flash-image-preview`
+- `black-forest-labs/flux.2-pro`
+- Other OpenRouter image-capable model IDs
+
+Notes:
+
+- OpenRouter image generation uses `/chat/completions`, not the OpenAI `/images` endpoints
+- If `--ref` is used, choose a multimodal model that supports image input and image output
+- `--imageSize` maps to OpenRouter `imageGenerationOptions.size`; `--size <WxH>` is converted to the nearest OpenRouter size and inferred aspect ratio when possible
+
+### Replicate Models
+
+Supported model formats:
+
+- `owner/name` (recommended for official models), e.g. `google/nano-banana-pro`
+- `owner/name:version` (community models by version), e.g. `stability-ai/sdxl:<version>`
+
+Examples:
+
+```bash
+# Use Replicate default model
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider replicate
+
+# Override model explicitly
+${BUN_X} {baseDir}/scripts/main.ts --prompt "A cat" --image out.png --provider replicate --model google/nano-banana
+```
+
+## Provider Selection
+
+1. `--ref` provided + no `--provider` → auto-select Google first, then OpenAI, then Azure, then OpenRouter, then Replicate, then Seedream, then MiniMax (MiniMax subject reference is more specialized toward character/portrait consistency)
+2. `--provider` specified → use it (if `--ref`, must be `google`, `openai`, `azure`, `openrouter`, `replicate`, `seedream`, or `minimax`)
+3. Only one API key available → use that provider
+4. Multiple available → default to Google
+
+## Quality Presets
+
+| Preset | Google imageSize | OpenAI Size | OpenRouter size | Replicate resolution | Use Case |
+|--------|------------------|-------------|-----------------|----------------------|----------|
+| `normal` | 1K | 1024px | 1K | 1K | Quick previews |
+| `2k` (default) | 2K | 2048px | 2K | 2K | Covers, illustrations, infographics |
+
+**Google/OpenRouter imageSize**: Can be overridden with `--imageSize 1K|2K|4K`
+
+## Aspect Ratios
+
+Supported: `1:1`, `16:9`, `9:16`, `4:3`, `3:4`, `2.35:1`
+
+- Google multimodal: uses `imageConfig.aspectRatio`
+- OpenAI: maps to closest supported size
+- OpenRouter: sends `imageGenerationOptions.aspect_ratio`; if only `--size <WxH>` is given, aspect ratio is inferred automatically
+- Replicate: passes `aspect_ratio` to model; when `--ref` is provided without `--ar`, defaults to `match_input_image`
+- MiniMax: sends official `aspect_ratio` values directly; if `--size <WxH>` is given without `--ar`, `width` / `height` are sent for `image-01`
+
+## Generation Mode
+
+**Default**: Sequential generation.
+
+**Batch Parallel Generation**: When `--batchfile` contains 2 or more pending tasks, the script automatically enables parallel generation.
+
+| Mode | When to Use |
+|------|-------------|
+| Sequential (default) | Normal usage, single images, small batches |
+| Parallel batch | Batch mode with 2+ tasks |
+
+Execution choice:
+
+| Situation | Preferred approach | Why |
+|-----------|--------------------|-----|
+| One image, or 1-2 simple images | Sequential | Lower coordination overhead and easier debugging |
+| Multiple images already have saved prompt files | Batch (`--batchfile`) | Reuses finalized prompts, applies shared throttling/retries, and gives predictable throughput |
+| Each image still needs separate reasoning, prompt writing, or style exploration | Subagents | The work is still exploratory, so each image may need independent analysis before generation |
+| Output comes from `baoyu-article-illustrator` with `outline.md` + `prompts/` | Batch (`build-batch.ts` -> `--batchfile`) | That workflow already produces prompt files, so direct batch execution is the intended path |
+
+Rule of thumb:
+
+- Prefer batch over subagents once prompt files are already saved and the task is "generate all of these"
+- Use subagents only when generation is coupled with per-image thinking, rewriting, or divergent creative exploration
+
+Parallel behavior:
+
+- Default worker count is automatic, capped by config, built-in default 10
+- Provider-specific throttling is applied only in batch mode, and the built-in defaults are tuned for faster throughput while still avoiding obvious RPM bursts
+- You can override worker count with `--jobs <count>`
+- Each image retries automatically up to 3 attempts
+- Final output includes success count, failure count, and per-image failure reasons
+
+## Error Handling
+
+- Missing API key → error with setup instructions
+- Generation failure → auto-retry up to 3 attempts per image
+- Invalid aspect ratio → warning, proceed with default
+- Reference images with unsupported provider/model → error with fix hint
+
+## Extension Support
+
+Custom configurations via EXTEND.md. See **Preferences** section for paths and supported options.

skills/baoyu-image-gen/references/config/first-time-setup.md

@@ -0,0 +1,316 @@
+---
+name: first-time-setup
+description: First-time setup and default model selection flow for baoyu-image-gen
+---
+
+# First-Time Setup
+
+## Overview
+
+Triggered when:
+1. No EXTEND.md found → full setup (provider + model + preferences)
+2. EXTEND.md found but `default_model.[provider]` is null → model selection only
+
+## Setup Flow
+
+```
+No EXTEND.md found          EXTEND.md found, model null
+        │                            │
+        ▼                            ▼
+┌─────────────────────┐    ┌──────────────────────┐
+│ AskUserQuestion     │    │ AskUserQuestion      │
+│ (full setup)        │    │ (model only)         │
+└─────────────────────┘    └──────────────────────┘
+        │                            │
+        ▼                            ▼
+┌─────────────────────┐    ┌──────────────────────┐
+│ Create EXTEND.md    │    │ Update EXTEND.md     │
+└─────────────────────┘    └──────────────────────┘
+        │                            │
+        ▼                            ▼
+    Continue                     Continue
+```
+
+## Flow 1: No EXTEND.md (Full Setup)
+
+**Language**: Use user's input language or saved language preference.
+
+Use AskUserQuestion with ALL questions in ONE call:
+
+### Question 1: Default Provider
+
+```yaml
+header: "Provider"
+question: "Default image generation provider?"
+options:
+  - label: "Google (Recommended)"
+    description: "Gemini multimodal - high quality, reference images, flexible sizes"
+  - label: "OpenAI"
+    description: "GPT Image - consistent quality, reliable output"
+  - label: "Azure OpenAI"
+    description: "Azure-hosted GPT Image deployments with resource-specific routing"
+  - label: "OpenRouter"
+    description: "Router for Gemini/FLUX/OpenAI-compatible image models"
+  - label: "DashScope"
+    description: "Alibaba Cloud - Qwen-Image, strong Chinese/English text rendering"
+  - label: "MiniMax"
+    description: "MiniMax image generation with subject-reference character workflows"
+  - label: "Replicate"
+    description: "Community models - nano-banana-pro, flexible model selection"
+```
+
+### Question 2: Default Google Model
+
+Only show if user selected Google or auto-detect (no explicit provider).
+
+```yaml
+header: "Google Model"
+question: "Default Google image generation model?"
+options:
+  - label: "gemini-3-pro-image-preview (Recommended)"
+    description: "Highest quality, best for production use"
+  - label: "gemini-3.1-flash-image-preview"
+    description: "Fast generation, good quality, lower cost"
+  - label: "gemini-3-flash-preview"
+    description: "Fast generation, balanced quality and speed"
+```
+
+### Question 2b: Default OpenRouter Model
+
+Only show if user selected OpenRouter.
+
+```yaml
+header: "OpenRouter Model"
+question: "Default OpenRouter image generation model?"
+options:
+  - label: "google/gemini-3.1-flash-image-preview (Recommended)"
+    description: "Best general-purpose OpenRouter image model with reference-image workflows"
+  - label: "google/gemini-2.5-flash-image-preview"
+    description: "Fast Gemini preview model on OpenRouter"
+  - label: "black-forest-labs/flux.2-pro"
+    description: "Strong text-to-image quality through OpenRouter"
+```
+
+### Question 2c: Default Azure Deployment
+
+Only show if user selected Azure OpenAI.
+
+```yaml
+header: "Azure Deploy"
+question: "Default Azure image deployment name?"
+options:
+  - label: "gpt-image-1.5 (Recommended)"
+    description: "Best default if your Azure deployment uses the same name"
+  - label: "gpt-image-1"
+    description: "Previous GPT Image deployment name"
+```
+
+### Question 2d: Default MiniMax Model
+
+Only show if user selected MiniMax.
+
+```yaml
+header: "MiniMax Model"
+question: "Default MiniMax image generation model?"
+options:
+  - label: "image-01 (Recommended)"
+    description: "Best default, supports aspect ratios and custom width/height"
+  - label: "image-01-live"
+    description: "Faster variant, use aspect ratio instead of custom size"
+```
+
+### Question 3: Default Quality
+
+```yaml
+header: "Quality"
+question: "Default image quality?"
+options:
+  - label: "2k (Recommended)"
+    description: "2048px - covers, illustrations, infographics"
+  - label: "normal"
+    description: "1024px - quick previews, drafts"
+```
+
+### Question 4: Save Location
+
+```yaml
+header: "Save"
+question: "Where to save preferences?"
+options:
+  - label: "Project (Recommended)"
+    description: ".baoyu-skills/ (this project only)"
+  - label: "User"
+    description: "~/.baoyu-skills/ (all projects)"
+```
+
+### Save Locations
+
+| Choice | Path | Scope |
+|--------|------|-------|
+| Project | `.baoyu-skills/baoyu-image-gen/EXTEND.md` | Current project |
+| User | `$HOME/.baoyu-skills/baoyu-image-gen/EXTEND.md` | All projects |
+
+### EXTEND.md Template
+
+```yaml
+---
+version: 1
+default_provider: [selected provider or null]
+default_quality: [selected quality]
+default_aspect_ratio: null
+default_image_size: null
+default_model:
+  google: [selected google model or null]
+  openai: null
+  azure: [selected azure deployment or null]
+  openrouter: [selected openrouter model or null]
+  dashscope: null
+  minimax: [selected minimax model or null]
+  replicate: null
+---
+```
+
+## Flow 2: EXTEND.md Exists, Model Null
+
+When EXTEND.md exists but `default_model.[current_provider]` is null, ask ONLY the model question for the current provider.
+
+### Google Model Selection
+
+```yaml
+header: "Google Model"
+question: "Choose a default Google image generation model?"
+options:
+  - label: "gemini-3-pro-image-preview (Recommended)"
+    description: "Highest quality, best for production use"
+  - label: "gemini-3.1-flash-image-preview"
+    description: "Fast generation, good quality, lower cost"
+  - label: "gemini-3-flash-preview"
+    description: "Fast generation, balanced quality and speed"
+```
+
+### OpenAI Model Selection
+
+```yaml
+header: "OpenAI Model"
+question: "Choose a default OpenAI image generation model?"
+options:
+  - label: "gpt-image-1.5 (Recommended)"
+    description: "Latest GPT Image model, high quality"
+  - label: "gpt-image-1"
+    description: "Previous generation GPT Image model"
+```
+
+### Azure Deployment Selection
+
+```yaml
+header: "Azure Deploy"
+question: "Choose a default Azure image deployment name?"
+options:
+  - label: "gpt-image-1.5 (Recommended)"
+    description: "Use when your Azure deployment name matches the GPT-image-1.5 model"
+  - label: "gpt-image-1"
+    description: "Use when your Azure deployment name matches GPT-image-1"
+```
+
+Notes for Azure setup:
+
+- In `baoyu-image-gen`, Azure `--model` / `default_model.azure` should be the Azure deployment name, not just the underlying model family.
+- If the deployment name is custom, save that exact deployment name in `default_model.azure`.
+
+### OpenRouter Model Selection
+
+```yaml
+header: "OpenRouter Model"
+question: "Choose a default OpenRouter image generation model?"
+options:
+  - label: "google/gemini-3.1-flash-image-preview (Recommended)"
+    description: "Recommended for image output and reference-image edits"
+  - label: "google/gemini-2.5-flash-image-preview"
+    description: "Fast preview-oriented image generation"
+  - label: "black-forest-labs/flux.2-pro"
+    description: "High-quality text-to-image through OpenRouter"
+```
+
+### DashScope Model Selection
+
+```yaml
+header: "DashScope Model"
+question: "Choose a default DashScope image generation model?"
+options:
+  - label: "qwen-image-2.0-pro (Recommended)"
+    description: "Best DashScope model for text rendering and custom sizes"
+  - label: "qwen-image-2.0"
+    description: "Faster 2.0 variant with flexible output size"
+  - label: "qwen-image-max"
+    description: "Legacy Qwen model with five fixed output sizes"
+  - label: "qwen-image-plus"
+    description: "Legacy Qwen model, same current capability as qwen-image"
+  - label: "z-image-turbo"
+    description: "Legacy DashScope model for compatibility"
+  - label: "z-image-ultra"
+    description: "Legacy DashScope model, higher quality but slower"
+```
+
+Notes for DashScope setup:
+
+- Prefer `qwen-image-2.0-pro` when the user needs custom `--size`, uncommon ratios like `21:9`, or strong Chinese/English text rendering.
+- `qwen-image-max` / `qwen-image-plus` / `qwen-image` only support five fixed sizes: `1664*928`, `1472*1104`, `1328*1328`, `1104*1472`, `928*1664`.
+- In `baoyu-image-gen`, `quality` is a compatibility preset. It is not a native DashScope parameter.
+
+### Replicate Model Selection
+
+```yaml
+header: "Replicate Model"
+question: "Choose a default Replicate image generation model?"
+options:
+  - label: "google/nano-banana-pro (Recommended)"
+    description: "Google's fast image model on Replicate"
+  - label: "google/nano-banana"
+    description: "Google's base image model on Replicate"
+```
+
+### MiniMax Model Selection
+
+```yaml
+header: "MiniMax Model"
+question: "Choose a default MiniMax image generation model?"
+options:
+  - label: "image-01 (Recommended)"
+    description: "Best general-purpose MiniMax image model with custom width/height support"
+  - label: "image-01-live"
+    description: "Lower-latency MiniMax image model using aspect ratios"
+```
+
+Notes for MiniMax setup:
+
+- `image-01` is the safest default. It supports official `aspect_ratio` values and documented custom `width` / `height` output sizes.
+- `image-01-live` is useful when the user prefers faster generation and can work with aspect-ratio-based sizing.
+- MiniMax subject reference currently uses `subject_reference[].type = character`; docs recommend front-facing portrait references in JPG/JPEG/PNG under 10MB.
+
+### Update EXTEND.md
+
+After user selects a model:
+
+1. Read existing EXTEND.md
+2. If `default_model:` section exists → update the provider-specific key
+3. If `default_model:` section missing → add the full section:
+
+```yaml
+default_model:
+  google: [value or null]
+  openai: [value or null]
+  azure: [value or null]
+  openrouter: [value or null]
+  dashscope: [value or null]
+  minimax: [value or null]
+  replicate: [value or null]
+```
+
+Only set the selected provider's model; leave others as their current value or null.
+
+## After Setup
+
+1. Create directory if needed
+2. Write/update EXTEND.md with frontmatter
+3. Confirm: "Preferences saved to [path]"
+4. Continue with image generation

skills/baoyu-image-gen/references/config/preferences-schema.md

@@ -0,0 +1,121 @@
+---
+name: preferences-schema
+description: EXTEND.md YAML schema for baoyu-image-gen user preferences
+---
+
+# Preferences Schema
+
+## Full Schema
+
+```yaml
+---
+version: 1
+
+default_provider: null      # google|openai|azure|openrouter|dashscope|minimax|replicate|null (null = auto-detect)
+
+default_quality: null       # normal|2k|null (null = use default: 2k)
+
+default_aspect_ratio: null  # "16:9"|"1:1"|"4:3"|"3:4"|"2.35:1"|null
+
+default_image_size: null    # 1K|2K|4K|null (Google/OpenRouter, overrides quality)
+
+default_model:
+  google: null              # e.g., "gemini-3-pro-image-preview", "gemini-3.1-flash-image-preview"
+  openai: null              # e.g., "gpt-image-1.5", "gpt-image-1"
+  azure: null               # Azure deployment name, e.g., "gpt-image-1.5" or "image-prod"
+  openrouter: null          # e.g., "google/gemini-3.1-flash-image-preview"
+  dashscope: null           # e.g., "qwen-image-2.0-pro"
+  minimax: null             # e.g., "image-01"
+  replicate: null           # e.g., "google/nano-banana-pro"
+
+batch:
+  max_workers: 10
+  provider_limits:
+    replicate:
+      concurrency: 5
+      start_interval_ms: 700
+    google:
+      concurrency: 3
+      start_interval_ms: 1100
+    openai:
+      concurrency: 3
+      start_interval_ms: 1100
+    azure:
+      concurrency: 3
+      start_interval_ms: 1100
+    openrouter:
+      concurrency: 3
+      start_interval_ms: 1100
+    dashscope:
+      concurrency: 3
+      start_interval_ms: 1100
+    minimax:
+      concurrency: 3
+      start_interval_ms: 1100
+---
+```
+
+## Field Reference
+
+| Field | Type | Default | Description |
+|-------|------|---------|-------------|
+| `version` | int | 1 | Schema version |
+| `default_provider` | string\|null | null | Default provider (null = auto-detect) |
+| `default_quality` | string\|null | null | Default quality (null = 2k) |
+| `default_aspect_ratio` | string\|null | null | Default aspect ratio |
+| `default_image_size` | string\|null | null | Google/OpenRouter image size (overrides quality) |
+| `default_model.google` | string\|null | null | Google default model |
+| `default_model.openai` | string\|null | null | OpenAI default model |
+| `default_model.azure` | string\|null | null | Azure default deployment name |
+| `default_model.openrouter` | string\|null | null | OpenRouter default model |
+| `default_model.dashscope` | string\|null | null | DashScope default model |
+| `default_model.minimax` | string\|null | null | MiniMax default model |
+| `default_model.replicate` | string\|null | null | Replicate default model |
+| `batch.max_workers` | int\|null | 10 | Batch worker cap |
+| `batch.provider_limits.<provider>.concurrency` | int\|null | provider default | Max simultaneous requests per provider |
+| `batch.provider_limits.<provider>.start_interval_ms` | int\|null | provider default | Minimum gap between request starts per provider |
+
+## Examples
+
+**Minimal**:
+```yaml
+---
+version: 1
+default_provider: google
+default_quality: 2k
+---
+```
+
+**Full**:
+```yaml
+---
+version: 1
+default_provider: google
+default_quality: 2k
+default_aspect_ratio: "16:9"
+default_image_size: 2K
+default_model:
+  google: "gemini-3-pro-image-preview"
+  openai: "gpt-image-1.5"
+  azure: "gpt-image-1.5"
+  openrouter: "google/gemini-3.1-flash-image-preview"
+  dashscope: "qwen-image-2.0-pro"
+  minimax: "image-01"
+  replicate: "google/nano-banana-pro"
+batch:
+  max_workers: 10
+  provider_limits:
+    replicate:
+      concurrency: 5
+      start_interval_ms: 700
+    azure:
+      concurrency: 3
+      start_interval_ms: 1100
+    openrouter:
+      concurrency: 3
+      start_interval_ms: 1100
+    minimax:
+      concurrency: 3
+      start_interval_ms: 1100
+---
+```

skills/baoyu-image-gen/scripts/main.test.ts

@@ -0,0 +1,412 @@
+import assert from "node:assert/strict";
+import fs from "node:fs/promises";
+import os from "node:os";
+import path from "node:path";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs, ExtendConfig } from "./types.ts";
+import {
+  createTaskArgs,
+  detectProvider,
+  getConfiguredMaxWorkers,
+  getConfiguredProviderRateLimits,
+  getWorkerCount,
+  isRetryableGenerationError,
+  loadBatchTasks,
+  mergeConfig,
+  normalizeOutputImagePath,
+  parseArgs,
+  parseSimpleYaml,
+} from "./main.ts";
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+async function makeTempDir(prefix: string): Promise<string> {
+  return fs.mkdtemp(path.join(os.tmpdir(), prefix));
+}
+
+test("parseArgs parses the main image-gen CLI flags", () => {
+  const args = parseArgs([
+    "--promptfiles",
+    "prompts/system.md",
+    "prompts/content.md",
+    "--image",
+    "out/hero",
+    "--provider",
+    "openai",
+    "--quality",
+    "2k",
+    "--imageSize",
+    "4k",
+    "--ref",
+    "ref/one.png",
+    "ref/two.jpg",
+    "--n",
+    "3",
+    "--jobs",
+    "5",
+    "--json",
+  ]);
+
+  assert.deepEqual(args.promptFiles, ["prompts/system.md", "prompts/content.md"]);
+  assert.equal(args.imagePath, "out/hero");
+  assert.equal(args.provider, "openai");
+  assert.equal(args.quality, "2k");
+  assert.equal(args.imageSize, "4K");
+  assert.deepEqual(args.referenceImages, ["ref/one.png", "ref/two.jpg"]);
+  assert.equal(args.n, 3);
+  assert.equal(args.jobs, 5);
+  assert.equal(args.json, true);
+});
+
+test("parseArgs falls back to positional prompt and rejects invalid provider", () => {
+  const positional = parseArgs(["draw", "a", "cat"]);
+  assert.equal(positional.prompt, "draw a cat");
+
+  assert.throws(
+    () => parseArgs(["--provider", "stability"]),
+    /Invalid provider/,
+  );
+});
+
+test("parseSimpleYaml parses nested defaults and provider limits", () => {
+  const yaml = `
+version: 2
+default_provider: openrouter
+default_quality: normal
+default_aspect_ratio: '16:9'
+default_image_size: 2K
+default_model:
+  google: gemini-3-pro-image-preview
+  openai: gpt-image-1.5
+  azure: image-prod
+  minimax: image-01
+batch:
+  max_workers: 8
+  provider_limits:
+    google:
+      concurrency: 2
+      start_interval_ms: 900
+    openai:
+      concurrency: 4
+    minimax:
+      concurrency: 2
+      start_interval_ms: 1400
+    azure:
+      concurrency: 1
+      start_interval_ms: 1500
+`;
+
+  const config = parseSimpleYaml(yaml);
+
+  assert.equal(config.version, 2);
+  assert.equal(config.default_provider, "openrouter");
+  assert.equal(config.default_quality, "normal");
+  assert.equal(config.default_aspect_ratio, "16:9");
+  assert.equal(config.default_image_size, "2K");
+  assert.equal(config.default_model?.google, "gemini-3-pro-image-preview");
+  assert.equal(config.default_model?.openai, "gpt-image-1.5");
+  assert.equal(config.default_model?.azure, "image-prod");
+  assert.equal(config.default_model?.minimax, "image-01");
+  assert.equal(config.batch?.max_workers, 8);
+  assert.deepEqual(config.batch?.provider_limits?.google, {
+    concurrency: 2,
+    start_interval_ms: 900,
+  });
+  assert.deepEqual(config.batch?.provider_limits?.openai, {
+    concurrency: 4,
+  });
+  assert.deepEqual(config.batch?.provider_limits?.minimax, {
+    concurrency: 2,
+    start_interval_ms: 1400,
+  });
+  assert.deepEqual(config.batch?.provider_limits?.azure, {
+    concurrency: 1,
+    start_interval_ms: 1500,
+  });
+});
+
+test("mergeConfig only fills values missing from CLI args", () => {
+  const merged = mergeConfig(
+    makeArgs({
+      provider: "openai",
+      quality: null,
+      aspectRatio: null,
+      imageSize: "4K",
+    }),
+    {
+      default_provider: "google",
+      default_quality: "2k",
+      default_aspect_ratio: "3:2",
+      default_image_size: "2K",
+    } satisfies Partial<ExtendConfig>,
+  );
+
+  assert.equal(merged.provider, "openai");
+  assert.equal(merged.quality, "2k");
+  assert.equal(merged.aspectRatio, "3:2");
+  assert.equal(merged.imageSize, "4K");
+});
+
+test("detectProvider rejects non-ref-capable providers and prefers Google first when multiple keys exist", (t) => {
+  assert.throws(
+    () =>
+      detectProvider(
+        makeArgs({
+          provider: "dashscope",
+          referenceImages: ["ref.png"],
+        }),
+      ),
+    /Reference images require a ref-capable provider/,
+  );
+
+  useEnv(t, {
+    GOOGLE_API_KEY: "google-key",
+    OPENAI_API_KEY: "openai-key",
+    OPENROUTER_API_KEY: null,
+    DASHSCOPE_API_KEY: null,
+    MINIMAX_API_KEY: null,
+    REPLICATE_API_TOKEN: null,
+    JIMENG_ACCESS_KEY_ID: null,
+    JIMENG_SECRET_ACCESS_KEY: null,
+    ARK_API_KEY: null,
+  });
+  assert.equal(detectProvider(makeArgs()), "google");
+});
+
+test("detectProvider selects an available ref-capable provider for reference-image tasks", (t) => {
+  useEnv(t, {
+    GOOGLE_API_KEY: null,
+    OPENAI_API_KEY: "openai-key",
+    AZURE_OPENAI_API_KEY: null,
+    AZURE_OPENAI_BASE_URL: null,
+    OPENROUTER_API_KEY: null,
+    DASHSCOPE_API_KEY: null,
+    MINIMAX_API_KEY: null,
+    REPLICATE_API_TOKEN: null,
+    JIMENG_ACCESS_KEY_ID: null,
+    JIMENG_SECRET_ACCESS_KEY: null,
+    ARK_API_KEY: null,
+  });
+  assert.equal(
+    detectProvider(makeArgs({ referenceImages: ["ref.png"] })),
+    "openai",
+  );
+});
+
+test("detectProvider selects Azure when only Azure credentials are configured", (t) => {
+  useEnv(t, {
+    GOOGLE_API_KEY: null,
+    OPENAI_API_KEY: null,
+    AZURE_OPENAI_API_KEY: "azure-key",
+    AZURE_OPENAI_BASE_URL: "https://example.openai.azure.com",
+    OPENROUTER_API_KEY: null,
+    DASHSCOPE_API_KEY: null,
+    MINIMAX_API_KEY: null,
+    REPLICATE_API_TOKEN: null,
+    JIMENG_ACCESS_KEY_ID: null,
+    JIMENG_SECRET_ACCESS_KEY: null,
+    ARK_API_KEY: null,
+  });
+
+  assert.equal(detectProvider(makeArgs()), "azure");
+  assert.equal(
+    detectProvider(makeArgs({ referenceImages: ["ref.png"] })),
+    "azure",
+  );
+});
+
+test("detectProvider infers Seedream from model id and allows Seedream reference-image workflows", (t) => {
+  useEnv(t, {
+    GOOGLE_API_KEY: null,
+    OPENAI_API_KEY: null,
+    OPENROUTER_API_KEY: null,
+    DASHSCOPE_API_KEY: null,
+    MINIMAX_API_KEY: null,
+    REPLICATE_API_TOKEN: null,
+    JIMENG_ACCESS_KEY_ID: null,
+    JIMENG_SECRET_ACCESS_KEY: null,
+    ARK_API_KEY: "ark-key",
+  });
+
+  assert.equal(
+    detectProvider(
+      makeArgs({
+        model: "doubao-seedream-4-5-251128",
+        referenceImages: ["ref.png"],
+      }),
+    ),
+    "seedream",
+  );
+
+  assert.equal(
+    detectProvider(
+      makeArgs({
+        provider: "seedream",
+        referenceImages: ["ref.png"],
+      }),
+    ),
+    "seedream",
+  );
+});
+
+test("detectProvider selects MiniMax when only MiniMax credentials are configured or the model id matches", (t) => {
+  useEnv(t, {
+    GOOGLE_API_KEY: null,
+    OPENAI_API_KEY: null,
+    AZURE_OPENAI_API_KEY: null,
+    AZURE_OPENAI_BASE_URL: null,
+    OPENROUTER_API_KEY: null,
+    DASHSCOPE_API_KEY: null,
+    MINIMAX_API_KEY: "minimax-key",
+    REPLICATE_API_TOKEN: null,
+    JIMENG_ACCESS_KEY_ID: null,
+    JIMENG_SECRET_ACCESS_KEY: null,
+    ARK_API_KEY: null,
+  });
+
+  assert.equal(detectProvider(makeArgs()), "minimax");
+  assert.equal(detectProvider(makeArgs({ referenceImages: ["ref.png"] })), "minimax");
+  assert.equal(detectProvider(makeArgs({ model: "image-01-live" })), "minimax");
+});
+
+test("batch worker and provider-rate-limit configuration prefer env over EXTEND config", (t) => {
+  useEnv(t, {
+    BAOYU_IMAGE_GEN_MAX_WORKERS: "12",
+    BAOYU_IMAGE_GEN_GOOGLE_CONCURRENCY: "5",
+    BAOYU_IMAGE_GEN_GOOGLE_START_INTERVAL_MS: "450",
+  });
+
+  const extendConfig: Partial<ExtendConfig> = {
+    batch: {
+      max_workers: 7,
+      provider_limits: {
+        google: {
+          concurrency: 2,
+          start_interval_ms: 900,
+        },
+        minimax: {
+          concurrency: 1,
+          start_interval_ms: 1500,
+        },
+      },
+    },
+  };
+
+  assert.equal(getConfiguredMaxWorkers(extendConfig), 12);
+  assert.deepEqual(getConfiguredProviderRateLimits(extendConfig).google, {
+    concurrency: 5,
+    startIntervalMs: 450,
+  });
+  assert.deepEqual(getConfiguredProviderRateLimits(extendConfig).minimax, {
+    concurrency: 1,
+    startIntervalMs: 1500,
+  });
+});
+
+test("loadBatchTasks and createTaskArgs resolve batch-relative paths", async (t) => {
+  const root = await makeTempDir("baoyu-image-gen-batch-");
+  t.after(() => fs.rm(root, { recursive: true, force: true }));
+
+  const batchFile = path.join(root, "jobs", "batch.json");
+  await fs.mkdir(path.dirname(batchFile), { recursive: true });
+  await fs.writeFile(
+    batchFile,
+    JSON.stringify({
+      jobs: 2,
+      tasks: [
+        {
+          id: "hero",
+          promptFiles: ["prompts/hero.md"],
+          image: "out/hero",
+          ref: ["refs/hero.png"],
+          ar: "16:9",
+        },
+      ],
+    }),
+  );
+
+  const loaded = await loadBatchTasks(batchFile);
+  assert.equal(loaded.jobs, 2);
+  assert.equal(loaded.batchDir, path.dirname(batchFile));
+  assert.equal(loaded.tasks[0]?.id, "hero");
+
+  const taskArgs = createTaskArgs(
+    makeArgs({
+      provider: "replicate",
+      quality: "2k",
+      json: true,
+    }),
+    loaded.tasks[0]!,
+    loaded.batchDir,
+  );
+
+  assert.deepEqual(taskArgs.promptFiles, [
+    path.join(loaded.batchDir, "prompts/hero.md"),
+  ]);
+  assert.equal(taskArgs.imagePath, path.join(loaded.batchDir, "out/hero"));
+  assert.deepEqual(taskArgs.referenceImages, [
+    path.join(loaded.batchDir, "refs/hero.png"),
+  ]);
+  assert.equal(taskArgs.provider, "replicate");
+  assert.equal(taskArgs.aspectRatio, "16:9");
+  assert.equal(taskArgs.quality, "2k");
+  assert.equal(taskArgs.json, true);
+});
+
+test("path normalization, worker count, and retry classification follow expected rules", () => {
+  assert.match(normalizeOutputImagePath("out/sample"), /out[\\/]+sample\.png$/);
+  assert.match(normalizeOutputImagePath("out/sample", ".jpg"), /out[\\/]+sample\.jpg$/);
+  assert.match(normalizeOutputImagePath("out/sample.webp"), /out[\\/]+sample\.webp$/);
+
+  assert.equal(getWorkerCount(8, null, 3), 3);
+  assert.equal(getWorkerCount(2, 6, 5), 2);
+  assert.equal(getWorkerCount(5, 0, 4), 1);
+
+  assert.equal(isRetryableGenerationError(new Error("API error (401): denied")), false);
+  assert.equal(isRetryableGenerationError(new Error("socket hang up")), true);
+});

skills/baoyu-image-gen/scripts/providers/azure.test.ts

@@ -0,0 +1,188 @@
+import assert from "node:assert/strict";
+import fs from "node:fs/promises";
+import os from "node:os";
+import path from "node:path";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  generateImage,
+  getDefaultModel,
+  parseAzureBaseURL,
+  validateArgs,
+} from "./azure.ts";
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+async function makeTempDir(prefix: string): Promise<string> {
+  return fs.mkdtemp(path.join(os.tmpdir(), prefix));
+}
+
+test("Azure endpoint parsing and default deployment selection follow env precedence", (t) => {
+  assert.deepEqual(parseAzureBaseURL("https://example.openai.azure.com"), {
+    resourceBaseURL: "https://example.openai.azure.com/openai",
+    deployment: null,
+  });
+  assert.deepEqual(
+    parseAzureBaseURL("https://example.openai.azure.com/openai/deployments/from-url"),
+    {
+      resourceBaseURL: "https://example.openai.azure.com/openai",
+      deployment: "from-url",
+    },
+  );
+
+  useEnv(t, {
+    AZURE_OPENAI_BASE_URL: "https://example.openai.azure.com/openai/deployments/from-url",
+    AZURE_OPENAI_DEPLOYMENT: "explicit-deploy",
+    AZURE_OPENAI_IMAGE_MODEL: "env-fallback",
+  });
+  assert.equal(getDefaultModel(), "explicit-deploy");
+});
+
+test("Azure validateArgs rejects unsupported edit input formats before the API call", () => {
+  assert.doesNotThrow(() =>
+    validateArgs("demo-deployment", makeArgs({ referenceImages: ["hero.png", "photo.jpeg"] })),
+  );
+  assert.throws(
+    () => validateArgs("demo-deployment", makeArgs({ referenceImages: ["hero.webp"] })),
+    /PNG or JPG\/JPEG/,
+  );
+});
+
+test("Azure image generation routes model to deployment and sends mapped quality", async (t) => {
+  useEnv(t, {
+    AZURE_OPENAI_API_KEY: "azure-key",
+    AZURE_OPENAI_BASE_URL: "https://example.openai.azure.com/openai/deployments/default-deploy",
+    AZURE_API_VERSION: null,
+    AZURE_OPENAI_DEPLOYMENT: null,
+    AZURE_OPENAI_IMAGE_MODEL: null,
+  });
+
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const calls: Array<{ url: string; body: string }> = [];
+  globalThis.fetch = async (input, init) => {
+    calls.push({
+      url: String(input),
+      body: String(init?.body ?? ""),
+    });
+    return Response.json({
+      data: [{ b64_json: Buffer.from("azure-image").toString("base64") }],
+    });
+  };
+
+  const bytes = await generateImage(
+    "A calm lake at sunset",
+    "custom-deploy",
+    makeArgs({ quality: "normal" }),
+  );
+
+  assert.equal(Buffer.from(bytes).toString("utf8"), "azure-image");
+  assert.equal(
+    calls[0]?.url,
+    "https://example.openai.azure.com/openai/deployments/custom-deploy/images/generations?api-version=2025-04-01-preview",
+  );
+
+  const body = JSON.parse(calls[0]!.body) as Record<string, string>;
+  assert.equal(body.quality, "medium");
+  assert.equal(body.size, "1024x1024");
+});
+
+test("Azure image edits include quality in multipart requests", async (t) => {
+  const root = await makeTempDir("baoyu-image-gen-azure-");
+  t.after(() => fs.rm(root, { recursive: true, force: true }));
+
+  const pngPath = path.join(root, "ref.png");
+  const jpgPath = path.join(root, "ref.jpg");
+  await fs.writeFile(pngPath, "png-bytes");
+  await fs.writeFile(jpgPath, "jpg-bytes");
+
+  useEnv(t, {
+    AZURE_OPENAI_API_KEY: "azure-key",
+    AZURE_OPENAI_BASE_URL: "https://example.openai.azure.com",
+    AZURE_API_VERSION: "2025-04-01-preview",
+    AZURE_OPENAI_DEPLOYMENT: null,
+    AZURE_OPENAI_IMAGE_MODEL: null,
+  });
+
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const calls: Array<{ url: string; form: FormData }> = [];
+  globalThis.fetch = async (input, init) => {
+    calls.push({
+      url: String(input),
+      form: init?.body as FormData,
+    });
+    return Response.json({
+      data: [{ b64_json: Buffer.from("edited-image").toString("base64") }],
+    });
+  };
+
+  const bytes = await generateImage(
+    "Add warm lighting",
+    "edit-deploy",
+    makeArgs({
+      quality: "2k",
+      referenceImages: [pngPath, jpgPath],
+    }),
+  );
+
+  assert.equal(Buffer.from(bytes).toString("utf8"), "edited-image");
+  assert.equal(
+    calls[0]?.url,
+    "https://example.openai.azure.com/openai/deployments/edit-deploy/images/edits?api-version=2025-04-01-preview",
+  );
+  assert.equal(calls[0]?.form.get("quality"), "high");
+  assert.equal(calls[0]?.form.get("size"), "1024x1024");
+  assert.equal(calls[0]?.form.getAll("image[]").length, 2);
+});

skills/baoyu-image-gen/scripts/providers/azure.ts

@@ -0,0 +1,192 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+import type { CliArgs } from "../types";
+import { getOpenAISize, extractImageFromResponse } from "./openai.ts";
+
+type OpenAIImageResponse = { data: Array<{ url?: string; b64_json?: string }> };
+type AzureEndpoint = {
+  resourceBaseURL: string;
+  deployment: string | null;
+};
+
+const DEFAULT_AZURE_API_VERSION = "2025-04-01-preview";
+const AZURE_EDIT_IMAGE_EXTENSIONS = new Set([".png", ".jpg", ".jpeg"]);
+
+export function parseAzureBaseURL(url: string): AzureEndpoint {
+  const parsed = new URL(url);
+  const trimmedPath = parsed.pathname.replace(/\/+$/, "");
+  const deploymentMatch = trimmedPath.match(/^(.*?)(?:\/openai)?\/deployments\/([^/]+)$/);
+
+  if (deploymentMatch) {
+    parsed.pathname = `${deploymentMatch[1] || ""}/openai`;
+    return {
+      resourceBaseURL: parsed.toString().replace(/\/+$/, ""),
+      deployment: decodeURIComponent(deploymentMatch[2]!),
+    };
+  }
+
+  parsed.pathname = trimmedPath.endsWith("/openai") ? trimmedPath : `${trimmedPath}/openai`;
+  return {
+    resourceBaseURL: parsed.toString().replace(/\/+$/, ""),
+    deployment: null,
+  };
+}
+
+export function getDefaultModel(): string {
+  const explicitDeployment = process.env.AZURE_OPENAI_DEPLOYMENT?.trim();
+  if (explicitDeployment) return explicitDeployment;
+
+  const baseURL = process.env.AZURE_OPENAI_BASE_URL;
+  if (baseURL) {
+    try {
+      const { deployment } = parseAzureBaseURL(baseURL);
+      if (deployment) return deployment;
+    } catch {
+      // Ignore invalid URLs here so the required-env check can raise the user-facing error later.
+    }
+  }
+
+  return process.env.AZURE_OPENAI_IMAGE_MODEL || "gpt-image-1.5";
+}
+
+function getEndpoint(): AzureEndpoint {
+  const url = process.env.AZURE_OPENAI_BASE_URL;
+  if (!url) {
+    throw new Error(
+      "AZURE_OPENAI_BASE_URL is required. Set it to your Azure resource or deployment endpoint, e.g.: https://your-resource.openai.azure.com or https://your-resource.openai.azure.com/openai/deployments/your-deployment"
+    );
+  }
+  return parseAzureBaseURL(url);
+}
+
+function getApiKey(): string {
+  const key = process.env.AZURE_OPENAI_API_KEY;
+  if (!key) {
+    throw new Error(
+      "AZURE_OPENAI_API_KEY is required. Get it from Azure Portal → your OpenAI resource → Keys and Endpoint."
+    );
+  }
+  return key;
+}
+
+function getApiVersion(): string {
+  return process.env.AZURE_API_VERSION || DEFAULT_AZURE_API_VERSION;
+}
+
+function getDeployment(model: string): string {
+  const deployment = model.trim();
+  if (!deployment) {
+    throw new Error(
+      "Azure deployment name is required. Use --model <deployment>, AZURE_OPENAI_DEPLOYMENT, AZURE_OPENAI_IMAGE_MODEL, or embed the deployment in AZURE_OPENAI_BASE_URL."
+    );
+  }
+  return deployment;
+}
+
+function buildURL(deployment: string, pathSuffix: string): string {
+  const { resourceBaseURL } = getEndpoint();
+  return `${resourceBaseURL}/deployments/${encodeURIComponent(deployment)}${pathSuffix}?api-version=${getApiVersion()}`;
+}
+
+function authHeaders(): Record<string, string> {
+  return { "api-key": getApiKey() };
+}
+
+function getAzureQuality(quality: CliArgs["quality"]): "medium" | "high" {
+  return quality === "2k" ? "high" : "medium";
+}
+
+export function validateArgs(_model: string, args: CliArgs): void {
+  for (const refPath of args.referenceImages) {
+    const ext = path.extname(refPath).toLowerCase();
+    if (!AZURE_EDIT_IMAGE_EXTENSIONS.has(ext)) {
+      throw new Error(
+        `Azure OpenAI reference images must be PNG or JPG/JPEG. Unsupported file: ${refPath}`
+      );
+    }
+  }
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const deployment = getDeployment(model);
+  const size = args.size || getOpenAISize(model, args.aspectRatio, args.quality);
+
+  if (args.referenceImages.length > 0) {
+    return generateWithAzureEdits(prompt, deployment, size, args.referenceImages, args.quality);
+  }
+
+  return generateWithAzureGenerations(prompt, deployment, size, args.quality);
+}
+
+async function generateWithAzureGenerations(
+  prompt: string,
+  deployment: string,
+  size: string,
+  quality: CliArgs["quality"]
+): Promise<Uint8Array> {
+  const body: Record<string, any> = {
+    prompt,
+    size,
+    n: 1,
+    quality: getAzureQuality(quality),
+  };
+
+  const res = await fetch(buildURL(deployment, "/images/generations"), {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      ...authHeaders(),
+    },
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`Azure OpenAI API error: ${err}`);
+  }
+
+  const result = (await res.json()) as OpenAIImageResponse;
+  return extractImageFromResponse(result);
+}
+
+async function generateWithAzureEdits(
+  prompt: string,
+  deployment: string,
+  size: string,
+  referenceImages: string[],
+  quality: CliArgs["quality"]
+): Promise<Uint8Array> {
+  const form = new FormData();
+  form.append("prompt", prompt);
+  form.append("size", size);
+  form.append("n", "1");
+  form.append("quality", getAzureQuality(quality));
+
+  for (const refPath of referenceImages) {
+    const bytes = await readFile(refPath);
+    const filename = path.basename(refPath);
+    const mimeType = path.extname(filename).toLowerCase() === ".png" ? "image/png" : "image/jpeg";
+    const blob = new Blob([bytes], { type: mimeType });
+    form.append("image[]", blob, filename);
+  }
+
+  const res = await fetch(buildURL(deployment, "/images/edits"), {
+    method: "POST",
+    headers: {
+      ...authHeaders(),
+    },
+    body: form,
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`Azure OpenAI edits API error: ${err}`);
+  }
+
+  const result = (await res.json()) as OpenAIImageResponse;
+  return extractImageFromResponse(result);
+}

skills/baoyu-image-gen/scripts/providers/dashscope.test.ts

@@ -0,0 +1,148 @@
+import assert from "node:assert/strict";
+import test, { type TestContext } from "node:test";
+
+import {
+  getDefaultModel,
+  getModelFamily,
+  getQwen2SizeFromAspectRatio,
+  getSizeFromAspectRatio,
+  normalizeSize,
+  parseAspectRatio,
+  parseSize,
+  resolveSizeForModel,
+} from "./dashscope.ts";
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+test("DashScope default model prefers env override and otherwise uses qwen-image-2.0-pro", (t) => {
+  useEnv(t, { DASHSCOPE_IMAGE_MODEL: null });
+  assert.equal(getDefaultModel(), "qwen-image-2.0-pro");
+
+  process.env.DASHSCOPE_IMAGE_MODEL = "qwen-image-max";
+  assert.equal(getDefaultModel(), "qwen-image-max");
+});
+
+test("DashScope aspect-ratio parsing accepts numeric ratios only", () => {
+  assert.deepEqual(parseAspectRatio("3:2"), { width: 3, height: 2 });
+  assert.equal(parseAspectRatio("square"), null);
+  assert.equal(parseAspectRatio("-1:2"), null);
+});
+
+test("DashScope model family routing distinguishes qwen-2.0, fixed-size qwen, and legacy models", () => {
+  assert.equal(getModelFamily("qwen-image-2.0-pro"), "qwen2");
+  assert.equal(getModelFamily("qwen-image"), "qwenFixed");
+  assert.equal(getModelFamily("z-image-turbo"), "legacy");
+  assert.equal(getModelFamily("wanx-v1"), "legacy");
+});
+
+test("Legacy DashScope size selection keeps the previous quality-based heuristic", () => {
+  assert.equal(getSizeFromAspectRatio(null, "normal"), "1024*1024");
+  assert.equal(getSizeFromAspectRatio("16:9", "normal"), "1280*720");
+  assert.equal(getSizeFromAspectRatio("16:9", "2k"), "2048*1152");
+  assert.equal(getSizeFromAspectRatio("invalid", "2k"), "1536*1536");
+});
+
+test("Qwen 2.0 recommended sizes follow the official common-ratio table", () => {
+  assert.equal(getQwen2SizeFromAspectRatio(null, "normal"), "1024*1024");
+  assert.equal(getQwen2SizeFromAspectRatio(null, "2k"), "1536*1536");
+  assert.equal(getQwen2SizeFromAspectRatio("16:9", "normal"), "1280*720");
+  assert.equal(getQwen2SizeFromAspectRatio("21:9", "2k"), "2048*872");
+});
+
+test("Qwen 2.0 derives free-form sizes within pixel budget for uncommon ratios", () => {
+  const size = getQwen2SizeFromAspectRatio("5:2", "normal");
+  const parsed = parseSize(size);
+  assert.ok(parsed);
+  assert.ok(parsed.width * parsed.height >= 512 * 512);
+  assert.ok(parsed.width * parsed.height <= 2048 * 2048);
+  assert.ok(Math.abs(parsed.width / parsed.height - 2.5) < 0.08);
+});
+
+test("resolveSizeForModel validates explicit qwen-image-2.0 sizes by total pixels", () => {
+  assert.equal(
+    resolveSizeForModel("qwen-image-2.0-pro", {
+      size: "2048x872",
+      aspectRatio: null,
+      quality: "2k",
+    }),
+    "2048*872",
+  );
+
+  assert.throws(
+    () =>
+      resolveSizeForModel("qwen-image-2.0-pro", {
+        size: "4096x4096",
+        aspectRatio: null,
+        quality: "2k",
+      }),
+    /total pixels between/,
+  );
+});
+
+test("resolveSizeForModel enforces fixed sizes for qwen-image-max/plus/image", () => {
+  assert.equal(
+    resolveSizeForModel("qwen-image-max", {
+      size: null,
+      aspectRatio: "1:1",
+      quality: "2k",
+    }),
+    "1328*1328",
+  );
+
+  assert.equal(
+    resolveSizeForModel("qwen-image", {
+      size: "1664x928",
+      aspectRatio: "9:16",
+      quality: "normal",
+    }),
+    "1664*928",
+  );
+
+  assert.throws(
+    () =>
+      resolveSizeForModel("qwen-image-max", {
+        size: null,
+        aspectRatio: "21:9",
+        quality: "2k",
+      }),
+    /supports only fixed ratios/,
+  );
+
+  assert.throws(
+    () =>
+      resolveSizeForModel("qwen-image-plus", {
+        size: "1024x1024",
+        aspectRatio: null,
+        quality: "2k",
+      }),
+    /support only these sizes/,
+  );
+});
+
+test("DashScope size normalization converts WxH into provider format", () => {
+  assert.equal(normalizeSize("1024x1024"), "1024*1024");
+  assert.equal(normalizeSize("2048*1152"), "2048*1152");
+});

skills/baoyu-image-gen/scripts/providers/dashscope.ts

@@ -0,0 +1,463 @@
+import type { CliArgs, Quality } from "../types";
+
+type DashScopeModelFamily = "qwen2" | "qwenFixed" | "legacy";
+
+type DashScopeModelSpec = {
+  family: DashScopeModelFamily;
+  defaultSize: string;
+};
+
+const DEFAULT_MODEL = "qwen-image-2.0-pro";
+const MIN_QWEN_2_TOTAL_PIXELS = 512 * 512;
+const MAX_QWEN_2_TOTAL_PIXELS = 2048 * 2048;
+const SIZE_STEP = 16;
+const QWEN_NEGATIVE_PROMPT =
+  "低分辨率，低画质，肢体畸形，手指畸形，画面过饱和，蜡像感，人脸无细节，过度光滑，画面具有AI感，构图混乱，文字模糊，扭曲";
+
+const QWEN_2_TARGET_PIXELS: Record<Quality, number> = {
+  normal: 1024 * 1024,
+  "2k": 1536 * 1536,
+};
+
+const QWEN_2_RECOMMENDED: Record<string, Record<Quality, string>> = {
+  "1:1": { normal: "1024*1024", "2k": "1536*1536" },
+  "2:3": { normal: "768*1152", "2k": "1024*1536" },
+  "3:2": { normal: "1152*768", "2k": "1536*1024" },
+  "3:4": { normal: "960*1280", "2k": "1080*1440" },
+  "4:3": { normal: "1280*960", "2k": "1440*1080" },
+  "9:16": { normal: "720*1280", "2k": "1080*1920" },
+  "16:9": { normal: "1280*720", "2k": "1920*1080" },
+  "21:9": { normal: "1344*576", "2k": "2048*872" },
+};
+
+const QWEN_FIXED_SIZES_BY_RATIO: Record<string, string> = {
+  "16:9": "1664*928",
+  "4:3": "1472*1104",
+  "1:1": "1328*1328",
+  "3:4": "1104*1472",
+  "9:16": "928*1664",
+};
+
+const QWEN_FIXED_SIZES = Object.values(QWEN_FIXED_SIZES_BY_RATIO);
+
+const LEGACY_STANDARD_SIZES: [number, number][] = [
+  [1024, 1024],
+  [1280, 720],
+  [720, 1280],
+  [1024, 768],
+  [768, 1024],
+  [1536, 1024],
+  [1024, 1536],
+  [1536, 864],
+  [864, 1536],
+];
+
+const LEGACY_STANDARD_SIZES_2K: [number, number][] = [
+  [1536, 1536],
+  [2048, 1152],
+  [1152, 2048],
+  [1536, 1024],
+  [1024, 1536],
+  [1536, 864],
+  [864, 1536],
+  [2048, 2048],
+];
+
+const QWEN_2_SPEC: DashScopeModelSpec = {
+  family: "qwen2",
+  defaultSize: "1024*1024",
+};
+
+const QWEN_FIXED_SPEC: DashScopeModelSpec = {
+  family: "qwenFixed",
+  defaultSize: QWEN_FIXED_SIZES_BY_RATIO["16:9"],
+};
+
+const LEGACY_SPEC: DashScopeModelSpec = {
+  family: "legacy",
+  defaultSize: "1536*1536",
+};
+
+const MODEL_SPEC_ALIASES: Record<string, DashScopeModelSpec> = {
+  "qwen-image-2.0-pro": QWEN_2_SPEC,
+  "qwen-image-2.0-pro-2026-03-03": QWEN_2_SPEC,
+  "qwen-image-2.0": QWEN_2_SPEC,
+  "qwen-image-2.0-2026-03-03": QWEN_2_SPEC,
+  "qwen-image-max": QWEN_FIXED_SPEC,
+  "qwen-image-max-2025-12-30": QWEN_FIXED_SPEC,
+  "qwen-image-plus": QWEN_FIXED_SPEC,
+  "qwen-image-plus-2026-01-09": QWEN_FIXED_SPEC,
+  "qwen-image": QWEN_FIXED_SPEC,
+};
+
+export function getDefaultModel(): string {
+  return process.env.DASHSCOPE_IMAGE_MODEL || DEFAULT_MODEL;
+}
+
+function getApiKey(): string | null {
+  return process.env.DASHSCOPE_API_KEY || null;
+}
+
+function getBaseUrl(): string {
+  const base = process.env.DASHSCOPE_BASE_URL || "https://dashscope.aliyuncs.com";
+  return base.replace(/\/+$/g, "");
+}
+
+function getModelSpec(model: string): DashScopeModelSpec {
+  return MODEL_SPEC_ALIASES[model.trim().toLowerCase()] || LEGACY_SPEC;
+}
+
+export function getModelFamily(model: string): DashScopeModelFamily {
+  return getModelSpec(model).family;
+}
+
+function normalizeQuality(quality: CliArgs["quality"]): Quality {
+  return quality === "normal" ? "normal" : "2k";
+}
+
+export function parseAspectRatio(ar: string): { width: number; height: number } | null {
+  const match = ar.match(/^(\d+(?:\.\d+)?):(\d+(?:\.\d+)?)$/);
+  if (!match) return null;
+  const w = parseFloat(match[1]!);
+  const h = parseFloat(match[2]!);
+  if (w <= 0 || h <= 0) return null;
+  return { width: w, height: h };
+}
+
+export function normalizeSize(size: string): string {
+  return size.replace("x", "*");
+}
+
+export function parseSize(size: string): { width: number; height: number } | null {
+  const match = normalizeSize(size).match(/^(\d+)\*(\d+)$/);
+  if (!match) return null;
+  const width = Number(match[1]);
+  const height = Number(match[2]);
+  if (!Number.isFinite(width) || !Number.isFinite(height) || width <= 0 || height <= 0) {
+    return null;
+  }
+  return { width, height };
+}
+
+function formatSize(width: number, height: number): string {
+  return `${width}*${height}`;
+}
+
+function getRatioValue(ar: string): number | null {
+  const parsed = parseAspectRatio(ar);
+  if (!parsed) return null;
+  return parsed.width / parsed.height;
+}
+
+function findKnownRatioKey(ar: string, candidates: string[], tolerance = 0.02): string | null {
+  const targetRatio = getRatioValue(ar);
+  if (targetRatio == null) return null;
+
+  let bestKey: string | null = null;
+  let bestDiff = Infinity;
+
+  for (const candidate of candidates) {
+    const candidateRatio = getRatioValue(candidate);
+    if (candidateRatio == null) continue;
+    const diff = Math.abs(candidateRatio - targetRatio);
+    if (diff < bestDiff) {
+      bestDiff = diff;
+      bestKey = candidate;
+    }
+  }
+
+  return bestDiff <= tolerance ? bestKey : null;
+}
+
+function roundToStep(value: number): number {
+  return Math.max(SIZE_STEP, Math.round(value / SIZE_STEP) * SIZE_STEP);
+}
+
+function fitToPixelBudget(
+  width: number,
+  height: number,
+  minPixels: number,
+  maxPixels: number,
+): { width: number; height: number } {
+  let nextWidth = width;
+  let nextHeight = height;
+  let pixels = nextWidth * nextHeight;
+
+  if (pixels > maxPixels) {
+    const scale = Math.sqrt(maxPixels / pixels);
+    nextWidth *= scale;
+    nextHeight *= scale;
+  } else if (pixels < minPixels) {
+    const scale = Math.sqrt(minPixels / pixels);
+    nextWidth *= scale;
+    nextHeight *= scale;
+  }
+
+  let roundedWidth = roundToStep(nextWidth);
+  let roundedHeight = roundToStep(nextHeight);
+  pixels = roundedWidth * roundedHeight;
+
+  while (pixels > maxPixels && (roundedWidth > SIZE_STEP || roundedHeight > SIZE_STEP)) {
+    if (roundedWidth >= roundedHeight && roundedWidth > SIZE_STEP) {
+      roundedWidth -= SIZE_STEP;
+    } else if (roundedHeight > SIZE_STEP) {
+      roundedHeight -= SIZE_STEP;
+    } else {
+      break;
+    }
+    pixels = roundedWidth * roundedHeight;
+  }
+
+  while (pixels < minPixels) {
+    if (roundedWidth <= roundedHeight) {
+      roundedWidth += SIZE_STEP;
+    } else {
+      roundedHeight += SIZE_STEP;
+    }
+    pixels = roundedWidth * roundedHeight;
+  }
+
+  return { width: roundedWidth, height: roundedHeight };
+}
+
+export function getSizeFromAspectRatio(ar: string | null, quality: CliArgs["quality"]): string {
+  const normalizedQuality = normalizeQuality(quality);
+  const sizes = normalizedQuality === "2k" ? LEGACY_STANDARD_SIZES_2K : LEGACY_STANDARD_SIZES;
+  const defaultSize = normalizedQuality === "2k" ? "1536*1536" : "1024*1024";
+
+  if (!ar) return defaultSize;
+
+  const parsed = parseAspectRatio(ar);
+  if (!parsed) return defaultSize;
+
+  const targetRatio = parsed.width / parsed.height;
+  let best = defaultSize;
+  let bestDiff = Infinity;
+
+  for (const [width, height] of sizes) {
+    const diff = Math.abs(width / height - targetRatio);
+    if (diff < bestDiff) {
+      bestDiff = diff;
+      best = formatSize(width, height);
+    }
+  }
+
+  return best;
+}
+
+export function getQwen2SizeFromAspectRatio(ar: string | null, quality: CliArgs["quality"]): string {
+  const normalizedQuality = normalizeQuality(quality);
+
+  if (!ar) {
+    return QWEN_2_RECOMMENDED["1:1"][normalizedQuality];
+  }
+
+  const recommendedRatio = findKnownRatioKey(ar, Object.keys(QWEN_2_RECOMMENDED));
+  if (recommendedRatio) {
+    return QWEN_2_RECOMMENDED[recommendedRatio][normalizedQuality];
+  }
+
+  const parsed = parseAspectRatio(ar);
+  if (!parsed) {
+    return QWEN_2_RECOMMENDED["1:1"][normalizedQuality];
+  }
+
+  const targetRatio = parsed.width / parsed.height;
+  const targetPixels = QWEN_2_TARGET_PIXELS[normalizedQuality];
+  const rawWidth = Math.sqrt(targetPixels * targetRatio);
+  const rawHeight = Math.sqrt(targetPixels / targetRatio);
+  const fitted = fitToPixelBudget(
+    rawWidth,
+    rawHeight,
+    MIN_QWEN_2_TOTAL_PIXELS,
+    MAX_QWEN_2_TOTAL_PIXELS,
+  );
+
+  return formatSize(fitted.width, fitted.height);
+}
+
+function getQwenFixedSizeFromAspectRatio(ar: string | null, quality: CliArgs["quality"]): string {
+  if (quality === "normal") {
+    console.warn(
+      "DashScope qwen-image-max/plus/image models use fixed output sizes; --quality normal does not change the generated resolution."
+    );
+  }
+
+  if (!ar) return QWEN_FIXED_SPEC.defaultSize;
+
+  const ratioKey = findKnownRatioKey(ar, Object.keys(QWEN_FIXED_SIZES_BY_RATIO));
+  if (!ratioKey) {
+    throw new Error(
+      `DashScope model supports only fixed ratios ${Object.keys(QWEN_FIXED_SIZES_BY_RATIO).join(", ")}. ` +
+      `For custom ratios like "${ar}", use --model qwen-image-2.0-pro.`
+    );
+  }
+
+  return QWEN_FIXED_SIZES_BY_RATIO[ratioKey]!;
+}
+
+function validateSizeFormat(size: string): { width: number; height: number } {
+  const parsed = parseSize(size);
+  if (!parsed) {
+    throw new Error(`Invalid DashScope size "${size}". Expected <width>x<height> or <width>*<height>.`);
+  }
+  return parsed;
+}
+
+function validateQwen2Size(size: string): string {
+  const normalized = normalizeSize(size);
+  const parsed = validateSizeFormat(normalized);
+  const totalPixels = parsed.width * parsed.height;
+  if (totalPixels < MIN_QWEN_2_TOTAL_PIXELS || totalPixels > MAX_QWEN_2_TOTAL_PIXELS) {
+    throw new Error(
+      `DashScope qwen-image-2.0* models require total pixels between ${MIN_QWEN_2_TOTAL_PIXELS} ` +
+      `and ${MAX_QWEN_2_TOTAL_PIXELS}. Received ${normalized} (${totalPixels} pixels).`
+    );
+  }
+  return normalized;
+}
+
+function validateQwenFixedSize(size: string): string {
+  const normalized = normalizeSize(size);
+  validateSizeFormat(normalized);
+  if (!QWEN_FIXED_SIZES.includes(normalized)) {
+    throw new Error(
+      `DashScope qwen-image-max/plus/image models support only these sizes: ${QWEN_FIXED_SIZES.join(", ")}. ` +
+      `Received ${normalized}.`
+    );
+  }
+  return normalized;
+}
+
+export function resolveSizeForModel(
+  model: string,
+  args: Pick<CliArgs, "size" | "aspectRatio" | "quality">,
+): string {
+  const spec = getModelSpec(model);
+
+  if (args.size) {
+    if (spec.family === "qwen2") return validateQwen2Size(args.size);
+    if (spec.family === "qwenFixed") return validateQwenFixedSize(args.size);
+    validateSizeFormat(args.size);
+    return normalizeSize(args.size);
+  }
+
+  if (spec.family === "qwen2") {
+    return getQwen2SizeFromAspectRatio(args.aspectRatio, args.quality);
+  }
+
+  if (spec.family === "qwenFixed") {
+    return getQwenFixedSizeFromAspectRatio(args.aspectRatio, args.quality);
+  }
+
+  return getSizeFromAspectRatio(args.aspectRatio, args.quality);
+}
+
+function buildParameters(
+  family: DashScopeModelFamily,
+  size: string,
+): Record<string, unknown> {
+  const parameters: Record<string, unknown> = {
+    prompt_extend: false,
+    size,
+  };
+
+  if (family === "qwen2" || family === "qwenFixed") {
+    parameters.watermark = false;
+    parameters.negative_prompt = QWEN_NEGATIVE_PROMPT;
+  }
+
+  return parameters;
+}
+
+type DashScopeResponse = {
+  output?: {
+    result_image?: string;
+    choices?: Array<{
+      message?: {
+        content?: Array<{ image?: string }>;
+      };
+    }>;
+  };
+};
+
+async function extractImageFromResponse(result: DashScopeResponse): Promise<Uint8Array> {
+  let imageData: string | null = null;
+
+  if (result.output?.result_image) {
+    imageData = result.output.result_image;
+  } else if (result.output?.choices?.[0]?.message?.content) {
+    const content = result.output.choices[0].message.content;
+    for (const item of content) {
+      if (item.image) {
+        imageData = item.image;
+        break;
+      }
+    }
+  }
+
+  if (!imageData) {
+    console.error("Response:", JSON.stringify(result, null, 2));
+    throw new Error("No image in response");
+  }
+
+  if (imageData.startsWith("http://") || imageData.startsWith("https://")) {
+    const imgRes = await fetch(imageData);
+    if (!imgRes.ok) throw new Error("Failed to download image");
+    const buf = await imgRes.arrayBuffer();
+    return new Uint8Array(buf);
+  }
+
+  return Uint8Array.from(Buffer.from(imageData, "base64"));
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const apiKey = getApiKey();
+  if (!apiKey) throw new Error("DASHSCOPE_API_KEY is required");
+
+  if (args.referenceImages.length > 0) {
+    throw new Error(
+      "Reference images are not supported with DashScope provider in baoyu-image-gen. Use --provider google with a Gemini multimodal model."
+    );
+  }
+
+  const spec = getModelSpec(model);
+  const size = resolveSizeForModel(model, args);
+  const url = `${getBaseUrl()}/api/v1/services/aigc/multimodal-generation/generation`;
+
+  const body = {
+    model,
+    input: {
+      messages: [
+        {
+          role: "user",
+          content: [{ text: prompt }],
+        },
+      ],
+    },
+    parameters: buildParameters(spec.family, size),
+  };
+
+  console.log(`Generating image with DashScope (${model})...`, { family: spec.family, size });
+
+  const res = await fetch(url, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`DashScope API error (${res.status}): ${err}`);
+  }
+
+  const result = await res.json() as DashScopeResponse;
+  return extractImageFromResponse(result);
+}

skills/baoyu-image-gen/scripts/providers/google.test.ts

@@ -0,0 +1,126 @@
+import assert from "node:assert/strict";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  addAspectRatioToPrompt,
+  buildGoogleUrl,
+  buildPromptWithAspect,
+  extractInlineImageData,
+  extractPredictedImageData,
+  getGoogleImageSize,
+  isGoogleImagen,
+  isGoogleMultimodal,
+  normalizeGoogleModelId,
+} from "./google.ts";
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+test("Google provider helpers normalize model IDs and select image size defaults", () => {
+  assert.equal(
+    normalizeGoogleModelId("models/gemini-3.1-flash-image-preview"),
+    "gemini-3.1-flash-image-preview",
+  );
+  assert.equal(isGoogleMultimodal("models/gemini-3-pro-image-preview"), true);
+  assert.equal(isGoogleImagen("imagen-3.0-generate-002"), true);
+  assert.equal(getGoogleImageSize(makeArgs({ imageSize: null, quality: "2k" })), "2K");
+  assert.equal(getGoogleImageSize(makeArgs({ imageSize: "4K", quality: "normal" })), "4K");
+});
+
+test("Google URL builder appends v1beta when the base URL does not already include it", (t) => {
+  useEnv(t, { GOOGLE_BASE_URL: "https://generativelanguage.googleapis.com" });
+  assert.equal(
+    buildGoogleUrl("models/demo:generateContent"),
+    "https://generativelanguage.googleapis.com/v1beta/models/demo:generateContent",
+  );
+});
+
+test("Google URL and prompt helpers preserve existing v1beta paths and aspect hints", (t) => {
+  useEnv(t, { GOOGLE_BASE_URL: "https://example.com/custom/v1beta/" });
+  assert.equal(
+    buildGoogleUrl("/models/demo:predict"),
+    "https://example.com/custom/v1beta/models/demo:predict",
+  );
+
+  assert.equal(
+    addAspectRatioToPrompt("A city skyline", "16:9"),
+    "A city skyline Aspect ratio: 16:9.",
+  );
+  assert.equal(
+    buildPromptWithAspect("A city skyline", "16:9", "2k"),
+    "A city skyline Aspect ratio: 16:9. High resolution 2048px.",
+  );
+});
+
+test("Google response extractors find inline and predicted image payloads", () => {
+  assert.equal(
+    extractInlineImageData({
+      candidates: [
+        {
+          content: {
+            parts: [{ inlineData: { data: "inline-base64" } }],
+          },
+        },
+      ],
+    }),
+    "inline-base64",
+  );
+
+  assert.equal(
+    extractPredictedImageData({
+      predictions: [{ image: { imageBytes: "predicted-base64" } }],
+    }),
+    "predicted-base64",
+  );
+
+  assert.equal(
+    extractPredictedImageData({
+      generatedImages: [{ bytesBase64Encoded: "generated-base64" }],
+    }),
+    "generated-base64",
+  );
+});

skills/baoyu-image-gen/scripts/providers/google.ts

@@ -0,0 +1,349 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+import { execFileSync } from "node:child_process";
+import type { CliArgs } from "../types";
+
+const GOOGLE_MULTIMODAL_MODELS = [
+  "gemini-3-pro-image-preview",
+  "gemini-3-flash-preview",
+  "gemini-3.1-flash-image-preview",
+];
+const GOOGLE_IMAGEN_MODELS = [
+  "imagen-3.0-generate-002",
+  "imagen-3.0-generate-001",
+];
+
+export function getDefaultModel(): string {
+  return process.env.GOOGLE_IMAGE_MODEL || "gemini-3-pro-image-preview";
+}
+
+export function normalizeGoogleModelId(model: string): string {
+  return model.startsWith("models/") ? model.slice("models/".length) : model;
+}
+
+export function isGoogleMultimodal(model: string): boolean {
+  const normalized = normalizeGoogleModelId(model);
+  return GOOGLE_MULTIMODAL_MODELS.some((m) => normalized.includes(m));
+}
+
+export function isGoogleImagen(model: string): boolean {
+  const normalized = normalizeGoogleModelId(model);
+  return GOOGLE_IMAGEN_MODELS.some((m) => normalized.includes(m));
+}
+
+function getGoogleApiKey(): string | null {
+  return process.env.GOOGLE_API_KEY || process.env.GEMINI_API_KEY || null;
+}
+
+export function getGoogleImageSize(args: CliArgs): "1K" | "2K" | "4K" {
+  if (args.imageSize) return args.imageSize as "1K" | "2K" | "4K";
+  return args.quality === "2k" ? "2K" : "1K";
+}
+
+function getGoogleBaseUrl(): string {
+  const base =
+    process.env.GOOGLE_BASE_URL || "https://generativelanguage.googleapis.com";
+  return base.replace(/\/+$/g, "");
+}
+
+export function buildGoogleUrl(pathname: string): string {
+  const base = getGoogleBaseUrl();
+  const cleanedPath = pathname.replace(/^\/+/g, "");
+  if (base.endsWith("/v1beta")) return `${base}/${cleanedPath}`;
+  return `${base}/v1beta/${cleanedPath}`;
+}
+
+function toModelPath(model: string): string {
+  const modelId = normalizeGoogleModelId(model);
+  return `models/${modelId}`;
+}
+
+function getHttpProxy(): string | null {
+  return (
+    process.env.https_proxy ||
+    process.env.HTTPS_PROXY ||
+    process.env.http_proxy ||
+    process.env.HTTP_PROXY ||
+    process.env.ALL_PROXY ||
+    null
+  );
+}
+
+async function postGoogleJsonViaCurl<T>(
+  url: string,
+  apiKey: string,
+  body: unknown,
+): Promise<T> {
+  const proxy = getHttpProxy();
+  const bodyStr = JSON.stringify(body);
+  const args = [
+    "-s",
+    "--connect-timeout",
+    "30",
+    "--max-time",
+    "300",
+    ...(proxy ? ["-x", proxy] : []),
+    url,
+    "-H",
+    "Content-Type: application/json",
+    "-H",
+    `x-goog-api-key: ${apiKey}`,
+    "-d",
+    "@-",
+  ];
+
+  let result = "";
+  try {
+    result = execFileSync("curl", args, {
+      input: bodyStr,
+      encoding: "utf8",
+      maxBuffer: 100 * 1024 * 1024,
+      timeout: 310000,
+    });
+  } catch (error) {
+    const e = error as { message?: string; stderr?: string | Buffer };
+    const stderrText =
+      typeof e.stderr === "string"
+        ? e.stderr
+        : e.stderr
+          ? e.stderr.toString("utf8")
+          : "";
+    const details = stderrText.trim() || e.message || "curl request failed";
+    throw new Error(`Google API request failed via curl: ${details}`);
+  }
+
+  const parsed = JSON.parse(result) as any;
+  if (parsed.error) {
+    throw new Error(
+      `Google API error (${parsed.error.code}): ${parsed.error.message}`,
+    );
+  }
+  return parsed as T;
+}
+
+async function postGoogleJsonViaFetch<T>(
+  url: string,
+  apiKey: string,
+  body: unknown,
+): Promise<T> {
+  const res = await fetch(url, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      "x-goog-api-key": apiKey,
+    },
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`Google API error (${res.status}): ${err}`);
+  }
+
+  return (await res.json()) as T;
+}
+
+async function postGoogleJson<T>(pathname: string, body: unknown): Promise<T> {
+  const apiKey = getGoogleApiKey();
+  if (!apiKey) throw new Error("GOOGLE_API_KEY or GEMINI_API_KEY is required");
+
+  const url = buildGoogleUrl(pathname);
+  const proxy = getHttpProxy();
+
+  // When an HTTP proxy is detected, use curl instead of fetch.
+  // Bun's fetch has a known issue where long-lived connections through
+  // HTTP proxies get their sockets closed unexpectedly, causing image
+  // generation requests to fail with "socket connection was closed
+  // unexpectedly". Using curl as the HTTP client works around this.
+  if (proxy) {
+    return postGoogleJsonViaCurl<T>(url, apiKey, body);
+  }
+
+  return postGoogleJsonViaFetch<T>(url, apiKey, body);
+}
+
+export function buildPromptWithAspect(
+  prompt: string,
+  ar: string | null,
+  quality: CliArgs["quality"],
+): string {
+  let result = prompt;
+  if (ar) {
+    result += ` Aspect ratio: ${ar}.`;
+  }
+  if (quality === "2k") {
+    result += " High resolution 2048px.";
+  }
+  return result;
+}
+
+export function addAspectRatioToPrompt(prompt: string, ar: string | null): string {
+  if (!ar) return prompt;
+  return `${prompt} Aspect ratio: ${ar}.`;
+}
+
+async function readImageAsBase64(
+  p: string,
+): Promise<{ data: string; mimeType: string }> {
+  const buf = await readFile(p);
+  const ext = path.extname(p).toLowerCase();
+  let mimeType = "image/png";
+  if (ext === ".jpg" || ext === ".jpeg") mimeType = "image/jpeg";
+  else if (ext === ".gif") mimeType = "image/gif";
+  else if (ext === ".webp") mimeType = "image/webp";
+  return { data: buf.toString("base64"), mimeType };
+}
+
+export function extractInlineImageData(response: {
+  candidates?: Array<{
+    content?: { parts?: Array<{ inlineData?: { data?: string } }> };
+  }>;
+}): string | null {
+  for (const candidate of response.candidates || []) {
+    for (const part of candidate.content?.parts || []) {
+      const data = part.inlineData?.data;
+      if (typeof data === "string" && data.length > 0) return data;
+    }
+  }
+  return null;
+}
+
+export function extractPredictedImageData(response: {
+  predictions?: Array<any>;
+  generatedImages?: Array<any>;
+}): string | null {
+  const candidates = [
+    ...(response.predictions || []),
+    ...(response.generatedImages || []),
+  ];
+  for (const candidate of candidates) {
+    if (!candidate || typeof candidate !== "object") continue;
+    if (typeof candidate.imageBytes === "string") return candidate.imageBytes;
+    if (typeof candidate.bytesBase64Encoded === "string")
+      return candidate.bytesBase64Encoded;
+    if (typeof candidate.data === "string") return candidate.data;
+    const image = candidate.image;
+    if (image && typeof image === "object") {
+      if (typeof image.imageBytes === "string") return image.imageBytes;
+      if (typeof image.bytesBase64Encoded === "string")
+        return image.bytesBase64Encoded;
+      if (typeof image.data === "string") return image.data;
+    }
+  }
+  return null;
+}
+
+async function generateWithGemini(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+): Promise<Uint8Array> {
+  const promptWithAspect = addAspectRatioToPrompt(prompt, args.aspectRatio);
+  const parts: Array<{
+    text?: string;
+    inlineData?: { data: string; mimeType: string };
+  }> = [];
+  for (const refPath of args.referenceImages) {
+    const { data, mimeType } = await readImageAsBase64(refPath);
+    parts.push({ inlineData: { data, mimeType } });
+  }
+  parts.push({ text: promptWithAspect });
+
+  const imageConfig: { imageSize: "1K" | "2K" | "4K" } = {
+    imageSize: getGoogleImageSize(args),
+  };
+
+  console.log("Generating image with Gemini...", imageConfig);
+  const response = await postGoogleJson<{
+    candidates?: Array<{
+      content?: { parts?: Array<{ inlineData?: { data?: string } }> };
+    }>;
+  }>(`${toModelPath(model)}:generateContent`, {
+    contents: [
+      {
+        role: "user",
+        parts,
+      },
+    ],
+    generationConfig: {
+      responseModalities: ["IMAGE"],
+      imageConfig,
+    },
+  });
+  console.log("Generation completed.");
+
+  const imageData = extractInlineImageData(response);
+  if (imageData) return Uint8Array.from(Buffer.from(imageData, "base64"));
+
+  throw new Error("No image in response");
+}
+
+async function generateWithImagen(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+): Promise<Uint8Array> {
+  const fullPrompt = buildPromptWithAspect(
+    prompt,
+    args.aspectRatio,
+    args.quality,
+  );
+  const imageSize = getGoogleImageSize(args);
+  if (imageSize === "4K") {
+    console.error(
+      "Warning: Imagen models do not support 4K imageSize, using 2K instead.",
+    );
+  }
+
+  const parameters: Record<string, unknown> = {
+    sampleCount: args.n,
+  };
+  if (args.aspectRatio) {
+    parameters.aspectRatio = args.aspectRatio;
+  }
+  if (imageSize === "1K" || imageSize === "2K") {
+    parameters.imageSize = imageSize;
+  } else {
+    parameters.imageSize = "2K";
+  }
+
+  const response = await postGoogleJson<{
+    predictions?: Array<any>;
+    generatedImages?: Array<any>;
+  }>(`${toModelPath(model)}:predict`, {
+    instances: [
+      {
+        prompt: fullPrompt,
+      },
+    ],
+    parameters,
+  });
+
+  const imageData = extractPredictedImageData(response);
+  if (imageData) return Uint8Array.from(Buffer.from(imageData, "base64"));
+
+  throw new Error("No image in response");
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+): Promise<Uint8Array> {
+  if (isGoogleImagen(model)) {
+    if (args.referenceImages.length > 0) {
+      throw new Error(
+        "Reference images are not supported with Imagen models. Use gemini-3-pro-image-preview, gemini-3-flash-preview, or gemini-3.1-flash-image-preview.",
+      );
+    }
+    return generateWithImagen(prompt, model, args);
+  }
+
+  if (!isGoogleMultimodal(model) && args.referenceImages.length > 0) {
+    throw new Error(
+      "Reference images are only supported with Gemini multimodal models. Use gemini-3-pro-image-preview, gemini-3-flash-preview, or gemini-3.1-flash-image-preview.",
+    );
+  }
+
+  return generateWithGemini(prompt, model, args);
+}

skills/baoyu-image-gen/scripts/providers/jimeng.test.ts

@@ -0,0 +1,114 @@
+import assert from "node:assert/strict";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import { generateImage } from "./jimeng.ts";
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+test("Jimeng submit request uses prompt field expected by current API", async (t) => {
+  useEnv(t, {
+    JIMENG_ACCESS_KEY_ID: "test-access-key",
+    JIMENG_SECRET_ACCESS_KEY: "test-secret-key",
+    JIMENG_BASE_URL: null,
+    JIMENG_REGION: null,
+  });
+
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const calls: Array<{
+    input: string;
+    init?: RequestInit;
+  }> = [];
+
+  globalThis.fetch = async (input, init) => {
+    calls.push({
+      input: String(input),
+      init,
+    });
+
+    if (calls.length === 1) {
+      return Response.json({
+        code: 10000,
+        data: {
+          task_id: "task-123",
+        },
+      });
+    }
+
+    return Response.json({
+      code: 10000,
+      data: {
+        status: "done",
+        binary_data_base64: [Buffer.from("jimeng-image").toString("base64")],
+      },
+    });
+  };
+
+  const image = await generateImage(
+    "A quiet bamboo forest",
+    "jimeng_t2i_v40",
+    makeArgs({ quality: "normal" }),
+  );
+
+  assert.equal(Buffer.from(image).toString("utf8"), "jimeng-image");
+  assert.equal(calls.length, 2);
+  assert.equal(
+    calls[0]?.input,
+    "https://visual.volcengineapi.com/?Action=CVSync2AsyncSubmitTask&Version=2022-08-31",
+  );
+
+  const submitBody = JSON.parse(String(calls[0]?.init?.body)) as Record<string, unknown>;
+  assert.equal(submitBody.req_key, "jimeng_t2i_v40");
+  assert.equal(submitBody.prompt, "A quiet bamboo forest");
+  assert.ok(!("prompt_text" in submitBody));
+  assert.equal(submitBody.width, 1024);
+  assert.equal(submitBody.height, 1024);
+});

skills/baoyu-image-gen/scripts/providers/jimeng.ts

@@ -0,0 +1,467 @@
+import type { CliArgs } from "../types";
+import * as crypto from "node:crypto";
+
+type JimengSizePreset = "normal" | "2k" | "4k";
+
+export function getDefaultModel(): string {
+  return process.env.JIMENG_IMAGE_MODEL || "jimeng_t2i_v40";
+}
+
+function getAccessKey(): string | null {
+  return process.env.JIMENG_ACCESS_KEY_ID || null;
+}
+
+function getSecretKey(): string | null {
+  return process.env.JIMENG_SECRET_ACCESS_KEY || null;
+}
+
+function getRegion(): string {
+  return process.env.JIMENG_REGION || "cn-north-1";
+}
+
+function getBaseUrl(): string {
+  return process.env.JIMENG_BASE_URL || "https://visual.volcengineapi.com";
+}
+
+function resolveEndpoint(query: Record<string, string>): {
+  url: string;
+  host: string;
+  canonicalUri: string;
+} {
+  let baseUrl: URL;
+  try {
+    baseUrl = new URL(getBaseUrl());
+  } catch {
+    throw new Error(`Invalid JIMENG_BASE_URL: ${getBaseUrl()}`);
+  }
+
+  baseUrl.search = "";
+  for (const [key, value] of Object.entries(query).sort(([a], [b]) => a.localeCompare(b))) {
+    baseUrl.searchParams.set(key, value);
+  }
+
+  return {
+    url: baseUrl.toString(),
+    host: baseUrl.host,
+    canonicalUri: baseUrl.pathname || "/",
+  };
+}
+
+/**
+ * Volcengine HMAC-SHA256 signature generation
+ * Following the official documentation at:
+ * https://www.volcengine.com/docs/85621/1817045
+ */
+function generateSignature(
+  method: string,
+  query: Record<string, string>,
+  headers: Record<string, string>,
+  body: string,
+  accessKey: string,
+  secretKey: string,
+  region: string,
+  service: string,
+  canonicalUri: string
+): string {
+  // 1. Create canonical request
+  // Sort query parameters alphabetically
+  const sortedQuery = Object.entries(query)
+    .sort(([a], [b]) => a.localeCompare(b))
+    .map(([k, v]) => `${encodeURIComponent(k)}=${encodeURIComponent(v)}`)
+    .join("&");
+
+  // Sort headers alphabetically and create canonical headers
+  const sortedHeaders = Object.entries(headers)
+    .sort(([a], [b]) => a.localeCompare(b))
+    .map(([k, v]) => `${k.toLowerCase()}:${v.trim()}\n`)
+    .join("");
+
+  const signedHeaders = Object.keys(headers)
+    .sort()
+    .map(k => k.toLowerCase())
+    .join(";");
+
+  const hashedPayload = crypto.createHash("sha256").update(body, "utf8").digest("hex");
+
+  const canonicalRequest = [
+    method,
+    canonicalUri,
+    sortedQuery,
+    sortedHeaders,
+    signedHeaders,
+    hashedPayload,
+  ].join("\n");
+
+  const hashedCanonicalRequest = crypto
+    .createHash("sha256")
+    .update(canonicalRequest, "utf8")
+    .digest("hex");
+
+  // 2. Create string to sign
+  const algorithm = "HMAC-SHA256";
+  const timestamp = headers["X-Date"] || headers["x-date"];
+  if (!timestamp) {
+    throw new Error("Jimeng signature generation requires an X-Date header.");
+  }
+  const dateStamp = timestamp.slice(0, 8);
+
+  const credentialScope = `${dateStamp}/${region}/${service}/request`;
+
+  const stringToSign = [
+    algorithm,
+    timestamp,
+    credentialScope,
+    hashedCanonicalRequest,
+  ].join("\n");
+
+  // 3. Calculate signature
+  const kDate = crypto
+    .createHmac("sha256", secretKey)
+    .update(dateStamp)
+    .digest();
+
+  const kRegion = crypto.createHmac("sha256", kDate).update(region).digest();
+  const kService = crypto.createHmac("sha256", kRegion).update(service).digest();
+  const kSigning = crypto.createHmac("sha256", kService).update("request").digest();
+
+  const signature = crypto
+    .createHmac("sha256", kSigning)
+    .update(stringToSign)
+    .digest("hex");
+
+  // 4. Create authorization header
+  return `${algorithm} Credential=${accessKey}/${credentialScope}, SignedHeaders=${signedHeaders}, Signature=${signature}`;
+}
+
+/**
+ * Parse aspect ratio string like "16:9", "1:1", "4:3" into width and height
+ */
+function parseAspectRatio(ar: string): { width: number; height: number } | null {
+  const match = ar.match(/^(\d+(?:\.\d+)?):(\d+(?:\.\d+)?)$/);
+  if (!match) return null;
+  const w = parseFloat(match[1]!);
+  const h = parseFloat(match[2]!);
+  if (w <= 0 || h <= 0) return null;
+  return { width: w, height: h };
+}
+
+/**
+ * Supported size presets for different quality levels
+ * Based on Volcengine Jimeng documentation
+ */
+const SIZE_PRESETS: Record<string, Record<string, string>> = {
+  normal: {
+    "1:1": "1024x1024",
+    "4:3": "1360x1020",
+    "16:9": "1536x864",
+    "3:2": "1440x960",
+    "21:9": "1920x824",
+  },
+  "2k": {
+    "1:1": "2048x2048",
+    "4:3": "2304x1728",
+    "16:9": "2560x1440",
+    "3:2": "2496x1664",
+    "21:9": "3024x1296",
+  },
+  "4k": {
+    "1:1": "4096x4096",
+    "4:3": "4694x3520",
+    "16:9": "5404x3040",
+    "3:2": "4992x3328",
+    "21:9": "6198x2656",
+  },
+};
+
+function normalizeDimensions(value: string): string | null {
+  const match = value.trim().match(/^(\d+)\s*[xX*]\s*(\d+)$/);
+  if (!match) return null;
+  return `${match[1]}x${match[2]}`;
+}
+
+function getClosestPresetSize(ar: string | null, qualityLevel: JimengSizePreset): string {
+  const presets = SIZE_PRESETS[qualityLevel];
+  const defaultSize = presets["1:1"]!;
+
+  if (!ar) return defaultSize;
+
+  const parsed = parseAspectRatio(ar);
+  if (!parsed) return defaultSize;
+
+  const targetRatio = parsed.width / parsed.height;
+  let bestMatch = defaultSize;
+  let bestDiff = Infinity;
+
+  for (const [ratio, size] of Object.entries(presets)) {
+    const [w, h] = ratio.split(":").map(Number);
+    const presetRatio = w / h;
+    const diff = Math.abs(presetRatio - targetRatio);
+    if (diff < bestDiff) {
+      bestDiff = diff;
+      bestMatch = size;
+    }
+  }
+
+  return bestMatch;
+}
+
+function normalizeImageSizePreset(imageSize: string, ar: string | null): string | null {
+  const preset = imageSize.trim().toUpperCase();
+  if (preset === "1K") return getClosestPresetSize(ar, "normal");
+  if (preset === "2K") return getClosestPresetSize(ar, "2k");
+  if (preset === "4K") return getClosestPresetSize(ar, "4k");
+  return normalizeDimensions(imageSize);
+}
+
+function getImageSize(ar: string | null, quality: CliArgs["quality"], imageSize?: string | null): string {
+  if (imageSize) {
+    const normalizedSize = normalizeImageSizePreset(imageSize, ar);
+    if (normalizedSize) return normalizedSize;
+  }
+
+  // Default to 2K quality if not specified
+  const qualityLevel: JimengSizePreset = quality === "normal" ? "normal" : "2k";
+  return getClosestPresetSize(ar, qualityLevel);
+}
+
+/**
+ * Step 1: Submit async task to Volcengine Jimeng API
+ */
+async function submitTask(
+  prompt: string,
+  model: string,
+  size: string,
+  accessKey: string,
+  secretKey: string,
+  region: string
+): Promise<string> {
+  // Query parameters for submit endpoint
+  const query = {
+    Action: "CVSync2AsyncSubmitTask",
+    Version: "2022-08-31",
+  };
+  const endpoint = resolveEndpoint(query);
+
+  // Request body - Jimeng API expects width/height as separate integers
+  const [width, height] = size.split("x").map(Number);
+  const bodyObj = {
+    req_key: model,
+    prompt,
+    // Use separate width and height parameters instead of size string
+    width: width,
+    height: height,
+    // Optional: seed for reproducibility
+    // seed: Math.floor(Math.random() * 999999),
+  };
+
+  const body = JSON.stringify(bodyObj);
+
+  // Headers
+  const timestampHeader = new Date().toISOString().replace(/[:\-]|\.\d{3}/g, "");
+  const headers = {
+    "Content-Type": "application/json",
+    "X-Date": timestampHeader,
+    "Host": endpoint.host,
+  };
+
+  // Generate signature
+  const authorization = generateSignature(
+    "POST",
+    query,
+    headers,
+    body,
+    accessKey,
+    secretKey,
+    region,
+    "cv",
+    endpoint.canonicalUri
+  );
+
+  console.error(`Submitting task to Jimeng (${model})...`, { width, height });
+
+  const res = await fetch(endpoint.url, {
+    method: "POST",
+    headers: {
+      ...headers,
+      "Authorization": authorization,
+    },
+    body,
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`Jimeng API submit error (${res.status}): ${err}`);
+  }
+
+  const result = (await res.json()) as {
+    code?: number;
+    message?: string;
+    data?: {
+      task_id?: string;
+    };
+  };
+
+  // Volcengine API returns code 10000 for success
+  if (result.code !== 10000 || !result.data?.task_id) {
+    console.error("Submit response:", JSON.stringify(result, null, 2));
+    throw new Error(`Failed to submit task: ${result.message || "Unknown error"}`);
+  }
+
+  return result.data.task_id;
+}
+
+/**
+ * Step 2: Poll for task result
+ * Returns image data directly as Uint8Array
+ */
+async function pollForResult(
+  taskId: string,
+  model: string,
+  accessKey: string,
+  secretKey: string,
+  region: string
+): Promise<Uint8Array> {
+  const maxAttempts = 60;
+  const pollIntervalMs = 2000;
+
+  for (let attempt = 0; attempt < maxAttempts; attempt++) {
+    // Query parameters for result endpoint
+    const query = {
+      Action: "CVSync2AsyncGetResult",
+      Version: "2022-08-31",
+    };
+    const endpoint = resolveEndpoint(query);
+
+    // Request body - include req_key and task_id
+    const bodyObj = {
+      req_key: model,
+      task_id: taskId,
+    };
+
+    const body = JSON.stringify(bodyObj);
+
+    // Headers
+    const timestampHeader = new Date().toISOString().replace(/[:\-]|\.\d{3}/g, "");
+    const headers = {
+      "Content-Type": "application/json",
+      "X-Date": timestampHeader,
+      "Host": endpoint.host,
+    };
+
+    // Generate signature
+    const authorization = generateSignature(
+      "POST",
+      query,
+      headers,
+      body,
+      accessKey,
+      secretKey,
+      region,
+      "cv",
+      endpoint.canonicalUri
+    );
+
+    const res = await fetch(endpoint.url, {
+      method: "POST",
+      headers: {
+        ...headers,
+        "Authorization": authorization,
+      },
+      body,
+    });
+
+    if (!res.ok) {
+      const err = await res.text();
+      throw new Error(`Jimeng API poll error (${res.status}): ${err}`);
+    }
+
+    const result = (await res.json()) as {
+      code?: number;
+      message?: string;
+      data?: {
+        status?: string;
+        image_urls?: string[];
+        binary_data_base64?: string[];
+      };
+    };
+
+    // Volcengine API returns code 10000 for success
+    if (result.code === 10000 && result.data) {
+      const { status, image_urls, binary_data_base64 } = result.data;
+
+      // Check for base64 image data (preferred by Jimeng)
+      if (binary_data_base64 && binary_data_base64.length > 0) {
+        console.error("Image received as base64 data");
+        const base64Data = binary_data_base64[0]!;
+        // Convert base64 to Uint8Array
+        const binaryString = Buffer.from(base64Data, "base64").toString("binary");
+        const bytes = new Uint8Array(binaryString.length);
+        for (let i = 0; i < binaryString.length; i++) {
+          bytes[i] = binaryString.charCodeAt(i);
+        }
+        return bytes;
+      }
+
+      // Fallback to URL format
+      if (status === "done" && image_urls && image_urls.length > 0) {
+        // Download from URL
+        console.error(`Downloading image from ${image_urls[0]}...`);
+        const imgRes = await fetch(image_urls[0]!);
+        if (!imgRes.ok) {
+          throw new Error(`Failed to download image from ${image_urls[0]}`);
+        }
+        const buffer = await imgRes.arrayBuffer();
+        return new Uint8Array(buffer);
+      }
+
+      if (status === "in_queue" || status === "generating") {
+        console.error(`Task status: ${status} (${attempt + 1}/${maxAttempts})`);
+        await new Promise(resolve => setTimeout(resolve, pollIntervalMs));
+        continue;
+      }
+
+      if (status === "fail") {
+        throw new Error(`Jimeng task failed: ${result.message || "Generation failed"}`);
+      }
+    }
+
+    console.error("Poll response:", JSON.stringify(result, null, 2));
+    throw new Error(`Unexpected response during polling: ${result.message || "Unknown error"}`);
+  }
+
+  throw new Error("Task timeout: image generation took too long");
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  if (args.referenceImages.length > 0) {
+    throw new Error(
+      "Jimeng does not support reference images. Use --provider google, openai, openrouter, or replicate."
+    );
+  }
+
+  const accessKey = getAccessKey();
+  const secretKey = getSecretKey();
+  const region = getRegion();
+
+  if (!accessKey || !secretKey) {
+    throw new Error(
+      "JIMENG_ACCESS_KEY_ID and JIMENG_SECRET_ACCESS_KEY are required. " +
+      "Get your credentials from https://console.volcengine.com/iam/keymanage"
+    );
+  }
+
+  const size = getImageSize(args.aspectRatio, args.quality, args.imageSize);
+
+  // Step 1: Submit task
+  const taskId = await submitTask(prompt, model, size, accessKey, secretKey, region);
+
+  // Step 2: Poll for result (returns image data directly)
+  const imageData = await pollForResult(taskId, model, accessKey, secretKey, region);
+
+  console.error("Image generation complete!");
+  return imageData;
+}

skills/baoyu-image-gen/scripts/providers/minimax.test.ts

@@ -0,0 +1,171 @@
+import assert from "node:assert/strict";
+import fs from "node:fs/promises";
+import os from "node:os";
+import path from "node:path";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  buildMinimaxUrl,
+  buildRequestBody,
+  buildSubjectReference,
+  extractImageFromResponse,
+  parsePixelSize,
+  validateArgs,
+} from "./minimax.ts";
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+test("MiniMax URL builder normalizes /v1 suffixes", (t) => {
+  useEnv(t, { MINIMAX_BASE_URL: "https://api.minimax.io" });
+  assert.equal(buildMinimaxUrl(), "https://api.minimax.io/v1/image_generation");
+
+  process.env.MINIMAX_BASE_URL = "https://proxy.example.com/custom/v1/";
+  assert.equal(buildMinimaxUrl(), "https://proxy.example.com/custom/v1/image_generation");
+});
+
+test("MiniMax size parsing and validation follow documented constraints", () => {
+  assert.deepEqual(parsePixelSize("1536x1024"), { width: 1536, height: 1024 });
+  assert.deepEqual(parsePixelSize("1536*1024"), { width: 1536, height: 1024 });
+  assert.equal(parsePixelSize("wide"), null);
+
+  validateArgs("image-01", makeArgs({ size: "1536x1024", n: 9 }));
+
+  assert.throws(
+    () => validateArgs("image-01-live", makeArgs({ size: "1536x1024" })),
+    /only supported with model image-01/,
+  );
+  assert.throws(
+    () => validateArgs("image-01", makeArgs({ size: "1537x1024" })),
+    /divisible by 8/,
+  );
+  assert.throws(
+    () => validateArgs("image-01", makeArgs({ aspectRatio: "2.35:1" })),
+    /aspect_ratio must be one of/,
+  );
+  assert.throws(
+    () => validateArgs("image-01", makeArgs({ n: 10 })),
+    /at most 9 images/,
+  );
+});
+
+test("MiniMax request body maps aspect ratio, size, n, and subject references", async (t) => {
+  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "minimax-test-"));
+  t.after(() => fs.rm(dir, { recursive: true, force: true }));
+
+  const refPath = path.join(dir, "portrait.png");
+  await fs.writeFile(refPath, Buffer.from("portrait"));
+
+  const ratioBody = await buildRequestBody(
+    "A portrait by the window",
+    "image-01",
+    makeArgs({ aspectRatio: "16:9", n: 2, referenceImages: [refPath] }),
+  );
+  assert.equal(ratioBody.aspect_ratio, "16:9");
+  assert.equal(ratioBody.n, 2);
+  assert.equal(ratioBody.response_format, "base64");
+  assert.match(ratioBody.subject_reference?.[0]?.image_file || "", /^data:image\/png;base64,/);
+
+  const sizeBody = await buildRequestBody(
+    "A portrait by the window",
+    "image-01",
+    makeArgs({ size: "1536x1024" }),
+  );
+  assert.equal(sizeBody.width, 1536);
+  assert.equal(sizeBody.height, 1024);
+  assert.equal(sizeBody.aspect_ratio, undefined);
+});
+
+test("MiniMax subject references require supported file types", async (t) => {
+  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "minimax-ref-"));
+  t.after(() => fs.rm(dir, { recursive: true, force: true }));
+
+  const good = path.join(dir, "portrait.jpg");
+  const bad = path.join(dir, "portrait.webp");
+  await fs.writeFile(good, Buffer.from("portrait"));
+  await fs.writeFile(bad, Buffer.from("portrait"));
+
+  const subjectReference = await buildSubjectReference([good]);
+  assert.equal(subjectReference?.[0]?.type, "character");
+
+  await assert.rejects(
+    () => buildSubjectReference([bad]),
+    /only supports JPG, JPEG, or PNG/,
+  );
+});
+
+test("MiniMax response extraction supports base64 and URL payloads", async (t) => {
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const fromBase64 = await extractImageFromResponse({
+    data: {
+      image_base64: [Buffer.from("hello").toString("base64")],
+    },
+  });
+  assert.equal(Buffer.from(fromBase64).toString("utf8"), "hello");
+
+  globalThis.fetch = async () =>
+    new Response(Uint8Array.from([1, 2, 3]), {
+      status: 200,
+      headers: { "Content-Type": "image/jpeg" },
+    });
+
+  const fromUrl = await extractImageFromResponse({
+    data: {
+      image_urls: ["https://example.com/output.jpg"],
+    },
+  });
+  assert.deepEqual([...fromUrl], [1, 2, 3]);
+
+  await assert.rejects(
+    () => extractImageFromResponse({ base_resp: { status_code: 1001, status_msg: "blocked" } }),
+    /blocked/,
+  );
+});

skills/baoyu-image-gen/scripts/providers/minimax.ts

@@ -0,0 +1,220 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+
+import type { CliArgs } from "../types";
+
+const DEFAULT_MODEL = "image-01";
+const MAX_REFERENCE_IMAGE_BYTES = 10 * 1024 * 1024;
+const SUPPORTED_ASPECT_RATIOS = new Set(["1:1", "16:9", "4:3", "3:2", "2:3", "3:4", "9:16", "21:9"]);
+
+type MinimaxSubjectReference = {
+  type: "character";
+  image_file: string;
+};
+
+type MinimaxRequestBody = {
+  model: string;
+  prompt: string;
+  response_format: "base64";
+  aspect_ratio?: string;
+  width?: number;
+  height?: number;
+  n?: number;
+  subject_reference?: MinimaxSubjectReference[];
+};
+
+type MinimaxResponse = {
+  id?: string;
+  data?: {
+    image_urls?: string[];
+    image_base64?: string[];
+  };
+  base_resp?: {
+    status_code?: number;
+    status_msg?: string;
+  };
+};
+
+export function getDefaultModel(): string {
+  return process.env.MINIMAX_IMAGE_MODEL || DEFAULT_MODEL;
+}
+
+function getApiKey(): string | null {
+  return process.env.MINIMAX_API_KEY || null;
+}
+
+export function buildMinimaxUrl(): string {
+  const base = (process.env.MINIMAX_BASE_URL || "https://api.minimax.io").replace(/\/+$/g, "");
+  return base.endsWith("/v1") ? `${base}/image_generation` : `${base}/v1/image_generation`;
+}
+
+function getMimeType(filename: string): "image/jpeg" | "image/png" {
+  const ext = path.extname(filename).toLowerCase();
+  if (ext === ".jpg" || ext === ".jpeg") return "image/jpeg";
+  if (ext === ".png") return "image/png";
+  throw new Error(
+    `MiniMax subject_reference only supports JPG, JPEG, or PNG files: ${filename}`
+  );
+}
+
+export function parsePixelSize(size: string): { width: number; height: number } | null {
+  const match = size.trim().match(/^(\d+)\s*[xX*]\s*(\d+)$/);
+  if (!match) return null;
+
+  const width = parseInt(match[1]!, 10);
+  const height = parseInt(match[2]!, 10);
+  if (!Number.isFinite(width) || !Number.isFinite(height) || width <= 0 || height <= 0) {
+    return null;
+  }
+
+  return { width, height };
+}
+
+function validatePixelSize(width: number, height: number): void {
+  if (width < 512 || width > 2048 || height < 512 || height > 2048) {
+    throw new Error("MiniMax custom size must keep width and height between 512 and 2048.");
+  }
+  if (width % 8 !== 0 || height % 8 !== 0) {
+    throw new Error("MiniMax custom size requires width and height divisible by 8.");
+  }
+}
+
+export function validateArgs(model: string, args: CliArgs): void {
+  if (args.n > 9) {
+    throw new Error("MiniMax supports at most 9 images per request.");
+  }
+
+  if (args.aspectRatio && !SUPPORTED_ASPECT_RATIOS.has(args.aspectRatio)) {
+    throw new Error(
+      `MiniMax aspect_ratio must be one of: ${Array.from(SUPPORTED_ASPECT_RATIOS).join(", ")}.`
+    );
+  }
+
+  if (args.size && !args.aspectRatio) {
+    if (model !== "image-01") {
+      throw new Error("MiniMax custom --size is only supported with model image-01. Use --model image-01 or pass --ar instead.");
+    }
+    const parsed = parsePixelSize(args.size);
+    if (!parsed) {
+      throw new Error("MiniMax --size must be in WxH format, for example 1536x1024.");
+    }
+    validatePixelSize(parsed.width, parsed.height);
+  }
+}
+
+export async function buildSubjectReference(
+  referenceImages: string[],
+): Promise<MinimaxSubjectReference[] | undefined> {
+  if (referenceImages.length === 0) return undefined;
+
+  const subjectReference: MinimaxSubjectReference[] = [];
+  for (const refPath of referenceImages) {
+    const bytes = await readFile(refPath);
+    if (bytes.length > MAX_REFERENCE_IMAGE_BYTES) {
+      throw new Error(`MiniMax subject_reference images must be smaller than 10MB: ${refPath}`);
+    }
+
+    subjectReference.push({
+      type: "character",
+      image_file: `data:${getMimeType(refPath)};base64,${bytes.toString("base64")}`,
+    });
+  }
+
+  return subjectReference;
+}
+
+export async function buildRequestBody(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+): Promise<MinimaxRequestBody> {
+  validateArgs(model, args);
+
+  const body: MinimaxRequestBody = {
+    model,
+    prompt,
+    response_format: "base64",
+  };
+
+  if (args.aspectRatio) {
+    body.aspect_ratio = args.aspectRatio;
+  } else if (args.size) {
+    const parsed = parsePixelSize(args.size);
+    if (!parsed) {
+      throw new Error("MiniMax --size must be in WxH format, for example 1536x1024.");
+    }
+    body.width = parsed.width;
+    body.height = parsed.height;
+  }
+
+  if (args.n > 1) {
+    body.n = args.n;
+  }
+
+  const subjectReference = await buildSubjectReference(args.referenceImages);
+  if (subjectReference) {
+    body.subject_reference = subjectReference;
+  }
+
+  return body;
+}
+
+async function downloadImage(url: string): Promise<Uint8Array> {
+  const response = await fetch(url);
+  if (!response.ok) {
+    throw new Error(`Failed to download image from MiniMax: ${response.status}`);
+  }
+  return new Uint8Array(await response.arrayBuffer());
+}
+
+export async function extractImageFromResponse(result: MinimaxResponse): Promise<Uint8Array> {
+  const baseResp = result.base_resp;
+  if (baseResp && baseResp.status_code !== undefined && baseResp.status_code !== 0) {
+    throw new Error(baseResp.status_msg || `MiniMax API returned status_code=${baseResp.status_code}`);
+  }
+
+  const base64Image = result.data?.image_base64?.[0];
+  if (base64Image) {
+    return Uint8Array.from(Buffer.from(base64Image, "base64"));
+  }
+
+  const url = result.data?.image_urls?.[0];
+  if (url) {
+    return downloadImage(url);
+  }
+
+  throw new Error("No image data in MiniMax response");
+}
+
+export function getDefaultOutputExtension(): ".jpg" {
+  return ".jpg";
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const apiKey = getApiKey();
+  if (!apiKey) {
+    throw new Error("MINIMAX_API_KEY is required. Get one from https://platform.minimax.io/");
+  }
+
+  const body = await buildRequestBody(prompt, model, args);
+  const response = await fetch(buildMinimaxUrl(), {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify(body),
+  });
+
+  if (!response.ok) {
+    const err = await response.text();
+    throw new Error(`MiniMax API error (${response.status}): ${err}`);
+  }
+
+  const result = (await response.json()) as MinimaxResponse;
+  return extractImageFromResponse(result);
+}

skills/baoyu-image-gen/scripts/providers/openai.test.ts

@@ -0,0 +1,56 @@
+import assert from "node:assert/strict";
+import test from "node:test";
+
+import {
+  extractImageFromResponse,
+  getMimeType,
+  getOpenAISize,
+  parseAspectRatio,
+} from "./openai.ts";
+
+test("OpenAI aspect-ratio parsing and size selection match model families", () => {
+  assert.deepEqual(parseAspectRatio("16:9"), { width: 16, height: 9 });
+  assert.equal(parseAspectRatio("wide"), null);
+  assert.equal(parseAspectRatio("0:1"), null);
+
+  assert.equal(getOpenAISize("dall-e-3", "16:9", "2k"), "1792x1024");
+  assert.equal(getOpenAISize("dall-e-3", "9:16", "normal"), "1024x1792");
+  assert.equal(getOpenAISize("dall-e-2", "16:9", "2k"), "1024x1024");
+  assert.equal(getOpenAISize("gpt-image-1.5", "16:9", "2k"), "1536x1024");
+  assert.equal(getOpenAISize("gpt-image-1.5", "4:3", "2k"), "1024x1024");
+});
+
+test("OpenAI mime-type detection covers supported reference image extensions", () => {
+  assert.equal(getMimeType("frame.png"), "image/png");
+  assert.equal(getMimeType("frame.jpg"), "image/jpeg");
+  assert.equal(getMimeType("frame.webp"), "image/webp");
+  assert.equal(getMimeType("frame.gif"), "image/gif");
+});
+
+test("OpenAI response extraction supports base64 and URL download flows", async (t) => {
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const fromBase64 = await extractImageFromResponse({
+    data: [{ b64_json: Buffer.from("hello").toString("base64") }],
+  });
+  assert.equal(Buffer.from(fromBase64).toString("utf8"), "hello");
+
+  globalThis.fetch = async () =>
+    new Response(Uint8Array.from([1, 2, 3]), {
+      status: 200,
+      headers: { "Content-Type": "application/octet-stream" },
+    });
+
+  const fromUrl = await extractImageFromResponse({
+    data: [{ url: "https://example.com/image.png" }],
+  });
+  assert.deepEqual([...fromUrl], [1, 2, 3]);
+
+  await assert.rejects(
+    () => extractImageFromResponse({ data: [{}] }),
+    /No image in response/,
+  );
+});

skills/baoyu-image-gen/scripts/providers/openai.ts

@@ -0,0 +1,227 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+import type { CliArgs } from "../types";
+
+export function getDefaultModel(): string {
+  return process.env.OPENAI_IMAGE_MODEL || "gpt-image-1.5";
+}
+
+type OpenAIImageResponse = { data: Array<{ url?: string; b64_json?: string }> };
+
+export function parseAspectRatio(ar: string): { width: number; height: number } | null {
+  const match = ar.match(/^(\d+(?:\.\d+)?):(\d+(?:\.\d+)?)$/);
+  if (!match) return null;
+  const w = parseFloat(match[1]!);
+  const h = parseFloat(match[2]!);
+  if (w <= 0 || h <= 0) return null;
+  return { width: w, height: h };
+}
+
+type SizeMapping = {
+  square: string;
+  landscape: string;
+  portrait: string;
+};
+
+export function getOpenAISize(
+  model: string,
+  ar: string | null,
+  quality: CliArgs["quality"]
+): string {
+  const isDalle3 = model.includes("dall-e-3");
+  const isDalle2 = model.includes("dall-e-2");
+
+  if (isDalle2) {
+    return "1024x1024";
+  }
+
+  const sizes: SizeMapping = isDalle3
+    ? {
+        square: "1024x1024",
+        landscape: "1792x1024",
+        portrait: "1024x1792",
+      }
+    : {
+        square: "1024x1024",
+        landscape: "1536x1024",
+        portrait: "1024x1536",
+      };
+
+  if (!ar) return sizes.square;
+
+  const parsed = parseAspectRatio(ar);
+  if (!parsed) return sizes.square;
+
+  const ratio = parsed.width / parsed.height;
+
+  if (Math.abs(ratio - 1) < 0.1) return sizes.square;
+  if (ratio > 1.5) return sizes.landscape;
+  if (ratio < 0.67) return sizes.portrait;
+  return sizes.square;
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const baseURL = process.env.OPENAI_BASE_URL || "https://api.openai.com/v1";
+  const apiKey = process.env.OPENAI_API_KEY;
+
+  if (!apiKey) {
+    throw new Error(
+      "OPENAI_API_KEY is required. Codex/ChatGPT desktop login does not automatically grant OpenAI Images API access to this script."
+    );
+  }
+
+  if (process.env.OPENAI_IMAGE_USE_CHAT === "true") {
+    return generateWithChatCompletions(baseURL, apiKey, prompt, model);
+  }
+
+  const size = args.size || getOpenAISize(model, args.aspectRatio, args.quality);
+
+  if (args.referenceImages.length > 0) {
+    if (model.includes("dall-e-2") || model.includes("dall-e-3")) {
+      throw new Error(
+        "Reference images with OpenAI in this skill require GPT Image models. Use --model gpt-image-1.5 (or another gpt-image model)."
+      );
+    }
+    return generateWithOpenAIEdits(baseURL, apiKey, prompt, model, size, args.referenceImages, args.quality);
+  }
+
+  return generateWithOpenAIGenerations(baseURL, apiKey, prompt, model, size, args.quality);
+}
+
+async function generateWithChatCompletions(
+  baseURL: string,
+  apiKey: string,
+  prompt: string,
+  model: string
+): Promise<Uint8Array> {
+  const res = await fetch(`${baseURL}/chat/completions`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify({
+      model,
+      messages: [{ role: "user", content: prompt }],
+    }),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`OpenAI API error: ${err}`);
+  }
+
+  const result = (await res.json()) as { choices: Array<{ message: { content: string } }> };
+  const content = result.choices[0]?.message?.content ?? "";
+
+  const match = content.match(/data:image\/[^;]+;base64,([A-Za-z0-9+/=]+)/);
+  if (match) {
+    return Uint8Array.from(Buffer.from(match[1]!, "base64"));
+  }
+
+  throw new Error("No image found in chat completions response");
+}
+
+async function generateWithOpenAIGenerations(
+  baseURL: string,
+  apiKey: string,
+  prompt: string,
+  model: string,
+  size: string,
+  quality: CliArgs["quality"]
+): Promise<Uint8Array> {
+  const body: Record<string, any> = { model, prompt, size };
+
+  if (model.includes("dall-e-3")) {
+    body.quality = quality === "2k" ? "hd" : "standard";
+  }
+
+  const res = await fetch(`${baseURL}/images/generations`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`OpenAI API error: ${err}`);
+  }
+
+  const result = (await res.json()) as OpenAIImageResponse;
+  return extractImageFromResponse(result);
+}
+
+async function generateWithOpenAIEdits(
+  baseURL: string,
+  apiKey: string,
+  prompt: string,
+  model: string,
+  size: string,
+  referenceImages: string[],
+  quality: CliArgs["quality"]
+): Promise<Uint8Array> {
+  const form = new FormData();
+  form.append("model", model);
+  form.append("prompt", prompt);
+  form.append("size", size);
+
+  if (model.includes("gpt-image")) {
+    form.append("quality", quality === "2k" ? "high" : "medium");
+  }
+
+  for (const refPath of referenceImages) {
+    const bytes = await readFile(refPath);
+    const filename = path.basename(refPath);
+    const mimeType = getMimeType(filename);
+    const blob = new Blob([bytes], { type: mimeType });
+    form.append("image[]", blob, filename);
+  }
+
+  const res = await fetch(`${baseURL}/images/edits`, {
+    method: "POST",
+    headers: {
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: form,
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`OpenAI edits API error: ${err}`);
+  }
+
+  const result = (await res.json()) as OpenAIImageResponse;
+  return extractImageFromResponse(result);
+}
+
+export function getMimeType(filename: string): string {
+  const ext = path.extname(filename).toLowerCase();
+  if (ext === ".jpg" || ext === ".jpeg") return "image/jpeg";
+  if (ext === ".webp") return "image/webp";
+  if (ext === ".gif") return "image/gif";
+  return "image/png";
+}
+
+export async function extractImageFromResponse(result: OpenAIImageResponse): Promise<Uint8Array> {
+  const img = result.data[0];
+
+  if (img?.b64_json) {
+    return Uint8Array.from(Buffer.from(img.b64_json, "base64"));
+  }
+
+  if (img?.url) {
+    const imgRes = await fetch(img.url);
+    if (!imgRes.ok) throw new Error("Failed to download image");
+    const buf = await imgRes.arrayBuffer();
+    return new Uint8Array(buf);
+  }
+
+  throw new Error("No image in response");
+}

skills/baoyu-image-gen/scripts/providers/openrouter.test.ts

@@ -0,0 +1,168 @@
+import assert from "node:assert/strict";
+import test from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  buildContent,
+  buildRequestBody,
+  extractImageFromResponse,
+  getAspectRatio,
+  getImageSize,
+  validateArgs,
+} from "./openrouter.ts";
+
+const GEMINI_MODEL = "google/gemini-3.1-flash-image-preview";
+const GEMINI_25_MODEL = "google/gemini-2.5-flash-image";
+const GPT_5_IMAGE_MODEL = "openai/gpt-5-image";
+const OPENROUTER_AUTO_MODEL = "openrouter/auto";
+const FLUX_MODEL = "black-forest-labs/flux.2-pro";
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+test("OpenRouter request body uses image_config and string content for text-only prompts", () => {
+  const args = makeArgs({ aspectRatio: "16:9", quality: "2k" });
+  const body = buildRequestBody("hello", GEMINI_MODEL, args, []);
+
+  assert.deepEqual(body.image_config, {
+    image_size: "2K",
+    aspect_ratio: "16:9",
+  });
+  assert.deepEqual(body.provider, {
+    require_parameters: true,
+  });
+  assert.deepEqual(body.modalities, ["image", "text"]);
+  assert.equal(body.stream, false);
+  assert.equal(body.messages[0].content, "hello");
+});
+
+test("OpenRouter request body keeps text+image modalities for current text+image models", () => {
+  for (const model of [GEMINI_MODEL, GEMINI_25_MODEL, GPT_5_IMAGE_MODEL, OPENROUTER_AUTO_MODEL]) {
+    const body = buildRequestBody("hello", model, makeArgs({ quality: "2k" }), []);
+
+    assert.deepEqual(body.image_config, {
+      image_size: "2K",
+    });
+    assert.deepEqual(body.provider, {
+      require_parameters: true,
+    });
+    assert.deepEqual(body.modalities, ["image", "text"]);
+    assert.equal(body.messages[0].content, "hello");
+  }
+});
+
+test("OpenRouter request body uses image-only modalities for image-only models under CLI defaults", () => {
+  const body = buildRequestBody("hello", FLUX_MODEL, makeArgs({ quality: "2k" }), []);
+
+  assert.deepEqual(body.image_config, {
+    image_size: "2K",
+  });
+  assert.deepEqual(body.provider, {
+    require_parameters: true,
+  });
+  assert.deepEqual(body.modalities, ["image"]);
+  assert.equal(body.stream, false);
+  assert.equal(body.messages[0].content, "hello");
+});
+
+test("OpenRouter helper omits image_config when no size or quality is passed", () => {
+  const body = buildRequestBody("hello", FLUX_MODEL, makeArgs(), []);
+
+  assert.equal(body.image_config, undefined);
+  assert.equal(body.provider, undefined);
+  assert.deepEqual(body.modalities, ["image"]);
+  assert.equal(body.stream, false);
+  assert.equal(body.messages[0].content, "hello");
+});
+
+test("OpenRouter request body keeps multimodal array content when references are provided", () => {
+  const content = buildContent("hello", ["data:image/png;base64,abc"]);
+  assert.ok(Array.isArray(content));
+  assert.deepEqual(content[0], { type: "text", text: "hello" });
+  assert.deepEqual(content[1], {
+    type: "image_url",
+    image_url: { url: "data:image/png;base64,abc" },
+  });
+});
+
+test("OpenRouter size and aspect helpers infer supported values", () => {
+  assert.equal(getImageSize(makeArgs()), null);
+  assert.equal(getImageSize(makeArgs({ quality: "normal" })), "1K");
+  assert.equal(getImageSize(makeArgs({ size: "2048x1024" })), "2K");
+  assert.equal(getAspectRatio(GEMINI_MODEL, makeArgs({ size: "1600x900" })), "16:9");
+  assert.equal(getAspectRatio(GEMINI_MODEL, makeArgs({ size: "1024x4096" })), "1:4");
+  assert.equal(getAspectRatio(GEMINI_25_MODEL, makeArgs({ size: "1600x900" })), "16:9");
+  assert.equal(getAspectRatio(FLUX_MODEL, makeArgs({ size: "1024x4096" })), null);
+});
+
+test("OpenRouter validates explicit aspect ratios and inferred size ratios against model support", () => {
+  assert.doesNotThrow(() =>
+    validateArgs(GEMINI_MODEL, makeArgs({ aspectRatio: "1:4" })),
+  );
+  assert.doesNotThrow(() =>
+    validateArgs(GEMINI_MODEL, makeArgs({ size: "1024x4096" })),
+  );
+  assert.throws(
+    () => validateArgs(GEMINI_25_MODEL, makeArgs({ aspectRatio: "1:4" })),
+    /does not support aspect ratio 1:4/,
+  );
+  assert.throws(
+    () => validateArgs(FLUX_MODEL, makeArgs({ aspectRatio: "1:4" })),
+    /does not support aspect ratio 1:4/,
+  );
+  assert.throws(
+    () => validateArgs(GEMINI_MODEL, makeArgs({ size: "2048x1024" })),
+    /does not support size 2048x1024 \(aspect ratio 2:1\)/,
+  );
+});
+
+test("OpenRouter response extraction supports inline image data and finish_reason errors", async () => {
+  const bytes = await extractImageFromResponse({
+    choices: [
+      {
+        message: {
+          images: [
+            {
+              image_url: {
+                url: `data:image/png;base64,${Buffer.from("hello").toString("base64")}`,
+              },
+            },
+          ],
+        },
+      },
+    ],
+  });
+  assert.equal(Buffer.from(bytes).toString("utf8"), "hello");
+
+  await assert.rejects(
+    () =>
+      extractImageFromResponse({
+        choices: [
+          {
+            finish_reason: "error",
+            native_finish_reason: "MALFORMED_FUNCTION_CALL",
+            message: { content: null },
+          },
+        ],
+      }),
+    /finish_reason=MALFORMED_FUNCTION_CALL/,
+  );
+});

skills/baoyu-image-gen/scripts/providers/openrouter.ts

@@ -0,0 +1,369 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+import type { CliArgs } from "../types";
+
+const DEFAULT_MODEL = "google/gemini-3.1-flash-image-preview";
+const COMMON_ASPECT_RATIOS = [
+  "1:1",
+  "2:3",
+  "3:2",
+  "3:4",
+  "4:3",
+  "4:5",
+  "5:4",
+  "9:16",
+  "16:9",
+  "21:9",
+];
+const GEMINI_EXTENDED_ASPECT_RATIOS = ["1:4", "4:1", "1:8", "8:1"];
+
+type OpenRouterImageEntry = {
+  image_url?: string | { url?: string | null } | null;
+  imageUrl?: string | { url?: string | null } | null;
+};
+
+type OpenRouterMessagePart = {
+  type?: string;
+  text?: string;
+  image_url?: string | { url?: string | null } | null;
+  imageUrl?: string | { url?: string | null } | null;
+};
+
+type OpenRouterResponse = {
+  choices?: Array<{
+    finish_reason?: string | null;
+    native_finish_reason?: string | null;
+    message?: {
+      images?: OpenRouterImageEntry[];
+      content?: string | OpenRouterMessagePart[] | null;
+    };
+  }>;
+};
+
+export function getDefaultModel(): string {
+  return process.env.OPENROUTER_IMAGE_MODEL || DEFAULT_MODEL;
+}
+
+function normalizeModelId(model: string): string {
+  return model.trim().toLowerCase().split(":")[0]!;
+}
+
+function isTextAndImageModel(model: string): boolean {
+  const normalized = normalizeModelId(model);
+  if (normalized === "openrouter/auto") {
+    return true;
+  }
+
+  if (normalized.startsWith("google/gemini-") && normalized.includes("image")) {
+    return true;
+  }
+
+  if (normalized.startsWith("openai/gpt-") && normalized.includes("image")) {
+    return true;
+  }
+
+  return false;
+}
+
+function getSupportedAspectRatios(model: string): Set<string> {
+  const normalized = normalizeModelId(model);
+  if (normalized !== "google/gemini-3.1-flash-image-preview") {
+    return new Set(COMMON_ASPECT_RATIOS);
+  }
+
+  return new Set([...COMMON_ASPECT_RATIOS, ...GEMINI_EXTENDED_ASPECT_RATIOS]);
+}
+
+function getApiKey(): string | null {
+  return process.env.OPENROUTER_API_KEY || null;
+}
+
+function getBaseUrl(): string {
+  const base = process.env.OPENROUTER_BASE_URL || "https://openrouter.ai/api/v1";
+  return base.replace(/\/+$/g, "");
+}
+
+function getHeaders(apiKey: string): Record<string, string> {
+  const headers: Record<string, string> = {
+    "Content-Type": "application/json",
+    Authorization: `Bearer ${apiKey}`,
+  };
+
+  const referer = process.env.OPENROUTER_HTTP_REFERER?.trim();
+  if (referer) {
+    headers["HTTP-Referer"] = referer;
+  }
+
+  const title = process.env.OPENROUTER_TITLE?.trim();
+  if (title) {
+    headers["X-OpenRouter-Title"] = title;
+    headers["X-Title"] = title;
+  }
+
+  return headers;
+}
+
+function parsePixelSize(value: string): { width: number; height: number } | null {
+  const match = value.match(/^(\d+)\s*[xX]\s*(\d+)$/);
+  if (!match) return null;
+
+  const width = parseInt(match[1]!, 10);
+  const height = parseInt(match[2]!, 10);
+
+  if (!Number.isFinite(width) || !Number.isFinite(height) || width <= 0 || height <= 0) {
+    return null;
+  }
+
+  return { width, height };
+}
+
+function gcd(a: number, b: number): number {
+  let x = Math.abs(a);
+  let y = Math.abs(b);
+  while (y !== 0) {
+    const next = x % y;
+    x = y;
+    y = next;
+  }
+  return x || 1;
+}
+
+function inferAspectRatio(size: string | null): string | null {
+  if (!size) return null;
+  const parsed = parsePixelSize(size);
+  if (!parsed) return null;
+
+  const divisor = gcd(parsed.width, parsed.height);
+  return `${parsed.width / divisor}:${parsed.height / divisor}`;
+}
+
+function inferImageSize(size: string | null): "1K" | "2K" | "4K" | null {
+  if (!size) return null;
+  const parsed = parsePixelSize(size);
+  if (!parsed) return null;
+
+  const longestEdge = Math.max(parsed.width, parsed.height);
+  if (longestEdge <= 1024) return "1K";
+  if (longestEdge <= 2048) return "2K";
+  return "4K";
+}
+
+export function getImageSize(args: CliArgs): "1K" | "2K" | "4K" | null {
+  if (args.imageSize) return args.imageSize as "1K" | "2K" | "4K";
+
+  const inferredFromSize = inferImageSize(args.size);
+  if (inferredFromSize) return inferredFromSize;
+
+  if (args.quality === "normal") return "1K";
+  if (args.quality === "2k") return "2K";
+  return null;
+}
+
+export function getAspectRatio(model: string, args: CliArgs): string | null {
+  if (args.aspectRatio) return args.aspectRatio;
+
+  const inferred = inferAspectRatio(args.size);
+  if (!inferred || !getSupportedAspectRatios(model).has(inferred)) {
+    return null;
+  }
+
+  return inferred;
+}
+
+function getModalities(model: string): string[] {
+  return isTextAndImageModel(model) ? ["image", "text"] : ["image"];
+}
+
+export function validateArgs(model: string, args: CliArgs): void {
+  const requestedAspectRatio = args.aspectRatio || inferAspectRatio(args.size);
+  if (!requestedAspectRatio) {
+    return;
+  }
+
+  const supported = getSupportedAspectRatios(model);
+  if (supported.has(requestedAspectRatio)) {
+    return;
+  }
+
+  const requestedValue = args.aspectRatio
+    ? `aspect ratio ${requestedAspectRatio}`
+    : `size ${args.size} (aspect ratio ${requestedAspectRatio})`;
+
+  throw new Error(
+    `OpenRouter model ${model} does not support ${requestedValue}. Supported values: ${Array.from(supported).join(", ")}`
+  );
+}
+
+function getMimeType(filename: string): string {
+  const ext = path.extname(filename).toLowerCase();
+  if (ext === ".jpg" || ext === ".jpeg") return "image/jpeg";
+  if (ext === ".webp") return "image/webp";
+  if (ext === ".gif") return "image/gif";
+  return "image/png";
+}
+
+async function readImageAsDataUrl(filePath: string): Promise<string> {
+  const bytes = await readFile(filePath);
+  return `data:${getMimeType(filePath)};base64,${bytes.toString("base64")}`;
+}
+
+export function buildContent(
+  prompt: string,
+  referenceImages: string[],
+): string | Array<Record<string, unknown>> {
+  if (referenceImages.length === 0) {
+    return prompt;
+  }
+
+  const content: Array<Record<string, unknown>> = [{ type: "text", text: prompt }];
+
+  for (const imageUrl of referenceImages) {
+    content.push({
+      type: "image_url",
+      image_url: { url: imageUrl },
+    });
+  }
+
+  return content;
+}
+
+function extractImageUrl(entry: OpenRouterImageEntry | OpenRouterMessagePart): string | null {
+  const value = entry.image_url ?? entry.imageUrl;
+  if (!value) return null;
+  if (typeof value === "string") return value;
+  return value.url ?? null;
+}
+
+function decodeDataUrl(value: string): Uint8Array | null {
+  const match = value.match(/^data:image\/[^;]+;base64,([A-Za-z0-9+/=]+)$/);
+  if (!match) return null;
+  return Uint8Array.from(Buffer.from(match[1]!, "base64"));
+}
+
+async function downloadImage(value: string): Promise<Uint8Array> {
+  const inline = decodeDataUrl(value);
+  if (inline) return inline;
+
+  if (value.startsWith("http://") || value.startsWith("https://")) {
+    const response = await fetch(value);
+    if (!response.ok) {
+      throw new Error(`Failed to download OpenRouter image: ${response.status}`);
+    }
+    const buffer = await response.arrayBuffer();
+    return new Uint8Array(buffer);
+  }
+
+  return Uint8Array.from(Buffer.from(value, "base64"));
+}
+
+export async function extractImageFromResponse(result: OpenRouterResponse): Promise<Uint8Array> {
+  const choice = result.choices?.[0];
+  const message = choice?.message;
+
+  for (const image of message?.images ?? []) {
+    const imageUrl = extractImageUrl(image);
+    if (imageUrl) return downloadImage(imageUrl);
+  }
+
+  if (Array.isArray(message?.content)) {
+    for (const item of message.content) {
+      const imageUrl = extractImageUrl(item);
+      if (imageUrl) return downloadImage(imageUrl);
+
+      if (item.type === "text" && item.text) {
+        const inline = decodeDataUrl(item.text);
+        if (inline) return inline;
+      }
+    }
+  } else if (typeof message?.content === "string") {
+    const inline = decodeDataUrl(message.content);
+    if (inline) return inline;
+  }
+
+  const finishReason =
+    choice?.native_finish_reason || choice?.finish_reason || "unknown";
+  throw new Error(
+    `No image in OpenRouter response (finish_reason=${finishReason})`,
+  );
+}
+
+export function buildRequestBody(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+  referenceImages: string[],
+): Record<string, unknown> {
+  validateArgs(model, args);
+
+  const imageConfig: Record<string, string> = {};
+
+  const imageSize = getImageSize(args);
+  if (imageSize) {
+    imageConfig.image_size = imageSize;
+  }
+
+  const aspectRatio = getAspectRatio(model, args);
+  if (aspectRatio) {
+    imageConfig.aspect_ratio = aspectRatio;
+  }
+
+  const body: Record<string, unknown> = {
+    messages: [
+      {
+        role: "user",
+        content: buildContent(prompt, referenceImages),
+      },
+    ],
+    modalities: getModalities(model),
+    stream: false,
+  };
+
+  if (Object.keys(imageConfig).length > 0) {
+    body.image_config = imageConfig;
+    body.provider = {
+      require_parameters: true,
+    };
+  }
+
+  return body;
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const apiKey = getApiKey();
+  if (!apiKey) {
+    throw new Error("OPENROUTER_API_KEY is required. Get one at https://openrouter.ai/settings/keys");
+  }
+
+  const referenceImages: string[] = [];
+  for (const refPath of args.referenceImages) {
+    referenceImages.push(await readImageAsDataUrl(refPath));
+  }
+
+  const body = {
+    model,
+    ...buildRequestBody(prompt, model, args, referenceImages),
+  };
+
+  console.log(
+    `Generating image with OpenRouter (${model})...`,
+    (body.image_config as Record<string, string>),
+  );
+
+  const response = await fetch(`${getBaseUrl()}/chat/completions`, {
+    method: "POST",
+    headers: getHeaders(apiKey),
+    body: JSON.stringify(body),
+  });
+
+  if (!response.ok) {
+    const errorText = await response.text();
+    throw new Error(`OpenRouter API error (${response.status}): ${errorText}`);
+  }
+
+  const result = (await response.json()) as OpenRouterResponse;
+  return extractImageFromResponse(result);
+}

skills/baoyu-image-gen/scripts/providers/replicate.test.ts

@@ -0,0 +1,101 @@
+import assert from "node:assert/strict";
+import test from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  buildInput,
+  extractOutputUrl,
+  parseModelId,
+} from "./replicate.ts";
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+test("Replicate model parsing accepts official formats and rejects malformed ones", () => {
+  assert.deepEqual(parseModelId("google/nano-banana-pro"), {
+    owner: "google",
+    name: "nano-banana-pro",
+    version: null,
+  });
+  assert.deepEqual(parseModelId("owner/model:abc123"), {
+    owner: "owner",
+    name: "model",
+    version: "abc123",
+  });
+
+  assert.throws(
+    () => parseModelId("just-a-model-name"),
+    /Invalid Replicate model format/,
+  );
+});
+
+test("Replicate input builder maps aspect ratio, image count, quality, and refs", () => {
+  assert.deepEqual(
+    buildInput(
+      "A robot painter",
+      makeArgs({
+        aspectRatio: "16:9",
+        quality: "2k",
+        n: 3,
+      }),
+      ["data:image/png;base64,AAAA"],
+    ),
+    {
+      prompt: "A robot painter",
+      aspect_ratio: "16:9",
+      number_of_images: 3,
+      resolution: "2K",
+      output_format: "png",
+      image_input: ["data:image/png;base64,AAAA"],
+    },
+  );
+
+  assert.deepEqual(
+    buildInput("A robot painter", makeArgs({ quality: "normal" }), ["ref"]),
+    {
+      prompt: "A robot painter",
+      aspect_ratio: "match_input_image",
+      resolution: "1K",
+      output_format: "png",
+      image_input: ["ref"],
+    },
+  );
+});
+
+test("Replicate output extraction supports string, array, and object URLs", () => {
+  assert.equal(
+    extractOutputUrl({ output: "https://example.com/a.png" } as never),
+    "https://example.com/a.png",
+  );
+  assert.equal(
+    extractOutputUrl({ output: ["https://example.com/b.png"] } as never),
+    "https://example.com/b.png",
+  );
+  assert.equal(
+    extractOutputUrl({ output: { url: "https://example.com/c.png" } } as never),
+    "https://example.com/c.png",
+  );
+
+  assert.throws(
+    () => extractOutputUrl({ output: { invalid: true } } as never),
+    /Unexpected Replicate output format/,
+  );
+});

skills/baoyu-image-gen/scripts/providers/replicate.ts

@@ -0,0 +1,205 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+import type { CliArgs } from "../types";
+
+const DEFAULT_MODEL = "google/nano-banana-pro";
+const SYNC_WAIT_SECONDS = 60;
+const POLL_INTERVAL_MS = 2000;
+const MAX_POLL_MS = 300_000;
+
+export function getDefaultModel(): string {
+  return process.env.REPLICATE_IMAGE_MODEL || DEFAULT_MODEL;
+}
+
+function getApiToken(): string | null {
+  return process.env.REPLICATE_API_TOKEN || null;
+}
+
+function getBaseUrl(): string {
+  const base = process.env.REPLICATE_BASE_URL || "https://api.replicate.com";
+  return base.replace(/\/+$/g, "");
+}
+
+export function parseModelId(model: string): { owner: string; name: string; version: string | null } {
+  const [ownerName, version] = model.split(":");
+  const parts = ownerName!.split("/");
+  if (parts.length !== 2 || !parts[0] || !parts[1]) {
+    throw new Error(
+      `Invalid Replicate model format: "${model}". Expected "owner/name" or "owner/name:version".`
+    );
+  }
+  return { owner: parts[0], name: parts[1], version: version || null };
+}
+
+export function buildInput(prompt: string, args: CliArgs, referenceImages: string[]): Record<string, unknown> {
+  const input: Record<string, unknown> = { prompt };
+
+  if (args.aspectRatio) {
+    input.aspect_ratio = args.aspectRatio;
+  } else if (referenceImages.length > 0) {
+    input.aspect_ratio = "match_input_image";
+  }
+
+  if (args.n > 1) {
+    input.number_of_images = args.n;
+  }
+
+  if (args.quality === "normal") {
+    input.resolution = "1K";
+  } else if (args.quality === "2k") {
+    input.resolution = "2K";
+  }
+
+  input.output_format = "png";
+
+  if (referenceImages.length > 0) {
+    input.image_input = referenceImages;
+  }
+
+  return input;
+}
+
+async function readImageAsDataUrl(p: string): Promise<string> {
+  const buf = await readFile(p);
+  const ext = path.extname(p).toLowerCase();
+  let mimeType = "image/png";
+  if (ext === ".jpg" || ext === ".jpeg") mimeType = "image/jpeg";
+  else if (ext === ".gif") mimeType = "image/gif";
+  else if (ext === ".webp") mimeType = "image/webp";
+  return `data:${mimeType};base64,${buf.toString("base64")}`;
+}
+
+type PredictionResponse = {
+  id: string;
+  status: string;
+  output: unknown;
+  error: string | null;
+  urls?: { get?: string };
+};
+
+async function createPrediction(
+  apiToken: string,
+  model: { owner: string; name: string; version: string | null },
+  input: Record<string, unknown>,
+  sync: boolean
+): Promise<PredictionResponse> {
+  const baseUrl = getBaseUrl();
+
+  let url: string;
+  const body: Record<string, unknown> = { input };
+
+  if (model.version) {
+    url = `${baseUrl}/v1/predictions`;
+    body.version = model.version;
+  } else {
+    url = `${baseUrl}/v1/models/${model.owner}/${model.name}/predictions`;
+  }
+
+  const headers: Record<string, string> = {
+    Authorization: `Bearer ${apiToken}`,
+    "Content-Type": "application/json",
+  };
+
+  if (sync) {
+    headers["Prefer"] = `wait=${SYNC_WAIT_SECONDS}`;
+  }
+
+  const res = await fetch(url, {
+    method: "POST",
+    headers,
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const err = await res.text();
+    throw new Error(`Replicate API error (${res.status}): ${err}`);
+  }
+
+  return (await res.json()) as PredictionResponse;
+}
+
+async function pollPrediction(apiToken: string, getUrl: string): Promise<PredictionResponse> {
+  const start = Date.now();
+
+  while (Date.now() - start < MAX_POLL_MS) {
+    const res = await fetch(getUrl, {
+      headers: { Authorization: `Bearer ${apiToken}` },
+    });
+
+    if (!res.ok) {
+      const err = await res.text();
+      throw new Error(`Replicate poll error (${res.status}): ${err}`);
+    }
+
+    const prediction = (await res.json()) as PredictionResponse;
+
+    if (prediction.status === "succeeded") return prediction;
+    if (prediction.status === "failed" || prediction.status === "canceled") {
+      throw new Error(`Replicate prediction ${prediction.status}: ${prediction.error || "unknown error"}`);
+    }
+
+    await new Promise((r) => setTimeout(r, POLL_INTERVAL_MS));
+  }
+
+  throw new Error(`Replicate prediction timed out after ${MAX_POLL_MS / 1000}s`);
+}
+
+export function extractOutputUrl(prediction: PredictionResponse): string {
+  const output = prediction.output;
+
+  if (typeof output === "string") return output;
+
+  if (Array.isArray(output)) {
+    const first = output[0];
+    if (typeof first === "string") return first;
+  }
+
+  if (output && typeof output === "object" && "url" in output) {
+    const url = (output as Record<string, unknown>).url;
+    if (typeof url === "string") return url;
+  }
+
+  throw new Error(`Unexpected Replicate output format: ${JSON.stringify(output)}`);
+}
+
+async function downloadImage(url: string): Promise<Uint8Array> {
+  const res = await fetch(url);
+  if (!res.ok) throw new Error(`Failed to download image from Replicate: ${res.status}`);
+  const buf = await res.arrayBuffer();
+  return new Uint8Array(buf);
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs
+): Promise<Uint8Array> {
+  const apiToken = getApiToken();
+  if (!apiToken) throw new Error("REPLICATE_API_TOKEN is required. Get one at https://replicate.com/account/api-tokens");
+
+  const parsedModel = parseModelId(model);
+
+  const refDataUrls: string[] = [];
+  for (const refPath of args.referenceImages) {
+    refDataUrls.push(await readImageAsDataUrl(refPath));
+  }
+
+  const input = buildInput(prompt, args, refDataUrls);
+
+  console.log(`Generating image with Replicate (${model})...`);
+
+  let prediction = await createPrediction(apiToken, parsedModel, input, true);
+
+  if (prediction.status !== "succeeded") {
+    if (!prediction.urls?.get) {
+      throw new Error("Replicate prediction did not return a poll URL");
+    }
+    console.log("Waiting for prediction to complete...");
+    prediction = await pollPrediction(apiToken, prediction.urls.get);
+  }
+
+  console.log("Generation completed.");
+
+  const outputUrl = extractOutputUrl(prediction);
+  return downloadImage(outputUrl);
+}

skills/baoyu-image-gen/scripts/providers/seedream.test.ts

@@ -0,0 +1,244 @@
+import assert from "node:assert/strict";
+import fs from "node:fs/promises";
+import os from "node:os";
+import path from "node:path";
+import test, { type TestContext } from "node:test";
+
+import type { CliArgs } from "../types.ts";
+import {
+  buildImageInput,
+  buildRequestBody,
+  generateImage,
+  getDefaultOutputExtension,
+  resolveSeedreamSize,
+  validateArgs,
+} from "./seedream.ts";
+
+function makeArgs(overrides: Partial<CliArgs> = {}): CliArgs {
+  return {
+    prompt: null,
+    promptFiles: [],
+    imagePath: null,
+    provider: null,
+    model: null,
+    aspectRatio: null,
+    size: null,
+    quality: null,
+    imageSize: null,
+    referenceImages: [],
+    n: 1,
+    batchFile: null,
+    jobs: null,
+    json: false,
+    help: false,
+    ...overrides,
+  };
+}
+
+function useEnv(
+  t: TestContext,
+  values: Record<string, string | null>,
+): void {
+  const previous = new Map<string, string | undefined>();
+  for (const [key, value] of Object.entries(values)) {
+    previous.set(key, process.env[key]);
+    if (value == null) {
+      delete process.env[key];
+    } else {
+      process.env[key] = value;
+    }
+  }
+
+  t.after(() => {
+    for (const [key, value] of previous.entries()) {
+      if (value == null) {
+        delete process.env[key];
+      } else {
+        process.env[key] = value;
+      }
+    }
+  });
+}
+
+async function makeTempPng(t: TestContext, name: string): Promise<string> {
+  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "seedream-test-"));
+  t.after(() => fs.rm(dir, { recursive: true, force: true }));
+
+  const filePath = path.join(dir, name);
+  const png1x1 =
+    "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mP8/x8AAwMCAO+a7m0AAAAASUVORK5CYII=";
+  await fs.writeFile(filePath, Buffer.from(png1x1, "base64"));
+  return filePath;
+}
+
+test("Seedream request body and default extensions follow official model capabilities", () => {
+  const five = buildRequestBody(
+    "A robot illustrator",
+    "doubao-seedream-5-0-260128",
+    makeArgs(),
+  );
+  assert.equal(five.size, "2K");
+  assert.equal(five.response_format, "url");
+  assert.equal(five.output_format, "png");
+  assert.equal(getDefaultOutputExtension("doubao-seedream-5-0-260128"), ".png");
+
+  const fourFive = buildRequestBody(
+    "A robot illustrator",
+    "doubao-seedream-4-5-251128",
+    makeArgs(),
+  );
+  assert.equal(fourFive.size, "2K");
+  assert.equal(fourFive.response_format, "url");
+  assert.ok(!("output_format" in fourFive));
+  assert.equal(getDefaultOutputExtension("doubao-seedream-4-5-251128"), ".jpg");
+
+  assert.throws(
+    () =>
+      buildRequestBody(
+        "Change the bubbles into hearts",
+        "doubao-seededit-3-0-i2i-250628",
+        makeArgs({ referenceImages: ["ref.png"] }),
+        "data:image/png;base64,AAAA",
+      ),
+    /no longer supported/,
+  );
+});
+
+test("Seedream size selection validates model-specific presets", () => {
+  assert.equal(
+    resolveSeedreamSize("doubao-seedream-4-0-250828", makeArgs({ quality: "normal" })),
+    "1K",
+  );
+  assert.equal(
+    resolveSeedreamSize("doubao-seedream-3-0-t2i-250415", makeArgs({ quality: "2k" })),
+    "2048x2048",
+  );
+
+  assert.throws(
+    () =>
+      resolveSeedreamSize("doubao-seedream-5-0-260128", makeArgs({ size: "4K" })),
+    /only supports 2K, 3K/,
+  );
+  assert.throws(
+    () =>
+      resolveSeedreamSize("doubao-seedream-3-0-t2i-250415", makeArgs({ imageSize: "2K" })),
+    /only supports explicit WxH sizes/,
+  );
+  assert.throws(
+    () =>
+      resolveSeedreamSize("doubao-seededit-3-0-i2i-250628", makeArgs({ size: "1024x1024" })),
+    /no longer supported/,
+  );
+});
+
+test("Seedream reference-image support is model-specific", () => {
+  assert.doesNotThrow(() =>
+    validateArgs(
+      "doubao-seedream-5-0-260128",
+      makeArgs({ referenceImages: ["a.png", "b.png"] }),
+    ),
+  );
+
+  assert.throws(
+    () =>
+      validateArgs(
+        "doubao-seedream-3-0-t2i-250415",
+        makeArgs({ referenceImages: ["a.png"] }),
+      ),
+    /does not support reference images/,
+  );
+
+  assert.throws(
+    () =>
+      validateArgs(
+        "doubao-seededit-3-0-i2i-250628",
+        makeArgs(),
+      ),
+    /no longer supported/,
+  );
+
+  assert.throws(
+    () =>
+      validateArgs(
+        "ep-20260315171508-t8br2",
+        makeArgs({ referenceImages: ["a.png"] }),
+      ),
+    /require a known model ID/,
+  );
+});
+
+test("Seedream image input encodes local references as data URLs", async (t) => {
+  const refOne = await makeTempPng(t, "one.png");
+  const refTwo = await makeTempPng(t, "two.png");
+
+  const single = await buildImageInput("doubao-seedream-4-5-251128", [refOne]);
+  assert.match(String(single), /^data:image\/png;base64,/);
+
+  const multiple = await buildImageInput("doubao-seedream-5-0-260128", [refOne, refTwo]);
+  assert.ok(Array.isArray(multiple));
+  assert.equal(multiple.length, 2);
+});
+
+test("Seedream generateImage posts the documented response_format and downloads the returned URL", async (t) => {
+  useEnv(t, { ARK_API_KEY: "test-key", SEEDREAM_BASE_URL: null });
+
+  const originalFetch = globalThis.fetch;
+  t.after(() => {
+    globalThis.fetch = originalFetch;
+  });
+
+  const calls: Array<{
+    input: string;
+    init?: RequestInit;
+  }> = [];
+
+  globalThis.fetch = async (input, init) => {
+    calls.push({
+      input: String(input),
+      init,
+    });
+
+    if (calls.length === 1) {
+      return Response.json({
+        model: "doubao-seedream-4-5-251128",
+        created: 1740000000,
+        data: [
+          {
+            url: "https://example.com/generated-image",
+            size: "2048x2048",
+          },
+        ],
+        usage: {
+          generated_images: 1,
+          output_tokens: 1,
+          total_tokens: 1,
+        },
+      });
+    }
+
+    return new Response(Uint8Array.from([7, 8, 9]), {
+      status: 200,
+      headers: { "Content-Type": "image/jpeg" },
+    });
+  };
+
+  const image = await generateImage(
+    "A robot illustrator",
+    "doubao-seedream-4-5-251128",
+    makeArgs(),
+  );
+
+  assert.deepEqual([...image], [7, 8, 9]);
+  assert.equal(calls.length, 2);
+  assert.equal(
+    calls[0]?.input,
+    "https://ark.cn-beijing.volces.com/api/v3/images/generations",
+  );
+
+  const requestBody = JSON.parse(String(calls[0]?.init?.body)) as Record<string, unknown>;
+  assert.equal(requestBody.model, "doubao-seedream-4-5-251128");
+  assert.equal(requestBody.size, "2K");
+  assert.equal(requestBody.response_format, "url");
+  assert.ok(!("output_format" in requestBody));
+  assert.equal(calls[1]?.input, "https://example.com/generated-image");
+});

skills/baoyu-image-gen/scripts/providers/seedream.ts

@@ -0,0 +1,341 @@
+import path from "node:path";
+import { readFile } from "node:fs/promises";
+
+import type { CliArgs } from "../types";
+
+export type SeedreamModelFamily =
+  | "seedream5"
+  | "seedream45"
+  | "seedream40"
+  | "seedream30"
+  | "unknown";
+
+type SeedreamRequestImage = string | string[];
+
+type SeedreamRequestBody = {
+  model: string;
+  prompt: string;
+  size: string;
+  response_format: "url";
+  watermark: boolean;
+  image?: SeedreamRequestImage;
+  output_format?: "png";
+};
+
+type SeedreamImageResponse = {
+  model?: string;
+  created?: number;
+  data?: Array<{
+    url?: string;
+    b64_json?: string;
+    size?: string;
+    error?: {
+      code?: string;
+      message?: string;
+    };
+  }>;
+  usage?: {
+    generated_images: number;
+    output_tokens: number;
+    total_tokens: number;
+  };
+  error?: {
+    code?: string;
+    message?: string;
+  };
+};
+
+export function getDefaultModel(): string {
+  return process.env.SEEDREAM_IMAGE_MODEL || "doubao-seedream-5-0-260128";
+}
+
+function getApiKey(): string | null {
+  return process.env.ARK_API_KEY || null;
+}
+
+function getBaseUrl(): string {
+  return process.env.SEEDREAM_BASE_URL || "https://ark.cn-beijing.volces.com/api/v3";
+}
+
+function parsePixelSize(value: string): { width: number; height: number } | null {
+  const match = value.trim().match(/^(\d+)\s*[xX]\s*(\d+)$/);
+  if (!match) return null;
+
+  const width = parseInt(match[1]!, 10);
+  const height = parseInt(match[2]!, 10);
+  if (!Number.isFinite(width) || !Number.isFinite(height) || width <= 0 || height <= 0) {
+    return null;
+  }
+
+  return { width, height };
+}
+
+function normalizePixelSize(value: string): string | null {
+  const parsed = parsePixelSize(value);
+  if (!parsed) return null;
+  return `${parsed.width}x${parsed.height}`;
+}
+
+function normalizeSizePreset(value: string): string | null {
+  const upper = value.trim().toUpperCase();
+  if (upper === "ADAPTIVE") return "adaptive";
+  if (upper === "1K" || upper === "2K" || upper === "3K" || upper === "4K") return upper;
+  return null;
+}
+
+function normalizeSizeValue(value: string): string | null {
+  return normalizeSizePreset(value) ?? normalizePixelSize(value);
+}
+
+function getMimeType(filename: string): string {
+  const ext = path.extname(filename).toLowerCase();
+  if (ext === ".jpg" || ext === ".jpeg") return "image/jpeg";
+  if (ext === ".webp") return "image/webp";
+  if (ext === ".gif") return "image/gif";
+  if (ext === ".bmp") return "image/bmp";
+  if (ext === ".tiff" || ext === ".tif") return "image/tiff";
+  return "image/png";
+}
+
+async function readImageAsDataUrl(filePath: string): Promise<string> {
+  const bytes = await readFile(filePath);
+  return `data:${getMimeType(filePath)};base64,${bytes.toString("base64")}`;
+}
+
+export function getModelFamily(model: string): SeedreamModelFamily {
+  const normalized = model.trim();
+  if (/^doubao-seedream-5-0(?:-lite)?-\d+$/.test(normalized)) return "seedream5";
+  if (/^doubao-seedream-4-5-\d+$/.test(normalized)) return "seedream45";
+  if (/^doubao-seedream-4-0-\d+$/.test(normalized)) return "seedream40";
+  if (/^doubao-seedream-3-0-t2i-\d+$/.test(normalized)) return "seedream30";
+  return "unknown";
+}
+
+function isRemovedSeededitModel(model: string): boolean {
+  return /^doubao-seededit-3-0-i2i-\d+$/.test(model.trim());
+}
+
+function assertSupportedModel(model: string): void {
+  if (isRemovedSeededitModel(model)) {
+    throw new Error(
+      `${model} is no longer supported. SeedEdit 3.0 support has been removed from this tool; use Seedream 5.0/4.5/4.0/3.0 instead.`
+    );
+  }
+}
+
+export function supportsReferenceImages(model: string): boolean {
+  const family = getModelFamily(model);
+  return family === "seedream5" || family === "seedream45" || family === "seedream40";
+}
+
+function supportsOutputFormat(model: string): boolean {
+  return getModelFamily(model) === "seedream5";
+}
+
+export function getDefaultOutputExtension(model: string): ".png" | ".jpg" {
+  assertSupportedModel(model);
+  return supportsOutputFormat(model) ? ".png" : ".jpg";
+}
+
+export function getDefaultSeedreamSize(model: string, args: CliArgs): string {
+  assertSupportedModel(model);
+  const family = getModelFamily(model);
+
+  if (family === "seedream5") return "2K";
+  if (family === "seedream45") return "2K";
+  if (family === "seedream40") return args.quality === "normal" ? "1K" : "2K";
+  if (family === "seedream30") return args.quality === "2k" ? "2048x2048" : "1024x1024";
+  return "2K";
+}
+
+export function resolveSeedreamSize(model: string, args: CliArgs): string {
+  assertSupportedModel(model);
+  const family = getModelFamily(model);
+  const requested = args.size || args.imageSize || null;
+  const normalized = requested ? normalizeSizeValue(requested) : null;
+
+  if (!normalized) {
+    return getDefaultSeedreamSize(model, args);
+  }
+
+  if (family === "seedream30") {
+    const pixelSize = normalizePixelSize(normalized);
+    if (!pixelSize) {
+      throw new Error("Seedream 3.0 only supports explicit WxH sizes such as 1024x1024.");
+    }
+    return pixelSize;
+  }
+
+  if (family === "seedream5") {
+    if (normalized === "4K" || normalized === "1K" || normalized === "adaptive") {
+      throw new Error("Seedream 5.0 only supports 2K, 3K, or explicit WxH sizes.");
+    }
+    return normalized;
+  }
+
+  if (family === "seedream45") {
+    if (normalized === "1K" || normalized === "3K" || normalized === "adaptive") {
+      throw new Error("Seedream 4.5 only supports 2K, 4K, or explicit WxH sizes.");
+    }
+    return normalized;
+  }
+
+  if (family === "seedream40") {
+    if (normalized === "3K" || normalized === "adaptive") {
+      throw new Error("Seedream 4.0 only supports 1K, 2K, 4K, or explicit WxH sizes.");
+    }
+    return normalized;
+  }
+
+  if (normalized === "adaptive") {
+    throw new Error("Adaptive size is not supported by Seedream image generation.");
+  }
+
+  if (normalized === "1K" || normalized === "3K" || normalized === "4K") {
+    throw new Error(
+      "Unknown Seedream model ID. Use a documented model ID or pass an explicit WxH size instead of preset imageSize."
+    );
+  }
+
+  return normalized;
+}
+
+export function validateArgs(model: string, args: CliArgs): void {
+  assertSupportedModel(model);
+  const family = getModelFamily(model);
+  const refCount = args.referenceImages.length;
+
+  if (refCount === 0) {
+    resolveSeedreamSize(model, args);
+    return;
+  }
+
+  if (family === "unknown") {
+    throw new Error(
+      "Reference images with Seedream require a known model ID. Use Seedream 5.0/4.5/4.0 model IDs instead of an endpoint ID."
+    );
+  }
+
+  if (!supportsReferenceImages(model)) {
+    throw new Error(`${model} does not support reference images.`);
+  }
+
+  if ((family === "seedream5" || family === "seedream45" || family === "seedream40") && refCount > 14) {
+    throw new Error(`${model} supports at most 14 reference images.`);
+  }
+
+  resolveSeedreamSize(model, args);
+}
+
+export async function buildImageInput(
+  model: string,
+  referenceImages: string[],
+): Promise<SeedreamRequestImage | undefined> {
+  if (referenceImages.length === 0) return undefined;
+  assertSupportedModel(model);
+
+  const encoded = await Promise.all(referenceImages.map((refPath) => readImageAsDataUrl(refPath)));
+
+  return encoded.length === 1 ? encoded[0]! : encoded;
+}
+
+export function buildRequestBody(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+  imageInput?: SeedreamRequestImage,
+): SeedreamRequestBody {
+  validateArgs(model, args);
+
+  const requestBody: SeedreamRequestBody = {
+    model,
+    prompt,
+    size: resolveSeedreamSize(model, args),
+    response_format: "url",
+    watermark: false,
+  };
+
+  if (imageInput) {
+    requestBody.image = imageInput;
+  }
+
+  if (supportsOutputFormat(model)) {
+    requestBody.output_format = "png";
+  }
+
+  return requestBody;
+}
+
+async function downloadImage(url: string): Promise<Uint8Array> {
+  const imgResponse = await fetch(url);
+  if (!imgResponse.ok) {
+    throw new Error(`Failed to download image from ${url}`);
+  }
+
+  const buffer = await imgResponse.arrayBuffer();
+  return new Uint8Array(buffer);
+}
+
+export async function extractImageFromResponse(result: SeedreamImageResponse): Promise<Uint8Array> {
+  const first = result.data?.find((item) => item.url || item.b64_json || item.error);
+
+  if (!first) {
+    throw new Error("No image data in Seedream response");
+  }
+
+  if (first.error) {
+    throw new Error(first.error.message || "Seedream returned an image generation error");
+  }
+
+  if (first.b64_json) {
+    return Uint8Array.from(Buffer.from(first.b64_json, "base64"));
+  }
+
+  if (first.url) {
+    console.error(`Downloading image from ${first.url}...`);
+    return downloadImage(first.url);
+  }
+
+  throw new Error("No image URL or base64 data in Seedream response");
+}
+
+export async function generateImage(
+  prompt: string,
+  model: string,
+  args: CliArgs,
+): Promise<Uint8Array> {
+  const apiKey = getApiKey();
+  if (!apiKey) {
+    throw new Error(
+      "ARK_API_KEY is required. " +
+        "Get your API key from https://console.volcengine.com/ark"
+    );
+  }
+
+  validateArgs(model, args);
+  const imageInput = await buildImageInput(model, args.referenceImages);
+  const requestBody = buildRequestBody(prompt, model, args, imageInput);
+
+  console.error(`Calling Seedream API (${model}) with size: ${requestBody.size}`);
+
+  const response = await fetch(`${getBaseUrl()}/images/generations`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify(requestBody),
+  });
+
+  if (!response.ok) {
+    const err = await response.text();
+    throw new Error(`Seedream API error (${response.status}): ${err}`);
+  }
+
+  const result = (await response.json()) as SeedreamImageResponse;
+  if (result.error) {
+    throw new Error(result.error.message || "Seedream API returned an error");
+  }
+
+  return extractImageFromResponse(result);
+}

skills/baoyu-image-gen/scripts/types.ts

@@ -0,0 +1,82 @@
+export type Provider =
+  | "google"
+  | "openai"
+  | "openrouter"
+  | "dashscope"
+  | "minimax"
+  | "replicate"
+  | "jimeng"
+  | "seedream"
+  | "azure";
+export type Quality = "normal" | "2k";
+
+export type CliArgs = {
+  prompt: string | null;
+  promptFiles: string[];
+  imagePath: string | null;
+  provider: Provider | null;
+  model: string | null;
+  aspectRatio: string | null;
+  size: string | null;
+  quality: Quality | null;
+  imageSize: string | null;
+  referenceImages: string[];
+  n: number;
+  batchFile: string | null;
+  jobs: number | null;
+  json: boolean;
+  help: boolean;
+};
+
+export type BatchTaskInput = {
+  id?: string;
+  prompt?: string | null;
+  promptFiles?: string[];
+  image?: string;
+  provider?: Provider | null;
+  model?: string | null;
+  ar?: string | null;
+  size?: string | null;
+  quality?: Quality | null;
+  imageSize?: "1K" | "2K" | "4K" | null;
+  ref?: string[];
+  n?: number;
+};
+
+export type BatchFile =
+  | BatchTaskInput[]
+  | {
+      tasks: BatchTaskInput[];
+      jobs?: number | null;
+    };
+
+export type ExtendConfig = {
+  version: number;
+  default_provider: Provider | null;
+  default_quality: Quality | null;
+  default_aspect_ratio: string | null;
+  default_image_size: "1K" | "2K" | "4K" | null;
+  default_model: {
+    google: string | null;
+    openai: string | null;
+    openrouter: string | null;
+    dashscope: string | null;
+    minimax: string | null;
+    replicate: string | null;
+    jimeng: string | null;
+    seedream: string | null;
+    azure: string | null;
+  };
+  batch?: {
+    max_workers?: number | null;
+    provider_limits?: Partial<
+      Record<
+        Provider,
+        {
+          concurrency?: number | null;
+          start_interval_ms?: number | null;
+        }
+      >
+    >;
+  };
+};

skills/baoyu-translate/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: baoyu-translate
 description: Translates articles and documents between languages with three modes - quick (direct), normal (analyze then translate), and refined (analyze, translate, review, polish). Supports custom glossaries and terminology consistency via EXTEND.md. Use when user asks to "translate", "翻译", "精翻", "translate article", "translate to Chinese/English", "改成中文", "改成英文", "convert to Chinese", "localize", "本地化", or needs any document translation. Also triggers for "refined translation", "精细翻译", "proofread translation", "快速翻译", "快翻", "这篇文章翻译一下", or when a URL or file is provided with translation intent.
-version: 1.56.1
+version: 1.59.0
 metadata:
   openclaw:
     homepage: https://github.com/JimLiu/baoyu-skills#baoyu-translate
@@ -189,12 +189,12 @@ Before translating chunks:
    - Splits at markdown block boundaries to preserve structure
    - If a single block exceeds the threshold, falls back to line splitting, then word splitting
 4. **Assemble translation prompt**:
-   - 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 the resolved style preset (from `--style` flag, EXTEND.md `style` setting, or default `storytelling`), content background, merged glossary, and comprehension challenges
+   - 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
    - Save as `02-prompt.md` in the output directory (shared context only, no task instructions)
 5. **Draft translation via subagents** (if Agent tool available):
    - Spawn one subagent **per chunk**, all in parallel (Part 2 of the template)
-   - Each subagent reads `02-prompt.md` for shared context, translates its chunk, saves to `chunks/chunk-NN-draft.md`
-   - Terminology consistency is guaranteed by the shared `02-prompt.md` (glossary + comprehension challenges from analysis)
+   - 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`
+   - Consistency is guaranteed by the shared `02-prompt.md` (glossary, figurative language mapping, comprehension challenges, source voice, and translation challenges from analysis)
    - If no chunks (content under threshold): spawn one subagent for the entire source file
    - If Agent tool is unavailable, translate chunks sequentially inline using `02-prompt.md`
 6. **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)
@@ -206,26 +206,22 @@ Before translating chunks:
 
 **Translation principles** (apply to all modes):
 
+- **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?"
 - **Accuracy first**: Facts, data, and logic must match the original exactly
-- **Meaning over words**: Translate what the author means, not just what the words say. When a literal translation sounds unnatural or fails to convey the intended effect, restructure freely to express the same meaning in idiomatic target language
-- **Figurative language**: Interpret metaphors, idioms, and figurative expressions by their intended meaning rather than translating them word-for-word. When a source-language image does not carry the same connotation in the target language, replace it with a natural expression that conveys the same idea and emotional effect
-- **Emotional fidelity**: Preserve the emotional connotations of word choices, not just their dictionary meanings. Words that carry subjective feelings (e.g., "alarming", "haunting") should be rendered to evoke the same response in target-language readers
-- **Natural flow**: Use idiomatic target language word order and sentence patterns; break or restructure sentences freely when the source structure doesn't work naturally in the target language
-- **Terminology**: Use standard translations; annotate with original term in parentheses on first occurrence
+- **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
+- **Terminology**: Use standard translations consistently. First occurrence of specialized terms: annotate with original in parentheses
 - **Preserve format**: Keep all markdown formatting (headings, bold, italic, images, links, code blocks)
-- **Image-language awareness**: Preserve image references exactly during translation, but after the translation is complete, review referenced images and check whether their likely main text language still matches the translated article language
-- **Frontmatter transformation**: If the source has YAML frontmatter, preserve it in the translation with these changes: (1) Rename metadata fields that describe the *source* article — `url`→`sourceUrl`, `title`→`sourceTitle`, `description`→`sourceDescription`, `author`→`sourceAuthor`, `date`→`sourceDate`, and any similar origin-metadata fields — by adding a `source` prefix (camelCase). (2) Translate the values of text fields (title, description, etc.) and add them as new top-level fields. (3) Keep other fields (tags, categories, custom fields) as-is, translating their values where appropriate
-- **Respect original**: Maintain original meaning and intent; do not add, remove, or editorialize — but sentence structure and imagery may be adapted freely to serve the meaning
-- **Translator's notes**: For terms, concepts, or cultural references that target readers may not understand — due to jargon, cultural gaps, or domain-specific knowledge — add a concise explanatory note in parentheses immediately after the term. The note should explain *what it means* in plain language, not just provide the English original. Format: `译文（English original，通俗解释）`. Calibrate annotation depth to the target audience: general readers need more notes than technical readers. For short texts (< 5 sentences), further reduce annotations — only annotate non-common terms that the target audience is unlikely to know; skip terms that are widely recognized or self-explanatory in context. Only add notes where genuinely needed; do not over-annotate obvious terms.
+- **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
+- **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
 
 #### Quick Mode
 
-Translate directly → save to `translation.md`. No analysis file, but still apply all translation principles above — especially: interpret figurative language by meaning (not word-for-word), preserve emotional connotations, and restructure sentences for natural target-language flow.
+Translate directly → save to `translation.md`. Apply all translation principles above.
 
 #### Normal Mode
 
-1. **Analyze** → `01-analysis.md` (domain, tone, audience, terminology, reader comprehension challenges, figurative language & metaphor mapping)
-2. **Assemble prompt** → `02-prompt.md` (translation instructions with inlined style preset, content background, glossary, and comprehension challenges)
+1. **Analyze** → `01-analysis.md` (domain, tone, terminology, translation challenges)
+2. **Assemble prompt** → `02-prompt.md` (translation instructions with context, glossary, challenges)
 3. **Translate** (following `02-prompt.md`) → `translation.md`
 
 After completion, prompt user: "Translation saved. To further review and polish, reply **继续润色** or **refine**."
@@ -239,7 +235,7 @@ Full workflow for publication quality. See [references/refined-workflow.md](refe
 The 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.
 
 Steps and saved files (all in output directory):
-1. **Analyze** → `01-analysis.md` (domain, tone, terminology, reader comprehension challenges, figurative language & metaphor mapping)
+1. **Analyze** → `01-analysis.md` (domain, tone, terminology, translation challenges)
 2. **Assemble prompt** → `02-prompt.md` (translation instructions with inlined context)
 3. **Draft** → `03-draft.md` (initial translation with translator's notes; from subagent if chunked)
 4. **Critical review** → `04-critique.md` (diagnosis only: accuracy, Europeanized language, strategy execution, expression issues)

skills/baoyu-translate/references/refined-workflow.md

@@ -11,95 +11,38 @@ All intermediate results are saved as files in the output directory.
 
 ## Step 1: Content Analysis
 
-Before translating, deeply analyze the source material. Save analysis to `01-analysis.md` in the output directory. Focus on dimensions that directly inform translation quality.
+Before translating, analyze the source material. Save analysis to `01-analysis.md` in the output directory.
 
-### 1.1 Quick Summary
+### 1.1 Content Summary
 
-3-5 sentences capturing:
-- What is this content about?
-- What is the core argument?
-- What is the most valuable point?
+- What is this content about? What is the core argument?
+- Author background, stance, and writing context
+- Purpose and intended audience of the original
 
-### 1.2 Core Content
+### 1.2 Terminology
 
-- **Core argument**: One sentence summary
-- **Key concepts**: What key concepts does the author use? How are they defined?
-- **Structure**: How is the argument developed? How do sections connect?
-- **Evidence**: What specific examples, data, or authoritative citations are used?
-
-### 1.3 Background Context
-
-- **Author**: Who is the author? What is their background and stance?
-- **Writing context**: What phenomenon, trend, or debate is this responding to?
-- **Purpose**: What problem is the author trying to solve? Who are they trying to influence?
-- **Implicit assumptions**: What unstated premises underlie the argument?
-
-### 1.4 Terminology Extraction
-
-- List all technical terms, proper nouns, brand names, acronyms
+- List technical terms, proper nouns, brand names, acronyms
 - Cross-reference with loaded glossaries
-- For terms not in glossary, research standard translations
-- Record decisions in a working terminology table
+- For terms not in glossary, determine standard translations
+- Record in a terminology table
 
-### 1.5 Tone & Style
+### 1.3 Tone & Style
 
-- Is the original formal or conversational?
-- Does it use humor, metaphor, or cultural references?
+- Formal or conversational? Humor, metaphor, cultural references?
 - What register is appropriate for the translation given the target audience?
 
-### 1.6 Reader Comprehension Challenges
+### 1.4 Translation Challenges
 
-Identify points where target readers may struggle, calibrated to the target audience:
+Identify what may cause difficulty in translation:
 
-- **Domain jargon**: Technical terms that lack widely-known translations or are meaningless when translated literally
-- **Cultural references**: Idioms, historical events, pop culture, social norms specific to the source culture
-- **Implicit knowledge**: Background context the original author assumes but target readers may lack
-- **Wordplay & metaphors**: Figurative language that doesn't carry over across languages
-- **Named concepts**: Theories, effects, or phenomena with coined names (e.g., "comb-over effect", "Dunning-Kruger effect")
-- **Cognitive gaps**: Counterintuitive claims or expectations vs. reality that need framing for target readers
-
-For each identified challenge, note:
-1. The original term/passage
-2. Why it may confuse target readers
-3. A concise plain-language explanation to use as a translator's note
-
-### 1.7 Figurative Language & Metaphor Mapping
-
-Identify all metaphors, similes, idioms, and figurative expressions in the source. For each:
-
-1. **Original expression**: The exact phrase
-2. **Intended meaning**: What the author is actually communicating (the idea behind the image)
-3. **Literal translation risk**: Would a word-for-word translation sound unnatural, lose the connotation, or confuse target readers?
-4. **Target-language approach**: One of:
-   - **Interpret**: Discard the source image entirely, express the intended meaning directly in natural target language
-   - **Substitute**: Replace with a target-language idiom or image that conveys the same idea and emotional effect
-   - **Retain**: Keep the original image if it works equally well in the target language
-
-Also flag:
-- **Emotional connotations carried by word choice**: Words like "alarming" that convey subjective feeling, not just objective description — note the emotional effect to preserve
-- **Implied meanings**: Sentences where the surface meaning is simple but the implication is richer — note what the author really means so the translator can convey the full intent
-
-### 1.8 Structural & Creative Challenges
-
-- Complex sentence patterns (long subordinate clauses, nested modifiers, participial phrases) that need restructuring for natural target-language flow
-- Structural challenges (wordplay, ambiguity, puns that don't translate)
-- Content where the author's voice or humor requires creative adaptation
+- **Comprehension gaps**: Terms or references that target readers may not understand — note what explanation is needed
+- **Figurative language**: Metaphors, idioms, expressions that don't translate literally — note intended meaning and target-language approach (interpret / substitute / retain)
+- **Structural challenges**: Long complex sentences, wordplay, puns, or humor that needs creative adaptation
 
 **Save `01-analysis.md`** with:
 ```
-## Quick Summary
-[3-5 sentences]
-
-## Core Content
-Core argument: [one sentence]
-Key concepts: [list]
-Structure: [outline]
-
-## Background Context
-Author: [who, background, stance]
-Writing context: [what this responds to]
-Purpose: [goal and target audience]
-Implicit assumptions: [unstated premises]
+## Content Summary
+[Core argument, author, context, purpose]
 
 ## Terminology
 [term → translation, ...]
@@ -107,23 +50,21 @@ Implicit assumptions: [unstated premises]
 ## Tone & Style
 [assessment]
 
-## Comprehension Challenges
-- [term/passage] → [why confusing] → [proposed note]
+## Translation Challenges
+- [term/passage] → [challenge type] → [suggested approach]
 - ...
-
-## Figurative Language & Metaphor Mapping
-- [original expression] → [intended meaning] → [approach: interpret/substitute/retain] → [suggested rendering]
-- ...
-
-## Structural & Creative Challenges
-[sentence restructuring needs, wordplay, creative adaptation needs]
 ```
 
 ## Step 2: Assemble Translation Prompt
 
-Main agent reads `01-analysis.md` and assembles a complete translation prompt using [references/subagent-prompt-template.md](subagent-prompt-template.md). Inline the resolved style preset (from `--style` flag, EXTEND.md `style` setting, or default `storytelling`), content background, merged glossary, and comprehension challenges into the prompt. Save to `02-prompt.md`.
+Main agent reads `01-analysis.md` and assembles a complete translation prompt using [references/subagent-prompt-template.md](subagent-prompt-template.md). Inline the following from analysis:
 
-This prompt is used by the subagent (chunked) or by the main agent itself (non-chunked).
+- **Target style**: Resolved style preset + source voice assessment from §1.3
+- **Content background**: Summary from §1.1
+- **Glossary**: Merged glossary with analysis-extracted terms from §1.2
+- **Translation challenges**: All challenges from §1.4
+
+Save to `02-prompt.md`. This prompt is used by the subagent (chunked) or by the main agent itself (non-chunked).
 
 ## Step 3: Initial Draft
 
@@ -131,111 +72,54 @@ Save to `03-draft.md` in the output directory.
 
 For chunked content, the subagent produces this draft (merged from chunk translations). For non-chunked content, the main agent produces it directly.
 
-Translate the full content following `02-prompt.md`. Apply all **Translation principles** from SKILL.md Step 4, plus these step-specific guidelines:
-
-- Use the terminology decisions from Step 1 consistently
-- Match the identified tone and register
-- Follow the metaphor mapping from Step 1 for figurative language handling
-- Add translator's notes for comprehension challenges identified in Step 1
+Translate the full content following `02-prompt.md`. Apply all **Translation principles** from SKILL.md.
 
 ## Step 4: Critical Review
 
 The main agent critically reviews the draft against the source. Save review findings to `04-critique.md`. This step produces **diagnosis only** — no rewriting yet.
 
-### 4.1 Accuracy & Completeness
-- Compare each paragraph against the original, sentence by sentence
-- Verify all facts, numbers, dates, and proper nouns
-- Flag any content accidentally added, removed, or altered
-- Check that technical terms match glossary consistently throughout
-- Verify no paragraphs or sections were skipped
+### 4.1 Accuracy
 
-### 4.2 Europeanized Language Diagnosis (for CJK targets)
-- **Unnecessary connectives**: Overuse of 因此/然而/此外/另外 where context already implies the relationship
-- **Passive voice abuse**: Excessive 被/由/受到 where active voice is more natural
-- **Noun pile-up**: Long modifier chains that should be broken into shorter clauses
-- **Cleft sentences**: Unnatural "是...的" structures calqued from English "It is...that"
-- **Over-nominalization**: Abstract nouns where verbs or adjectives would be more natural (e.g., "进行了讨论" → "讨论了")
-- **Awkward pronouns**: Overuse of 他/她/它/我们/你 where they can be omitted
+- Compare each paragraph against the original
+- Verify facts, numbers, dates, proper nouns
+- Flag content accidentally added, removed, or altered
+- Check terminology consistency with glossary
 
-### 4.3 Figurative Language & Emotional Fidelity
-- Cross-check against the metaphor mapping in `01-analysis.md`: were all flagged metaphors/idioms handled per the recommended approach (interpret/substitute/retain)?
-- Flag any metaphors or figurative expressions that were translated literally and sound unnatural or lose the intended meaning in the target language
-- Check emotional connotations: do words that carry subjective feelings in the source (e.g., "alarming", "haunting", "striking") evoke the same response in the translation, or were they flattened into neutral/objective descriptions?
-- Flag implied meanings that were lost: sentences where the author's deeper intent was not conveyed because the translator stayed too close to the surface meaning
+### 4.2 Native Voice
 
-### 4.4 Strategy Execution
-- Were the translation strategies from `02-prompt.md` actually followed?
-- Did the translator apply the tone and register identified in analysis?
-- Were comprehension challenges from `01-analysis.md` addressed with appropriate notes?
-- Were glossary terms used consistently?
+- Flag sentences that read as "translated" rather than "written" — unnatural word order, calques, stiff phrasing
+- For CJK targets: check for unnecessary connectives (因此/然而/此外), passive voice abuse (被/由/受到), noun pile-ups, over-nominalization
+- Flag metaphors translated literally that sound unnatural in the target language
+- Check emotional connotations are preserved, not flattened
+- Note where sentence restructuring would improve readability
 
-### 4.5 Expression & Logic
-- Flag sentences that read like "translationese" — unnatural word order, calques, stiff phrasing
-- Check logical flow between sentences and paragraphs
-- Identify where sentence restructuring would improve readability
-- Note where the target language idiom was missed
+### 4.3 Notes & Adaptation
 
-### 4.6 Translator's Notes Quality
-- Are notes accurate, concise, and genuinely helpful?
-- Identify missed comprehension challenges that need notes
-- Flag over-annotations on terms obvious to the target audience
-- Check that cultural references are explained where needed
-
-### 4.7 Cultural Adaptation
-- Do metaphors and idioms work in the target language?
-- Are any references potentially confusing or offensive in the target culture?
-- Could any passage be misinterpreted due to cultural context differences?
+- Are translator's notes accurate, concise, and genuinely helpful?
+- Flag missed comprehension challenges that need notes, and over-annotations on obvious terms
+- Were translation strategies from `02-prompt.md` followed?
+- Do cultural references work in the target language?
 
 **Save `04-critique.md`** with:
 ```
-## Accuracy & Completeness
+## Accuracy
 - [issue]: [location] — [description]
-- ...
 
-## Europeanized Language Issues
-- [issue type]: [example from draft] → [suggested fix]
-- ...
+## Native Voice
+- [issue]: [example] → [suggested fix]
 
-## Figurative Language & Emotional Fidelity
-- [literal metaphor]: [original] → [draft rendering] → [suggested interpretation]
-- [flattened emotion]: [original word/phrase] → [draft rendering] → [how to restore emotional effect]
-- ...
-
-## Strategy Execution
-- [strategy]: [followed/missed] — [details]
-- ...
-
-## Expression & Logic
-- [location]: [problem] → [suggestion]
-- ...
-
-## Translator's Notes
-- [add/remove/revise]: [term] — [reason]
-- ...
-
-## Cultural Adaptation
-- [issue]: [description] — [suggestion]
-- ...
+## Notes & Adaptation
+- [add/remove/revise]: [term/passage] — [reason]
 
 ## Summary
-[Overall assessment: X critical issues, Y improvements, Z minor suggestions]
+[Overall assessment: X critical issues, Y improvements]
 ```
 
 ## Step 5: Revision
 
 Apply all findings from `04-critique.md` to produce a revised translation. Save to `05-revision.md`.
 
-The revision reads `03-draft.md` (the original draft) and `04-critique.md` (the review findings), and may also refer back to the source text and `01-analysis.md`:
-
-- Fix all accuracy issues identified in the critique
-- Rewrite Europeanized expressions into natural target-language patterns
-- Re-interpret literally translated metaphors and figurative expressions per the metaphor mapping; replace with natural target-language renderings that convey the intended meaning and emotional effect
-- Restore flattened emotional connotations: ensure words carrying subjective feelings evoke the same response as the source
-- Apply missed translation strategies
-- Restructure stiff or awkward sentences for fluency
-- Add, remove, or revise translator's notes per critique recommendations
-- Improve transitions between paragraphs
-- Adapt cultural references as suggested
+Read `03-draft.md` and `04-critique.md`, fix all accuracy issues, rewrite unnatural expressions, adjust notes, and improve flow.
 
 ## Step 6: Polish
 
@@ -244,21 +128,18 @@ Save final version to `translation.md`.
 Final pass on `05-revision.md` for publication quality:
 
 - Read the entire translation as a standalone piece — does it flow as native content?
-- Smooth any remaining rough transitions between paragraphs
-- Ensure the narrative voice is consistent throughout
-- Apply the selected translation style consistently: storytelling should flow like a narrative, formal should maintain neutral professionalism, humorous should land jokes naturally in the target language, etc.
-- Final scan for surviving literal metaphors or flattened emotions: any figurative expression that still reads as "translated" rather than "written" should be recast into natural target-language expression
-- Final consistency check on terminology across the full text
-- Verify formatting is preserved correctly (headings, bold, links, code blocks)
-- Remove any remaining traces of translationese
+- Smooth remaining rough transitions
+- Ensure consistent narrative voice and style throughout
+- Final terminology consistency check
+- Verify formatting is preserved correctly
 
 ## Subagent Responsibility
 
-Each subagent (one per chunk) is responsible **only** for producing the initial draft of its chunk (Step 3). The main agent assembles the shared prompt (Step 2), spawns all subagents in parallel, then takes over for critical review (Step 4), revision (Step 5), and polish (Step 6). The main agent may delegate revision or polish to subagents at its own discretion.
+Each subagent (one per chunk) is responsible **only** for producing the initial draft of its chunk (Step 3). The main agent assembles the shared prompt (Step 2), spawns all subagents in parallel, then takes over for critical review (Step 4), revision (Step 5), and polish (Step 6).
 
 ## Chunked Refined Translation
 
-When content exceeds the chunk threshold (see Defaults in SKILL.md) and uses refined mode:
+When content exceeds the chunk threshold and uses refined mode:
 
 1. Main agent runs analysis (Step 1) on the **entire** document first → `01-analysis.md`
 2. Main agent assembles translation prompt → `02-prompt.md`
@@ -267,7 +148,4 @@ When content exceeds the chunk threshold (see Defaults in SKILL.md) and uses ref
 5. Main agent critically reviews the merged draft → `04-critique.md`
 6. Main agent revises based on critique → `05-revision.md`
 7. Main agent polishes → `translation.md`
-7. Final cross-chunk consistency check:
-   - Check terminology consistency across chunk boundaries
-   - Verify narrative flow between chunks
-   - Fix any transition issues at chunk boundaries
+8. Final cross-chunk consistency check: terminology, narrative flow, transitions at chunk boundaries

skills/baoyu-translate/references/subagent-prompt-template.md

@@ -15,43 +15,39 @@ Replace `{placeholders}` with actual values. Omit sections marked "if analysis e
 ```markdown
 You are a professional translator. Your task is to translate markdown content from {source_lang} to {target_lang}.
 
-## Target Audience
+## Target Audience & Style
 
-{audience description}
+**Audience**: {audience description}
 
-## Translation Style
+**Target style**: {style description — e.g., "storytelling: engaging narrative flow, smooth transitions, vivid phrasing" or custom style from user}
 
-{style description — e.g., "storytelling: engaging narrative flow, smooth transitions, vivid phrasing" or custom style from user}
-
-Apply this style consistently: it determines the voice, tone, and sentence-level choices throughout the translation. Style is independent of audience — a technical audience can still get a storytelling-style translation, or a general audience can get a formal one.
+**Source voice** (from analysis, if exists): {Brief description of the original author's voice — formal/conversational, humor, register, sentence rhythm.}
 
 ## Content Background
 
-{Inlined from 01-analysis.md if analysis exists: quick summary, core argument, author background, writing context, tone assessment, figurative language & metaphor mapping.}
+{Inlined from 01-analysis.md if analysis exists: content summary, core argument, author background, context.}
 
 ## Glossary
 
-Apply these term translations consistently throughout. First occurrence of each term: include the original in parentheses after the translation.
+Apply these term translations consistently. First occurrence: include original in parentheses.
 
-{Merged glossary — combine built-in glossary + EXTEND.md glossary + terms extracted in analysis. One per line: English → Translation}
+{Merged glossary — one per line: English → Translation}
 
-## Comprehension Challenges
+## Translation Challenges
 
-The following terms or references may confuse target readers. Add translator's notes in parentheses where they appear: `译文（English original，通俗解释）`
+{Inlined from 01-analysis.md §1.4 if analysis exists. Comprehension gaps, figurative language, structural challenges with suggested approaches:}
 
-{Inlined from 01-analysis.md comprehension challenges section if analysis exists. Each entry: term → explanation to use as note.}
+- **{term/passage}**: {challenge type} → {suggested approach}
 
 ## Translation Principles
 
+Rewrite the content into natural, engaging {target_lang} — not merely translate it. Every sentence should read as if a skilled native writer composed it from scratch.
+
 - **Accuracy first**: Facts, data, and logic must match the original exactly
-- **Meaning over words**: Translate what the author means, not just what the words say. When a literal translation sounds unnatural or fails to convey the intended effect, restructure freely to express the same meaning in idiomatic {target_lang}
-- **Figurative language**: Interpret metaphors, idioms, and figurative expressions by their intended meaning. When a source-language image does not carry the same connotation in {target_lang}, replace it with a natural expression that conveys the same idea and emotional effect. Refer to the Figurative Language section in Content Background (if provided) for pre-analyzed metaphor mappings
-- **Emotional fidelity**: Preserve the emotional connotations of word choices, not just their dictionary meanings
-- **Natural flow**: Use idiomatic {target_lang} word order and sentence patterns; break or restructure sentences freely when the source structure doesn't work naturally
-- **Terminology**: Use glossary translations consistently; annotate with original term in parentheses on first occurrence
+- **Natural flow**: Use idiomatic {target_lang} word order. Break long source sentences into shorter, natural ones. Interpret metaphors and idioms by intended meaning, not word-for-word
+- **Terminology**: Use glossary translations consistently. Annotate with original in parentheses on first occurrence of specialized terms
 - **Preserve format**: Keep all markdown formatting (headings, bold, italic, images, links, code blocks)
-- **Respect original**: Maintain original meaning and intent; do not add, remove, or editorialize — but sentence structure and imagery may be adapted freely to serve the meaning
-- **Translator's notes**: For terms or cultural references listed in Comprehension Challenges above, add a concise explanatory note in parentheses. Only annotate where genuinely needed for the target audience.
+- **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
 ```
 
 ---
@@ -63,6 +59,9 @@ The following terms or references may confuse target readers. Add translator's n
 ```
 Read the translation instructions from: {output_dir}/02-prompt.md
 
+You are translating chunk {NN} of {total_chunks}.
+Context: {brief description of what this chunk covers and where it sits in the overall argument}
+
 Translate this chunk:
 1. Read `{output_dir}/chunks/chunk-{NN}.md`
 2. Translate following the instructions in 02-prompt.md