@renoise/video-maker 0.1.3 → 0.2.0

package/skills/gemini-gen/scripts/gemini.mjs

@@ -0,0 +1,220 @@
+#!/usr/bin/env node
+
+/**
+ * Gemini API client via Renoise gateway.
+ * Zero npm dependencies — uses native fetch.
+ *
+ * Usage:
+ *   # Text only
+ *   node gemini.mjs "Explain quantum computing"
+ *
+ *   # With image(s)
+ *   node gemini.mjs --file photo.jpg "Describe this product"
+ *   node gemini.mjs --file a.jpg --file b.jpg "Compare these two"
+ *
+ *   # With video
+ *   node gemini.mjs --file clip.mp4 --resolution low "Summarize this clip"
+ *
+ *   # With uploaded file URI (from file upload skill)
+ *   node gemini.mjs --file-uri "https://...fileUri" --file-mime video/mp4 "Analyze this video"
+ *
+ *   # JSON output mode
+ *   node gemini.mjs --json "Return a JSON object with name and age"
+ *
+ * Options:
+ *   --file <path>         Attach a local file (image/video). Repeatable.
+ *   --file-uri <uri>      Attach an uploaded file by URI. Requires --file-mime.
+ *   --file-mime <mime>     MIME type for --file-uri.
+ *   --resolution <level>  Media resolution: low|medium|high|ultra_high (default: medium)
+ *   --model <name>        Model name (default: gemini-3.1-pro)
+ *   --temperature <n>     Temperature (default: 1.0)
+ *   --max-tokens <n>      Max output tokens (default: 8192)
+ *   --json                Request JSON response format
+ *
+ * Environment:
+ *   RENOISE_API_KEY       Required. Get one at https://www.renoise.ai
+ */
+
+import fs from "fs/promises";
+import path from "path";
+
+// --- Auth ---
+const RENOISE_API_KEY = process.env.RENOISE_API_KEY;
+if (!RENOISE_API_KEY) {
+  console.error(
+    "RENOISE_API_KEY not set. Get one at: https://www.renoise.ai"
+  );
+  process.exit(1);
+}
+
+// --- MIME detection ---
+const MIME_MAP = {
+  ".jpg": "image/jpeg",
+  ".jpeg": "image/jpeg",
+  ".png": "image/png",
+  ".webp": "image/webp",
+  ".gif": "image/gif",
+  ".mp4": "video/mp4",
+  ".mov": "video/quicktime",
+  ".webm": "video/webm",
+};
+
+function getMimeType(filePath) {
+  return (
+    MIME_MAP[path.extname(filePath).toLowerCase()] ?? "application/octet-stream"
+  );
+}
+
+// --- Resolution mapping ---
+const RESOLUTION_MAP = {
+  low: "media_resolution_low",
+  medium: "media_resolution_medium",
+  high: "media_resolution_high",
+  ultra_high: "media_resolution_ultra_high",
+};
+
+// --- Parse args ---
+function parseArgs(argv) {
+  const files = [];
+  let fileUri = null;
+  let fileMime = null;
+  let resolution = "medium";
+  let model = "gemini-3.1-pro";
+  let temperature = 1.0;
+  let maxTokens = 8192;
+  let json = false;
+  const textParts = [];
+
+  for (let i = 0; i < argv.length; i++) {
+    switch (argv[i]) {
+      case "--file":
+        files.push(argv[++i]);
+        break;
+      case "--file-uri":
+        fileUri = argv[++i];
+        break;
+      case "--file-mime":
+        fileMime = argv[++i];
+        break;
+      case "--resolution":
+        resolution = argv[++i];
+        break;
+      case "--model":
+        model = argv[++i];
+        break;
+      case "--temperature":
+        temperature = parseFloat(argv[++i]);
+        break;
+      case "--max-tokens":
+        maxTokens = parseInt(argv[++i], 10);
+        break;
+      case "--json":
+        json = true;
+        break;
+      default:
+        textParts.push(argv[i]);
+    }
+  }
+
+  return {
+    files,
+    fileUri,
+    fileMime,
+    resolution,
+    model,
+    temperature,
+    maxTokens,
+    json,
+    prompt: textParts.join(" "),
+  };
+}
+
+// --- Build parts ---
+async function buildParts(opts) {
+  const parts = [];
+  const resLevel = RESOLUTION_MAP[opts.resolution] ?? RESOLUTION_MAP.medium;
+
+  // Local files → inline base64
+  for (const filePath of opts.files) {
+    const data = await fs.readFile(filePath);
+    parts.push({
+      inlineData: {
+        mimeType: getMimeType(filePath),
+        data: data.toString("base64"),
+      },
+      mediaResolution: { level: resLevel },
+    });
+  }
+
+  // Uploaded file URI
+  if (opts.fileUri) {
+    parts.push({
+      fileData: {
+        mimeType: opts.fileMime ?? "application/octet-stream",
+        fileUri: opts.fileUri,
+      },
+    });
+  }
+
+  // Text prompt (always last)
+  if (opts.prompt) {
+    parts.push({ text: opts.prompt });
+  }
+
+  return parts;
+}
+
+// --- Main ---
+async function main() {
+  const opts = parseArgs(process.argv.slice(2));
+
+  if (!opts.prompt && opts.files.length === 0 && !opts.fileUri) {
+    console.error(
+      "Usage: node gemini.mjs [--file <path>] [--resolution low|medium|high|ultra_high] <prompt>"
+    );
+    process.exit(1);
+  }
+
+  const endpoint = `https://staging--ujgsvru36x4korjj10nq.edgespark.app/api/public/llm/proxy/v1beta/models/${opts.model}:generateContent?key=${RENOISE_API_KEY}`;
+
+  const parts = await buildParts(opts);
+
+  const body = {
+    contents: [{ role: "user", parts }],
+    generationConfig: {
+      temperature: opts.temperature,
+      maxOutputTokens: opts.maxTokens,
+    },
+  };
+
+  if (opts.json) {
+    body.generationConfig.responseMimeType = "application/json";
+  }
+
+  const res = await fetch(endpoint, {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify(body),
+  });
+
+  if (!res.ok) {
+    const errText = await res.text();
+    console.error(`Gemini API error ${res.status}: ${errText}`);
+    process.exit(1);
+  }
+
+  const data = await res.json();
+  const text = data.candidates?.[0]?.content?.parts?.[0]?.text;
+
+  if (!text) {
+    console.error("No text in response:", JSON.stringify(data, null, 2));
+    process.exit(1);
+  }
+
+  console.log(text);
+}
+
+main().catch((err) => {
+  console.error("ERROR:", err.message);
+  process.exit(1);
+});

package/skills/renoise-gen/SKILL.md

@@ -13,6 +13,8 @@ metadata:
 
 Generate AI videos and images through the Renoise platform.
 
+> **IMPORTANT**: The Renoise website is **https://www.renoise.ai** — NOT renoise.com. Always use `renoise.ai` when referencing the platform URL.
+
 ## Supported Models
 
 | Model | Type | Description |
@@ -96,7 +98,7 @@ See `${CLAUDE_SKILL_DIR}/references/video-capabilities.md` for details.
 
 CLI path: `${CLAUDE_SKILL_DIR}/renoise-cli.mjs` (Node.js 18+)
 
-API Key and base URL are configured via environment variables (`RENOISE_API_KEY`, `RENOISE_BASE_URL`).
+API Key and base URL are configured via environment variables (`RENOISE_API_KEY`, `RENOISE_BASE_URL`). Get your API key at https://www.renoise.ai (NOT renoise.com).
 
 ## CLI Commands
 

package/skills/short-film-editor/SKILL.md

@@ -1,15 +1,17 @@
 ---
 name: short-film-editor
 description: >
-  短片剪辑师：音乐先行，按节拍卡点拆分故事为多段（5-15s不等），
-  生成HTML分镜预览供确认，维护角色/画风一致性，批量生成并输出组装指南。
-  触发词：short film, multi-clip, 短片, 1分钟视频, story video, 多段视频
+  Short film editor: music-first workflow, splits story into beat-synced segments (5-15s each),
+  generates HTML storyboard preview for confirmation, maintains character/style consistency,
+  batch generates and outputs assembly guide.
+  Use when user says "short film", "multi-clip", "story video", "multi-segment video",
+  "1-minute video".
 allowed-tools: Bash, Read
 metadata:
   author: renoise
   version: 0.1.0
   category: video-production
-  tags: [short-film, multi-clip, narrative, story, 短片]
+  tags: [short-film, multi-clip, narrative, story]
 ---
 
 # Short Film Editor
@@ -86,7 +88,7 @@ You are a short film editor specializing in multi-clip AI video production. You
 
 **If user has or wants music**, follow Steps 1-3 below.
 
-**If user skips music** (e.g. "先不用音乐" / "no music for now"), skip to **Step 4 — Manual Rhythm**:
+**If user skips music** (e.g. "no music for now"), skip to **Step 4 — Manual Rhythm**:
 - Define segments based on narrative pacing, not equal splits.
 - Vary durations: establishing shots 7-10s, action bursts 5-6s, aftermath/resolution 5-7s.
 - Aim for 4-7 segments depending on total duration.
@@ -149,7 +151,7 @@ Output JSON:
 | Climax / collision | 6-8s | Intense, dense |
 | Aftermath / resolution | 5-7s | Slow, lingering |
 
-**In-clip cutting (镜内切换)**:
+**In-clip cutting**:
 Real films average 2-4s per camera angle (action films: 1-2s). Seedance minimum is 5s per clip, so use **time-annotated camera changes within each clip** to simulate fast cutting:
 
 ```
@@ -247,13 +249,14 @@ Store each prompt in the shot's `prompt` field in `project.json`.
 
 ### Step 2 — Generate Reference Images
 
-Three image sources (ask user preference, or default to Gemini):
+Three image sources (ask user preference, or default to Renoise):
 
-**Option A — Gemini (default, automatic, free)**:
+**Option A — Renoise (default)**:
+Use `renoise-gen` with `nano-banana-2` model to generate a reference image. Prompt should describe the shot scene + character appearance + key action + lighting (NO camera movement). Save the result to `${PROJECT_DIR}/storyboard/${shot_id}.png`.
 ```bash
-npx tsx ${CLAUDE_PLUGIN_ROOT}/skills/scene-generate/scripts/generate-scene.ts \
-  "<shot scene + character appearance + key action + lighting — NO camera movement>" \
-  "${PROJECT_DIR}/storyboard/${shot_id}.png"
+node ${CLAUDE_PLUGIN_ROOT}/skills/renoise-gen/renoise-cli.mjs task generate \
+  --model nano-banana-2 --resolution 2k --ratio 16:9 \
+  --prompt "<shot scene + character appearance + key action + lighting>"
 ```
 
 **Option B — Midjourney (higher quality, recommended for stylized projects)**:
@@ -271,12 +274,14 @@ Submit all shots in parallel via `/v1/tob/diffusion`, poll for completion, downl
 **Option C — User-provided**:
 User manually places reference images in `${PROJECT_DIR}/storyboard/S1.png`, `S2.png`, etc.
 
-**Option D — Gemini Grid Storyboard (recommended for best consistency)**:
-Generate ALL shots in a single grid image so characters and style are naturally consistent across panels, then split into individual reference images.
+**Option D — Renoise Grid Storyboard (recommended for best consistency)**:
+Generate ALL shots in a single grid image via `renoise-gen` `nano-banana-2` so characters and style are naturally consistent across panels, then split into individual reference images.
 
-1. Generate a single N-panel grid with Gemini:
-   ```
-   Prompt: "Generate a single N-panel [manga/cinematic] storyboard grid image.
+1. Generate a single N-panel grid:
+   ```bash
+   node ${CLAUDE_PLUGIN_ROOT}/skills/renoise-gen/renoise-cli.mjs task generate \
+     --model nano-banana-2 --resolution 2k --ratio 16:9 \
+     --prompt "Generate a single N-panel [manga/cinematic] storyboard grid image.
    Layout: 2 rows x 4 columns grid with thin white borders.
    The SAME two characters must appear consistently across all panels:
    Character A: [verbatim from Character Bible]
@@ -312,18 +317,12 @@ Reference image prompts should include:
 
 ### Step 3 — Generate HTML Storyboard Preview
 
-```bash
-npx tsx ${CLAUDE_SKILL_DIR}/scripts/generate-storyboard-html.ts \
-  --project-file "${PROJECT_DIR}/project.json" \
-  --output "${PROJECT_DIR}/storyboard.html"
-```
-
-This generates a single self-contained HTML file with:
+Generate a single self-contained HTML file from `${PROJECT_DIR}/project.json` and save it to `${PROJECT_DIR}/storyboard.html`. The HTML should include:
 - **Header**: Project title, total duration, clip count, BPM, character summary, style summary
 - **Music timeline**: Visual bar showing sections and cut points
 - **Shot cards**: One card per shot with reference image, scene/action, dialogue/beats, continuity, and collapsible Seedance prompt
 - **Reference images**: Base64-embedded inline (single-file, shareable). Use `--skip-images` to skip Gemini generation and read existing images from `storyboard/` directory (useful when using Midjourney or user-provided images).
-- **UI language**: Matches user's language (default: Chinese). White theme, bold confident design.
+- **UI language**: English. White theme, bold confident design.
 - **Responsive**: Viewable on phone
 
 Open for preview:

package/skills/short-film-editor/references/continuity-guide.md

@@ -129,7 +129,7 @@ AI-generated clips will achieve ~80% visual consistency when following these tec
 - Subtle speed adjustments for timing
 - Audio continuity (shared BGM) creates perceived visual continuity
 
-## Grid Storyboard Method (宫格分镜法)
+## Grid Storyboard Method
 
 ### Why One Image > Many Images
 
@@ -147,7 +147,7 @@ When generating reference images for each shot independently (even with the same
 ### Workflow
 
 1. Write a single prompt describing all panels with verbatim character descriptions
-2. Generate one grid image via Gemini (`gemini-3-pro-image-preview`)
+2. Generate one grid image via `renoise-gen` (`nano-banana-2`)
 3. Split into individual panels: `bash split-grid.sh grid.png storyboard/ 2 4`
 4. Upload each panel as material for Image-to-Video generation
 5. Each Seedance clip now has a visual anchor from the same source

package/skills/tiktok-content-maker/SKILL.md

@@ -3,8 +3,8 @@ name: tiktok-content-maker
 description: >
   TikTok e-commerce short video script generator. Analyzes product photos,
   generates 15s video scripts with video prompts and English dialogue.
-  Use when user says "TikTok product video", "ecommerce video", "电商视频",
-  "带货视频", "商品视频", "拍商品". Do NOT use for non-ecommerce videos or
+  Use when user says "TikTok product video", "ecommerce video",
+  "product video", "sales video", "shoot product". Do NOT use for non-ecommerce videos or
   general creative direction (use director instead).
 allowed-tools: Bash, Read
 metadata:
@@ -14,130 +14,127 @@ metadata:
   tags: [product, ecommerce, tiktok]
 ---
 
-# Content Maker — 电商短视频脚本 + 视频生成
+# Content Maker — E-commerce Short Video Script + Generation
 
 ## Overview
 
-电商短视频全流程工具：用户提供商品图（+ 可选模特图）→ 分析商品信息 → 生成 15 秒 TikTok 脚本（视频 prompt，含英文台词嵌入）→ 提交视频生成任务。
+End-to-end e-commerce short video tool: user provides product images (+ optional model images) → analyze product info → generate 15-second TikTok script (video prompt with embedded English dialogue) → submit video generation task.
 
 ## Workflow
 
-### Phase 1: 素材收集 & 商品分析
+### Phase 1: Material Collection & Product Analysis
 
-1. **收集素材路径**：向用户索要图片
-   - `商品图路径`（必需）：产品主图。**最佳：干净白底纯产品图，无文字/标注/装饰**。有营销文字覆盖的图会干扰模型。
-   - `模特图路径`（可选，仅供分析参考）：展示穿搭/使用效果的图。**注意：模特图仅用于理解产品使用方式，不上传到 Renoise**（隐私检测会拦截含真人面孔的图片）。
+1. **Collect material paths**: Ask user for images
+   - `Product image path` (required): Product hero image. **Best: clean white-background product photo with no text/labels/decorations**. Images with marketing text overlays will interfere with the model.
+   - `Model image path` (optional, for analysis reference only): Shows how the product is worn/used. **Note: Model images are only used to understand product usage — they are NOT uploaded to Renoise** (privacy detection will block images containing realistic human faces).
 
-2. **分析商品信息**：
-   - 如果有 Gemini API 可用，调用 Gemini 分析：
-     ```bash
-     cd ${CLAUDE_PLUGIN_ROOT} && npm install --silent && npx tsx ${CLAUDE_SKILL_DIR}/scripts/analyze-images.ts "<商品图路径>" "<模特图路径>"
-     ```
-   - 也可以直接通过 Read 工具查看图片，人工分析商品信息
-   - 需要提取：商品类型、颜色、材质、卖点、品牌调性、适用场景
-   - **（关键）从使用场景图中理解产品的正确使用方式**：
-     - 用户的姿势是什么？（站/坐/躺/走）
-     - 产品放在身体哪个位置？（手持/地面/桌面/身体下方）
-     - 产品与身体的交互方式？（用手按压 vs 用体重压 vs 穿戴 vs 涂抹）
-     - 使用场景在哪？（健身房/办公室/家里/户外）
-   - 如果用户提供了商品链接，用 WebFetch 抓取产品详情页补充理解
+2. **Analyze product info**:
+   - Use the `gemini-gen` skill to analyze product images — send the image(s) with a prompt requesting product analysis (type, color, material, selling points, brand tone, scene suggestions)
+   - Alternatively, view images directly via the Read tool and analyze manually
+   - Extract: product type, color, material, selling points, brand tone, applicable scenarios
+   - **(Critical) Understand correct product usage from lifestyle images**:
+     - What is the user's posture? (standing/sitting/lying/walking)
+     - Where is the product positioned on the body? (handheld/floor/table/under body)
+     - How does the product interact with the body? (hand pressure vs body weight vs wearing vs applying)
+     - Where is the usage scenario? (gym/office/home/outdoors)
+   - If the user provides a product link, use WebFetch to scrape product detail page for additional context
 
-3. **展示分析结果**，让用户确认或补充信息。分析结果中必须包含一条明确的「**使用方式描述**」，例如：
-   > 使用方式：将花生球放在地面/瑜伽垫上，用户躺在球上方，通过自身体重施压按摩脊柱两侧肌肉。花生形凹槽避开脊柱，两侧球体作用于竖脊肌。
+3. **Present analysis results** for user to confirm or supplement. Results must include a clear "**Usage description**", e.g.:
+   > Usage: Place the peanut ball on the floor/yoga mat, user lies on top of the ball, using body weight to massage the muscles along both sides of the spine. The peanut-shaped groove avoids the spine while the two ball ends work the erector spinae muscles.
 
-### Phase 2: 15 秒脚本 + Prompt 生成
+### Phase 2: 15-Second Script + Prompt Generation
 
-基于分析结果 + 参考指南，生成完整的 15 秒视频脚本。
+Based on analysis results + reference guide, generate a complete 15-second video script.
 
-**必须参考以下指南**（先 Read 再生成）：
-- `${CLAUDE_SKILL_DIR}/references/ecom-prompt-guide.md` — 电商视频 prompt 指南
+**Must reference the following guide** (Read before generating):
+- `${CLAUDE_SKILL_DIR}/references/ecom-prompt-guide.md` — E-commerce video prompt guide
 
-**Prompt 结构（3 个必需组成部分）：**
+**Prompt structure (3 required components):**
 
-#### Part A: 产品锚定（Prompt 开头，一句话）
+#### Part A: Product Anchoring (first line of prompt)
 
-产品外观靠参考图传达，prompt 里只需**一句话**说明产品是什么 + 用途：
+Product appearance is conveyed by the reference image. The prompt only needs **one sentence** stating what the product is + its use case:
 
 ```
 The product is a [brand] [product type] for [primary use case], shown in the reference image.
 The product must match the reference image exactly in every frame. Do not invent any packaging, box, or container unless the reference image shows one.
 ```
 
-**关键**：不要在 prompt 里重复描述颜色、材质、形状、logo — 这些信息已在参考图里。把 prompt 空间留给 hook 和画面叙事。
+**Key**: Do not repeat color, material, shape, or logo descriptions in the prompt — that information is already in the reference image. Save prompt space for the hook and visual narrative.
 
-#### Part B: 台词嵌入（贯穿全段）
+#### Part B: Dialogue Embedding (throughout)
 
-台词必须是英文，以强制口型同步格式嵌入叙事段落中：
+Dialogue must be in English, embedded in the narrative using forced lip-sync format:
 ```
 Spoken dialogue (say EXACTLY, word-for-word): "..."
 Mouth clearly visible when speaking, lip-sync aligned.
 ```
 
-**台词风格要求**：
-- **闺蜜聊天感**：像在跟朋友推荐，不像在念广告词
-- **高信息密度**：每句话都带具体信息（数字、对比、使用场景），没有废话
-- **不硬推销**：结尾不用 "link below" / "点击链接" 这种生硬 CTA，用自然的个人推荐收尾（如 "Best money I have spent this year"、"Trust me just start"）
+**Dialogue style requirements**:
+- **Best-friend casual tone**: Like recommending to a friend, not reading ad copy
+- **High information density**: Every sentence includes specific details (numbers, comparisons, usage scenarios) — no filler
+- **No hard sell**: Don't end with "link below" or generic CTAs. Use natural personal recommendations (e.g., "Best money I have spent this year", "Trust me just start")
 
-**台词节奏**（4 句，对应 4 个时间段）：
+**Dialogue pacing** (4 lines, matching 4 time segments):
 ```
-[0-3s]   Hook — 一句话喊停用户（痛点/悬念/结果前置）
-[3-8s]   卖点 — 具体参数 + 使用体验
-[8-12s]  场景 — 在哪用 + 便携性/多功能
-[12-15s] 收尾 — 个人真实推荐感，不硬推销
+[0-3s]   Hook — One sentence to stop the scroll (pain point / suspense / result-first)
+[3-8s]   Selling point — Specific specs + personal experience
+[8-12s]  Scene — Where to use + portability / versatility
+[12-15s] Close — Genuine personal recommendation, no hard sell
 ```
 
-#### Part C: 画面叙事（一段连续叙事）
+#### Part C: Visual Narrative (one continuous narrative)
 
-**视频结构（一个连续 15 秒视频）：**
+**Video structure (one continuous 15-second video):**
 ```
-[0-3s]   HOOK — 高冲击力开场。必须：快速运镜（whip pan / snap dolly in）+ 动态动作 + 立即开口说台词。绝对不能慢热。
-[3-8s]   SHOWCASE — 产品展示 + 模特互动。运镜变化展示材质细节。
-[8-12s]  SCENE — 生活场景使用。拉远到中景/全景。
-[12-15s] CLOSE — 模特面对镜头 + 产品在画面中 + 自然收尾。frame holds steady。
+[0-3s]   HOOK — High-impact opening. Must: fast camera movement (whip pan / snap dolly in) + dynamic action + start speaking immediately. Never start slow.
+[3-8s]   SHOWCASE — Product display + model interaction. Camera transitions to reveal material details.
+[8-12s]  SCENE — Real-life usage scenario. Pull back to medium/wide shot.
+[12-15s] CLOSE — Model faces camera + product in frame + natural ending. Frame holds steady.
 ```
 
-**输出 3 项内容：**
+**Output 3 items:**
 
-#### 1. Video Prompt（英文，含台词）
-导演口述式段落（6-10 句，每句只做一件事），包含：
-- 产品锚定（一句话，Part A）在最开头
-- 台词以 `Spoken dialogue (say EXACTLY, word-for-word):` 格式嵌入（Part B）
-- 每句台词后跟 `Mouth clearly visible when speaking, lip-sync aligned.`
-- Ad-6D Protocol 元素穿插
-- 模特外观一致性描述（性别、发型、肤色、体型、服装）
-- 运镜变化至少 3 次
-- 光线/氛围描述
+#### 1. Video Prompt (English, with dialogue)
+Director-dictation style paragraph (6-10 sentences, one thing per sentence), containing:
+- Product anchoring (one sentence, Part A) at the very beginning
+- Dialogue embedded with `Spoken dialogue (say EXACTLY, word-for-word):` format (Part B)
+- `Mouth clearly visible when speaking, lip-sync aligned.` after each dialogue line
+- Ad-6D Protocol elements interspersed
+- Model appearance consistency description (gender, hair, skin tone, body type, outfit)
+- At least 3 camera movement changes
+- Lighting/atmosphere description
 
-#### 2. 台词脚本（英文，标注时间段）
-单独列出 4 句台词及对应时间段，方便审阅。
+#### 2. Dialogue Script (English, with timestamps)
+List the 4 dialogue lines separately with their time segments for easy review.
 
-#### 3. BGM/音效建议
-- 推荐适合产品调性的音乐风格
-- 关键节点的音效提示
+#### 3. BGM / Sound Design Suggestions
+- Recommend music style matching the product tone
+- Key moment sound effect cues
 
-**参考示例**：Read `${CLAUDE_SKILL_DIR}/examples/dress-demo.md` 了解最新标准输出格式。
+**Reference example**: Read `${CLAUDE_SKILL_DIR}/examples/dress-demo.md` for the latest standard output format.
 
-### Phase 3: 用户确认
+### Phase 3: User Confirmation
 
-展示完整脚本后，询问用户：
-- 是否调整台词
-- 是否更换场景
-- 是否修改 prompt 细节
-- 确认后进入提交
+After presenting the full script, ask the user:
+- Whether to adjust dialogue
+- Whether to change the scene
+- Whether to modify prompt details
+- Proceed to submission after confirmation
 
-### Phase 4: 上传素材 + 提交视频生成任务
+### Phase 4: Upload Materials + Submit Video Generation Task
 
-用户确认脚本后，上传商品图并提交视频生成任务。
+After user confirms the script, upload the product image and submit the video generation task.
 
-**重要规则**：
-- 只上传商品图，**不上传模特/真人图**（隐私检测会拦截含真人面孔的图片，报错 `InputImageSensitiveContentDetected.PrivacyInformation`）
-- 模特外观完全靠 prompt 文字描述控制
-- 商品图最好用干净白底纯产品图，避免有营销文字覆盖的图
-- 批量生成时：商品图只需上传一次，复用 material ID 提交多个不同场景的任务
+**Important rules**:
+- Only upload product images — **never upload model/real person photos** (privacy detection will block images containing realistic human faces, error: `InputImageSensitiveContentDetected.PrivacyInformation`)
+- Model appearance is controlled entirely by prompt text description
+- Product images should ideally be clean white-background product photos, avoid images with marketing text overlays
+- For batch generation: upload the product image once, reuse the material ID to submit multiple tasks with different scenes
 
 ## Important Notes
 
-- 图片支持 jpg/jpeg/png/webp 格式
-- 视频 prompt 必须全英文
-- 台词必须英文，嵌入 prompt（`Spoken dialogue (say EXACTLY, word-for-word): "..."`）
-- **不输出单独的字幕文案** — 台词已在 prompt 中，不需要额外字幕层
+- Images support jpg/jpeg/png/webp formats
+- Video prompts must be entirely in English
+- Dialogue must be in English, embedded in the prompt (`Spoken dialogue (say EXACTLY, word-for-word): "..."`)
+- **Do not output separate subtitle text** — dialogue is already in the prompt, no additional subtitle layer needed