--- name: ggs-sales-report description: > Generate standardized Alibaba.com industry insight & business proposal presentations (.pptx) for the GGS sales team. Accepts a single JSON file, runs build-report.js to produce a branded PowerPoint with charts and tables. --- # Alibaba.com Industry Report Generator > **⛔ MANDATORY RULES — Read before doing ANYTHING else:** > > 1. **DO NOT write your own pptxgenjs/PPTX generation code.** A complete build script already exists. Your only job is to **assemble the JSON data file** and **run the build script**. > 2. **DO NOT copy example values from this document or module spec files into your output data.** All example values are placeholders. You must use real data from CRM tools or user-provided sources. > 3. **DO NOT skip reading the module spec files** in `modules/` — they define the exact JSON schema each module requires. --- ## Execution SOP (Step-by-Step) **You MUST follow these steps in order. Do NOT deviate.** ### Step 0 — Clarify Required Parameters (MANDATORY) 生成报告前，必须确认以下信息已知。**如果当前对话上下文中无法获取，必须先向用户提问，不得跳过或自行假设。** | 需用户提供 | 说明 | AI 处理方式 | |------|------|----------| | **客户所在国家** | 自然语言即可，如"越南"、"韩国" | 转换为 2 位国家码（`VN`、`KR` 等） | | **客户主营行业** | 自然语言描述，如"家具"、"服装" | 调用 `ggs-crm industry-mapping` 自动匹配 lv1/lv2 行业 ID | > ⚠️ **一级行业 ID 和二级行业 ID 由 AI 自动推断，无需用户提供。** 用户只需告知国家和行业描述即可。 **澄清话术（当国家或行业信息缺失时，向用户提问）：** > 为了生成报告，我需要确认以下信息： > 1. **客户所在国家**是哪个？（如：越南、韩国、印度） > 2. **客户主营行业**是什么？（如：家具、服装、电子产品） > > 我会根据行业描述自动匹配行业 ID，无需您手动提供 ID。 以上信息确认后，AI 执行： 1. 将国家名转为 2 位国家码 2. 运行 `ggs-crm industry-mapping`，根据行业描述匹配 `cate_lv1_id` 和 `cate_lv2_id` 3. 确认 lv1/lv2 ID 后，进入 Step 1 ### Step 1 — Fix Dependencies (run once per environment) If you see `ERR_DLOPEN_FAILED` or `sharp` errors, run: ```bash cd npm rebuild sharp # If still failing on macOS, try: npm remove sharp && npm install sharp ``` Verify with: `node -e "require('sharp'); console.log('sharp OK')"` ### Step 2 — Gather Data & Build JSON Create a single JSON file (e.g., `report-data.json`) following the schema below. Data sources: | Module | Data Source | |--------|-----------| | SWOT (swot_data) | LLM analysis of client company (web search + user info) | | Macro Trends (macro_data) | LLM web search for country export statistics | | Success Stories (success_stories) | **`ggs-crm crm-industry-success-story`** or user-provided | | Industry Outlook (industry_outlook_data) | **`ggs-crm crm-industry-traffic`** + **`crm-industry-buyer-country`** | | Trending Products (trending_products_data) | **`ggs-crm crm-industry-product`** | | Benchmarks (benchmark_data) | **`ggs-crm crm-industry-benchmark`** | | Action Plan (action_plan) | LLM-generated recommendations | **For Modules 03, 04, 05, 07: You MUST call the ggs-crm CLI tools to get real data.** See "CRM CLI Reference" section below. Do NOT fabricate these numbers. ### Step 3 — Run the Build Script ```bash node scripts/build-report.js --input report-data.json --output proposal.pptx ``` That's it. The script handles all HTML template rendering, chart generation, table layout, image processing, and PPTX assembly automatically. **Do NOT modify the script or write your own.** ### Step 4 — Deliver the .pptx File Return the generated `.pptx` file to the user. Do NOT return HTML, PDF, or markdown. --- ## Input JSON Schema A single JSON file drives the entire report. **Omit a key to skip that module.** Reference example: `scripts/sample-data.json` (read this file for the complete structure). ### Top-Level Fields | Field | Required | Description | |-------|----------|-------------| | `title` | Yes | Report title on cover (e.g., "Alibaba.com Business Proposal") | | `company_name` | Yes | Client company name | | `main_category` | Yes | Client's primary industry/category | | `target_country` | No | Client's country (e.g., "Vietnam") | | `target_industry` | No | Export industry for data lookups (defaults to `main_category`) | | `sales_name` | Yes | Sales representative name | | `sales_email` | Yes | Sales representative email | | `language` | No | Target language code (default: "en") | | `include_about` | No | Set `false` to skip About Alibaba.com module (default: true) | ### Module Data Keys | JSON Key | Module | Pages | |----------|--------|-------| | `swot_data` | SWOT Analysis | 2 | | `macro_data` | Macroeconomic Trends | 1 | | _(include_about)_ | About Alibaba.com | 2 (fixed content) | | `success_stories` | Seller Success Stories | 1–N | | `industry_outlook_data` | Local Industry Outlook | 2 | | `trending_products_data` | Trending Product Insights | 1–4 | | `benchmark_data` | Industry Leaders & Benchmarks | 1–5 | | `action_plan` | Your Action Plan | 1 | ### Sub-Schema Quick Reference **Read the module spec files for complete field definitions.** Each spec is in `modules/`: - `modules/module-swot.md` — SWOT schema - `modules/module-01-macro-trends.md` — macro_data schema - `modules/module-03-success-stories.md` — success_stories schema - `modules/module-04-industry-outlook.md` — industry_outlook_data schema (traffic_trends + buyer_analysis) - `modules/module-05-trending-products.md` — trending_products_data schema (4 sub-tables) - `modules/module-07-benchmarks.md` — benchmark_data schema (peers array) **`swot_data`**: `{ company_name, company_logo, industry, country, brief, strengths[], weaknesses[], opportunities[], threats[] }` **`macro_data`**: `{ country, industry, currency, unit, export_data[{year,value}], cagr, ai_summary, data_sources[] }` **`success_stories[]`**: `[{ image, category, description, company, country }]` **`industry_outlook_data`**: ```json { "country": "...", "industry": "...", "traffic_trends": { "data_range": "2025-04 ~ 2026-03", "months": ["2025-04", ...], "country_ipv": [8400, ...], "ggs_ipv": [720000, ...], "country_ipv_unit": "", "ggs_ipv_unit": "", "ai_summary": "..." }, "buyer_analysis": { "data_range": "...", "buyers": [{ "country": "...", "share": 28.5, "yoy": 18.7 }], "ai_summary": "..." } } ``` **`trending_products_data`**: Contains `hot_categories`, `hot_keywords`, `high_demand`, `high_growth`. Each has `items[]` with fields specific to its type (see module-05 spec), plus `data_source`, `data_note`, `ai_summary`. **`benchmark_data`**: `{ peers: [{ company_name, profile_similarity[], core_strategy[], storefront_highlights[], roi_takeaways, cooperation_plan:{...} }] }` **`action_plan`**: `{ subtitle, steps: [{ title, description }] }` --- ## CRM CLI Reference **These tools provide real Alibaba.com platform data. You MUST use them for Modules 04, 05, and 07.** ### ID Resolution Before calling any CRM tool, resolve the user's industry name to numeric IDs using: ```bash ggs-crm industry-mapping # 可配合 grep 过滤，例如查找食品饮料相关行业： ggs-crm industry-mapping | grep -E "Food|Beverage|食品|饮料" ``` Match the user's natural-language industry input against the returned `*_desc` columns to find the best-fit IDs (`cate_lv1_id`, `cate_lv2_id`, `cate_lv3_id`). ### Commands All commands share: `--industry-id, -i` (lv1), `--country, -c` (country code). Some commands also accept `--industry-lv2-id` (lv2). ```bash # Module 03 — Success stories ggs-crm crm-industry-success-story -t -i -c # Module 04 Page 1 — Traffic trends ggs-crm crm-industry-traffic -i --industry-lv2-id -c # Module 04 Page 2 — Buyer country analysis ggs-crm crm-industry-buyer-country -i --industry-lv2-id -c # Module 05 — Trending products ggs-crm crm-industry-product -i --industry-lv2-id -c # Module 07 — Industry benchmarks ggs-crm crm-industry-benchmark -i --industry-lv2-id -c ``` > ⚠️ **注意：** `crm-industry-success-story` 需要 `-t ` 参数（需先完成认证），且只需一级行业 ID（`-i`），不需要二级行业 ID。 ### IPV Unit Handling CRM traffic data may return unit indicators in the response: - `country_ipv_unit: ''` → absolute numbers - `ggs_ipv_unit: 'K'` → values are in thousands (build script auto-multiplies ×1000) - `ggs_ipv_unit: 'M'` → values are in millions (build script auto-multiplies ×1000000) **Pass these unit fields through to the JSON as-is.** The build script handles normalization. --- ## Data Processing Rules ### Currency Conversion Upstream monetary values are in **RMB**. Report displays **USD**. Fixed rate: 1 USD = 7.2 RMB. If `benchmark_data.peers[].cooperation_plan_currency` is `"RMB"`, build script auto-converts. Otherwise values are assumed USD. ### Data Desensitization (Module 07) Company names are auto-masked. USD amounts are auto-approximated (e.g., 38889 → "~39K"). The build script handles this — just pass raw values. --- ## Presentation Structure ``` Slide | Data Key -------------------------------------|--------- Cover Page | (top-level fields) SWOT Analysis (2 pages) | swot_data Table of Contents | (auto-generated) 01 Macroeconomic Trends | macro_data 02 About Alibaba.com (2 pages) | include_about 03 Seller Success Stories (1–N) | success_stories 04 Local Industry Outlook (2 pages) | industry_outlook_data 05 Trending Product Insights (1–4) | trending_products_data 06 Market Performance (placeholder) | — 07 Industry Leaders (1–5 pages) | benchmark_data 08 Your Action Plan | action_plan End Page (Thank You) | (top-level fields) ``` SWOT appears before TOC, is NOT numbered in TOC. All other modules get sequential numbering. If a module's sub-data is missing, a 404 error page is generated automatically. --- ## Multi-Language Support The build script (`build-report.js`) has a **built-in translation system** that handles all UI text automatically. You only need to set the `language` field in the top-level JSON. ### How It Works 1. **Language detection**: The script reads `data.language` (e.g., `"zh"`, `"es"`, `"fr"`) and selects the matching translation column from an internal `UI_TEXTS` dictionary. 2. **Automatic UI translation**: All framework text — module titles, section dividers, TOC heading, SWOT labels, cover labels, table headers, About Alibaba.com content, timeline milestones, Thank You page, error pages — is automatically translated. **You do NOT need to translate these yourself.** 3. **Fallback**: If a language is not in the dictionary, the script falls back to English (`"en"`). ### What You MUST Translate in the JSON Even though the build script translates UI framework text, the **data content** in the JSON must also be in the target language. When `language` is not `"en"`, translate ALL of the following before writing to the JSON: - `title` (report title) - `swot_data.brief`, `swot_data.strengths[]`, `swot_data.weaknesses[]`, `swot_data.opportunities[]`, `swot_data.threats[]` - `macro_data.ai_summary` - `success_stories[].description`, `success_stories[].category` - `industry_outlook_data.traffic_trends.ai_summary`, `industry_outlook_data.buyer_analysis.ai_summary` - `trending_products_data.*.ai_summary` - `benchmark_data.peers[].profile_similarity[]`, `core_strategy[]`, `storefront_highlights[]`, `roi_takeaways` - `action_plan.subtitle`, `action_plan.steps[].title`, `action_plan.steps[].description` ### Language Consistency Rule **ALL content must be in a single, consistent language.** No mixing. - If the user specifies a language (e.g., "用中文生成"), set `"language": "zh"` and write all JSON data content in Chinese. - If the user doesn't specify a language, detect the user's conversation language and use that for both `language` field and all data content. - The build script will handle UI framework text in the matching language automatically. ### Supported Language Codes | Code | Language | |------|----------| | `en` | English (default) | | `zh` | 中文 (Chinese Simplified) | | `es` | Español (Spanish) | | `fr` | Français (French) | | `pt` | Português (Portuguese) | | `ja` | 日本語 (Japanese) | | `ko` | 한국어 (Korean) | | `ar` | العربية (Arabic) | | `ru` | Русский (Russian) | | `vi` | Tiếng Việt (Vietnamese) | | `th` | ไทย (Thai) | | `id` | Bahasa Indonesia | | `tr` | Türkçe (Turkish) | ### Brand Names — Do NOT Translate These brand names must remain in English regardless of target language: **Alibaba.com**, **Trade Assurance**, **Smart Assistant**, **Alibaba Guaranteed** --- ## Build Architecture ``` alibaba-report/ ├── SKILL.md # This file ├── assets/ # Logos, backgrounds ├── slide-templates/ # HTML layout templates ├── modules/ # Per-module schema specs (READ THESE) ├── references/ # ggs_industry_mapping.csv ├── scripts/ │ ├── build-report.js # ★ Production build (JSON → PPTX) │ ├── test-build.js # Dev build with mock data │ └── sample-data.json # Complete example JSON └── node_modules/ # pptxgenjs, sharp, playwright ``` The build pipeline: JSON → HTML templates (variable substitution) → Playwright screenshot → PptxGenJS (charts/tables overlay) → .pptx output. **Everything is automated in `build-report.js`. You only need to provide the JSON.**