ricocc/rico-skills/rico-resource-article

Resource Recommendation Article Generator

This skill helps you create high-quality resource recommendation articles through a multi-turn interactive workflow, ensuring accurate content, complete structure, and professional style.

Core Features

• Intelligent Step-by-Step Generation: Collect candidates → User confirmation → Generate titles → Generate article
• Domain Adaptation: Auto-detect topic domain (Design/Development/Marketing/Product/Content) and apply professional terminology and writing style
• Broad Domain Support: Technical tools, design resources, development frameworks, online services, physical products, etc.
• Multiple Generation Modes: Smart step-by-step, one-shot, outline-based generation
• Professional Format Standards: Complete Front Matter and standard article structure
• Data Accuracy: Only use verifiable data (GitHub Stars, component counts, official metrics)

Triggers

This skill supports two trigger modes:

Mode 1: Explicit Command (rico article prefix)

• rico article [topic] / rico 文章 [主题]
• rico article 8 design tools / rico 文章 8个设计工具

Mode 2: Natural Language (explicit article intent required)

• "Generate a recommendation article about shadcn/ui"
• "写一篇工具推荐文章"
• "Create a resource list article"

Important: This skill ONLY triggers when user explicitly wants to generate an article. Do NOT trigger for simple questions like "recommend 8 tools" or "有哪些好用的工具" - answer directly.

Examples

Trigger examples:

• rico article 8 React hooks libraries
• rico 文章 10个设计工具
• Generate a recommendation article about shadcn/ui
• 写一篇 Figma 插件推荐文章

Do NOT trigger examples (just asking or conversing):

• "推荐 8 个组件库" - Answer directly
• "What are some good tools?" - Conversational inquiry
• "Generate 8 resources" - Unclear request without article intent
• "What plugins do you recommend?" - Question-only inquiry
• "生成 8 个资源" - 无明确文章生成意图的模糊请求
• "推荐几个好用的插件" - 单纯询问，不生成文章

🎯 Mode Selection and Priority

Default Mode: Intelligent Step-by-Step Generation

Mode Selection Priority:

1. User explicitly specifies mode → Use requested mode
2. User provides outline list → Use Outline-Based Generation
3. User only needs simple list → Use Quick Generation
4. Default fallback → Intelligent Step-by-Step Generation

When in doubt, always use Intelligent Step-by-Step Generation as it ensures highest quality through user confirmation.

🌟 Recommended Mode: Intelligent Step-by-Step Generation

This is the smartest, most controllable generation mode, ensuring article quality through multi-turn interaction.

Workflow

User inputs topic and quantity
  ↓
Step 0: Domain Detection (Auto-detect topic domain)
  ↓
Step 1: Collect candidate items (URL + recommendation reason + differentiating features + verifiable data)
  ↓
User confirms candidate list
  ↓
Step 2: Generate 6 title options (pain-point, exploratory, recommendation)
  ↓
User selects title
  ↓
Step 3: Generate complete article with domain-specific writing style
  ↓
Step 4: Ask about screenshots (Optional - ALWAYS ask after article generation)

Step 0: Domain Detection

Auto-detect the topic domain and apply corresponding writing template

Detection Rules

Detection Process

1. Extract keywords from user topic
2. Match against domain keyword library
3. Calculate confidence score based on keyword matches
4. If confidence ≥80%: Auto-select domain and inform user
5. If confidence 50-79%: Ask user to confirm
6. If confidence <50%: Use General template

Example Interaction

High Confidence (Auto-select):

### User Input
"生成一篇 Figma 插件推荐文章，推荐 6 个"
 
### AI Response
🔍 **检测到领域: Design (设计领域)**
 
基于关键词 "Figma" + "插件"，将使用设计领域专业模板，侧重：
-✅ 设计工作流集成
-✅ 创意效率提升
-✅ Design System 兼容性
 
[继续执行 Step 1: 收集候选项目...]

Medium Confidence (Ask User):

### User Input
"生成一篇图片压缩工具推荐文章"
 
### AI Response
🤔 **检测到多个可能的领域方向**
 
"图片压缩工具" 可以是：
-A. 面向设计师（设计工作流、批量处理、保持质量）
-B. 面向开发者（API 集成、自动化处理、性能优化）
 
请问您希望侧重哪个方向？(回复 A 或 B，或直接跳过使用通用模板)

Step 1: Collect Candidate Items

Based on user's topic requirements, collect matching projects/websites/tools and list them in the following format:

Candidate List Format:

**1. [Project Name]**
- URL: {{website link}}
- Recommendation reason: {{2-3 sentences explaining why it's worth recommending}}
- Differentiating features: {{How it differs from similar projects}}
- Data: {{Verifiable metrics - GitHub Stars, component counts, official numbers}}
 
**2. [Project Name]**
...

⚠️ DATA ACCURACY RULES (CRITICAL):

1. Only Use Verifiable Data

• ✅ USE: GitHub Stars, GitHub Forks, NPM weekly downloads, Component/Template counts (from official site)
• ✅ USE: Contributor count, Last updated date, License type, Tech stack details
• ❌ DON'T USE: User counts, Customer numbers, Efficiency percentages, Fake metrics
• ❌ DON'T USE: "100万+ 用户", "效率提升 10 倍", "转化率提升 25%" (unless officially verified)

2. How to Collect Data

• Visit GitHub page → Get Stars, Forks, Contributors, Last Updated
• Visit official site → Get component/template counts, feature lists
• Visit NPM page → Get weekly/monthly download counts
• If unsure about data → Either omit or mark as "approximate"

⚠️ CRITICAL RULES (MUST FOLLOW):

1. NEVER Hallucinate Projects or URLs

• ❌ DO NOT invent or fabricate project names, URLs, or websites
• ❌ DO NOT create fake resources or fictional tools
• ✅ ONLY recommend real, verifiable projects that actually exist
• ✅ If unsure about a project's existence, either skip it or ask the user

2. Verify URLs Before Listing

• ✅ Ensure URLs are accessible and valid
• ✅ If a URL seems suspicious or broken, note it for user verification
• ❌ DO NOT include projects with broken or dubious URLs

Important: If you encounter these situations, ask the user first:

• Topic direction is unclear
• Insufficient candidate items (inform how many you found)
• Multiple related sub-topics discovered, need to confirm focus
• Certain project URLs need verification
• Unable to find enough real, verifiable candidates (tell user and ask for guidance)

After listing candidates, ALWAYS wait for user confirmation, tell the user:

Above are {{number}} candidate items. Please confirm:
1. Do you need to adjust any items?
2. Do you need to add or remove any items?
3. Reply 'y' or 'yes' to confirm and proceed

Step 2: Generate Title Options

After user confirms the candidate list, generate 6 article titles in different styles based on these items:

### Please select your preferred title style:
 
**A. Pain-Point Type**
- A1. {{title}}
- A2. {{title}}
 
**B. Exploratory Type**
- B1. {{title}}
- B2. {{title}}
 
**C. Recommendation Type**
- C1. {{title}}
- C2. {{title}}
 
Please tell me which title you choose (e.g., A1), or suggest modifications.

Title Style References:

• Pain-Point Type: Say no to homogenization!, Break through bottlenecks!, Farewell to XXX
• Exploratory Type: Inspiration exploration!, In-depth analysis!, XXX exploration
• Recommendation Type: Curated, Essential, Must-know, Make XXX, XXX essentials

Step 3: Generate Complete Article

After user selects a title, generate the complete article.

First, determine the article folder path:

1. Translate the Chinese title to a descriptive English slug
2. Use lowercase, hyphen-separated format
3. Inform the user: "Article will be saved to: {{article-slug}}/index.md"
4. If using screenshots, they will be saved in the same folder

Example:

Title: 精选 4 款最受欢迎的渐变生成器，每个设计师都值得收藏
Slug: featured-4-most-popular-gradient-generators
Path: featured-4-most-popular-gradient-generators/index.md

Then proceed with generating the article content.

Article Structure Standards

File Organization

Article files should follow this structure:

{{article-slug}}/
└── index.md          # Article content
├── screenshot-1.png   # Screenshots (if any)
├── screenshot-2.png
└── ...

Path Explanation:

• Articles are saved in current working directory
• Folder name: {{article-slug}}/ (e.g., featured-4-most-popular-gradient-generators/)
• Article file: always index.md
• Screenshots: saved in the same folder as article

Cover Image Location (separate from article folder):

assets/blog/
└── cover-{{article-slug}}.jpg  # Manually added cover images

Naming rules:

• Article folder: lowercase English, hyphen-separated, descriptive and concise
• Example: shadcn-ui-component-libraries/, featured-4-most-popular-gradient-generators/
• Rule: Translate Chinese title to English, use lowercase, connect with hyphens
• Format: [adjective]-[number]-[category]-[items] or similar descriptive pattern
• Examples:
• "精选 4 款最受欢迎的渐变生成器" → featured-4-most-popular-gradient-generators/
• "设计师必备的 8 个 Figma 插件" → essential-8-figma-plugins-for-designers/
• "2026 年最佳 Vue 组件库" → best-vue-component-libraries-2026/
• File: always index.md
• Screenshots: saved in same folder as article, named after project (e.g., cssgradient-io.png)

1. Front Matter (Metadata)

---
title: {{user selected title}}
publishDate: {{today's date: YYYY-MM-DD}}
read: {{estimated reading time in minutes, typically 6-12}}
description: {{English subtitle, concise summary for SEO}}
tags:
  - {{English tag1}}
  - {{English tag2}}
  - {{Chinese tag1}}
  - {{Chinese tag2}}
img: "/assets/blog/cover-{{article-slug}}.jpg"
img_alt: "{{Chinese description of cover image}}"
---

Important Notes:

• img and img_alt are manually added cover images (not auto-generated)
• Image path should follow pattern: /assets/blog/cover-{{article-slug}}.jpg
• description should be in English, concise and SEO-friendly
• read is estimated reading time in minutes
• tags should include both Chinese and English keywords (4-6 items total)

2. Opening Paragraphs (3-4 paragraphs)

• Paragraph 1: Present industry background and current status
• Paragraph 2: Explain core value and significance of the topic
• Paragraph 3: Introduce the content to be recommended in the article
• Paragraph 4 (optional): Explain why these resources are being recommended

3. Body Content

Write content for each confirmed item in the following format:

### {{number}}. {{Project Name}}
 
-**官网**: [{{URL}}]({{URL}})
-**简介**: {{Core positioning, main functions, problems solved}}, covering {{specific content scope}}, {{special features/characteristics}}.
-**优势**: {{Why it's worth recommending, unique value, applicable scenarios, problems solved}}.

Body Requirements:

• Number of recommended items: As requested by user, typically 6-12
• Arrange in logical order (from basic to advanced, or from general to specialized)
• Description should explain core positioning and main content scope
• Advantages should highlight differentiated value and applicable scenarios
• Language should be professional but accessible, target reader appropriate