SKILL.md

Skill Name

product_to_assets_skill

Type

Multimodal Product → Marketing Asset Compiler Skill

1. Overview

This Skill converts any product input (URL, image, video, or text) into publish-ready marketing assets:

• Marketing Posters
• Product Videos (shot-based scripts)
• User Manuals

2. Supported Inputs

• URL (product pages, landing pages)
• Image (UI screenshots, PPT, design drafts)
• Video (product demos, walkthroughs)
• Text (product concept / PRD)

  小葱技能7w4.net持续更新中。

3. Input Schema

{
  "input_type": "url | image | video | text",
  "content": "string",
  "goal": "launch | explain | pitch | educate",
  "tone": "marketing | product | cinematic | technical"
}

4. Core Pipeline

Input → Multimodal Parsing → Product Representation Object (PRO) → Narrative Engine → Parallel Generation ├ Poster Engine ├ Video Engine └ Manual Engine → Validation → Export

5. Product Representation Object (PRO)

{
  "product_name": "",
  "core_function": "",
  "user_problem": "",
  "target_audience": "",
  "key_features": [],
  "workflow": [],
  "visual_style": "minimal | vibrant | enterprise | playful",
  "tone_cues": ""
}

6. Output Schema

{
  "product_brief": {},
  "poster": {},
  "video": {},
  "manual": {}
}

7. Poster Output

• headline
• subheadline
• layout type (hero / split / feature / before-after)
• visual prompt (AI image generation ready)
• 2–3 variants

8. Video Output

• 30s structure
• Scene-based breakdown:
• Hook (0–3s)
• Problem (3–10s)
• Solution (10–20s)
• Feature (20–25s)
• CTA (25–30s)

Each scene includes: - visual - voiceover - screen text

9. User Manual Output

• Quick Start
• Step-by-step guide
• FAQ
• Troubleshooting

10. Prompt Design Rules

• No hallucination
• Structured JSON output required
• Visual prompts must be generation-ready
• Maintain consistency across all assets

11. Error Handling

• URL failure → fallback to text extraction
• Image failure → OCR + vision fallback
• Video failure → keyframe extraction fallback
• Generation failure → retry with simplified prompt

12. Key Value

• Converts ANY product into marketing assets
• Multimodal input support
• Structured AI pipeline
• Ready for automation / SaaS integration

Version

v5.0 - Shareable Production Skill Definition

这个Skill的设计理念不错，能支持多种形式的产品输入并自动生成营销素材，输出结构也很完整。但目前更像是一个设计文档，缺少实际可用的代码实现。如果你需要一个完整的产品营销素材生成工具，这个版本可能还需要开发者进一步补充代码；如果只是作为设计参考，文档质量还是值得肯定的。整体属于半成品状态，潜力有但完成度一般。