Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

AI × Cross-Border E-Commerce Knowledge Hub

An AAAI China Chapter open-source project

A hands-on AI manual for cross-border e-commerce — 69 chapters covering every step from product research to growth, each with copy-paste prompts.

This is more than a book. dist/ is a plug-and-play agent capability package — a 100-entity domain ontology, 9 installable skills, and an MCP server integration. Claude Code users can install it with two commands; the source is on GitHub.

🌐 Every chapter is available in all three languages. Use the language switcher (top right) to jump between 中文 / EN / 日本語 on any page — it keeps you on the same chapter.

Try It First

Copy this into ChatGPT or Claude and get results in 30 seconds:

You are a senior cross-border e-commerce expert with deep knowledge of the Amazon marketplace.
I want to sell a portable neck fan on Amazon US.
Please provide a quick market feasibility analysis including:
1. Category characteristics (seasonality, competition level, price range)
2. Top 3 competitors' key selling points and main pain points from negative reviews
3. 3 potential differentiation angles
4. Risk alerts (compliance, patents, seasonal inventory risks)
Present key data comparisons in table format.

<data_discipline>
- Specific figures or facts about market data, search volume, competitor performance, regulatory text, or fee rates must come from what I supplied. **Don't fill gaps from memory** — these facts move fast and your version may be stale
- When you need a fact to make a judgment, tell me which official source to verify it against, then stop and ask me
- Tag every conclusion with its source: [supplied by me] or [model inference]
</data_discipline>

How This Hub Is Organized

Six tracks:

TrackForCovers
FoundationsEveryoneAI literacy, prompt engineering, agents, RAG, RPA
OperatorsOperations rolesProduct research, listings, ads, customer service, compliance, finance
DevelopersEngineersData pipelines, prediction models, RAG, agents, MCP
ManagersTeam leadsCapability assessment, team building, ROI, risk governance
MarketplacesAll rolesHands-on guides for 13 e-commerce platforms
Social MediaAll rolesAI playbooks for 7 social channels

Want to see what AI can (and can’t) do first? Start with the AI Landscape Assessment.

About the Prompts

All prompt templates in this hub were tested against:

  • The T2 workhorse tier of ChatGPT / Claude / Gemini — re-verified July 2026 (current model ids in the model matrix)
  • Claude (Opus 4 / Sonnet 4) — March 2026

Results can vary across models. If a prompt underperforms, try another model or add more context at the top of the prompt. AI models iterate quickly — re-validate prompts you rely on periodically.