WishlistAI Agent Skill

A wishlist management API that allows AI agents to help users create, manage, and share gift wishlists.

Quick Start

Base URL: https://wishlist-app-production.up.railway.app

Authentication:

•Login: POST /api/auth/login with { phoneNumber, password }
•Generate API Key: POST /api/users/me/apikey (Bearer token required)
•Use API Key: Add x-api-key: <your-key> header to all requests

Available Actions

Wishlists

Action	Endpoint	Description
List all	`GET /api/wishlists`	Get all user's wishlists
Create	`POST /api/wishlists`	Create new wishlist `{ title }`
Get one	`GET /api/wishlists/{id}`	Get wishlist with items
Delete	`DELETE /api/wishlists/{id}`	Delete a wishlist

Items

Action	Endpoint	Description
Add by name	`POST /api/wishlists/{id}/items`	Add item `{ name, price?, notes? }`
Add by URL	`POST /api/wishlists/{id}/items/url`	Scrape product from URL `{ url }`
Update	`PUT /api/items/{id}`	Update item details
Delete	`DELETE /api/items/{id}`	Remove item

Example Workflow

bash

# 1. Login and get token
curl -X POST https://wishlist-app-production.up.railway.app/api/auth/login \
  -H "Content-Type: application/json" \
  -d '{"phoneNumber":"0912345678","password":"Password123"}'

# 2. Generate API Key (use token from step 1)
curl -X POST https://wishlist-app-production.up.railway.app/api/users/me/apikey \
  -H "Authorization: Bearer <token>"

# 3. Create a wishlist
curl -X POST https://wishlist-app-production.up.railway.app/api/wishlists \
  -H "x-api-key: <api-key>" \
  -H "Content-Type: application/json" \
  -d '{"title":"Birthday Gifts"}'

# 4. Add an item from URL (auto-scrapes product info)
curl -X POST https://wishlist-app-production.up.railway.app/api/wishlists/1/items/url \
  -H "x-api-key: <api-key>" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://www.momoshop.com.tw/goods/GoodsDetail.jsp?i_code=14244558"}'

OpenAPI Specification

Full API documentation: /api/swagger.json

Supported E-commerce Platforms

•Shopee (蝦皮)
•Momo (momo購物網)
•PChome
•And any URL with Open Graph metadata

🛠️ 開發與品質守則 (Mandatory)

所有 AI 代理（包含 Claude CLI）在執行任何代碼修改或部署前，必須遵守以下守則：

•查閱參考資料：必須先讀取並理解 /.agent/skills/ 與 /.agent/workflows/ 中的所有檔案。
•品質保護網：這些文件定義了產品的架構標準與品質檢查流程，任何變更不得與其衝突。
•穩定性優先：嚴格執行 workflows/ 定義的 Pre-work 與檢查步驟。

Model Context Protocol (MCP) Integration

This application now supports MCP, allowing AI agents (like Claude) to access wishlist data directly as resources.

Resources

•wishlist://all: Returns a JSON list of all public wishlists and their items.

Tools

•create_wishlist: Creates a new wishlist. Parameters: { title, userId }.
•add_item: Adds an item to a wishlist. Parameters: { wishlistId, name, price?, link?, notes? }.

Running the MCP Server

To use this with Claude Desktop or other MCP clients, configure it with the following command:

bash

node /path/to/wishlist-app/server/dist/mcp/index.js

Note: Ensure you have built the project (npm run build) before running.