Agent Skills › firecrawl/web-agent › e-commerce

e-commerce

GitHub

用于导航电商网站，提取商品、价格、分类及库存数据。支持分页处理、变体解析及JS渲染站点，优先通过API或Sitemap获取数据，确保数据完整性与准确性。

agent-core/src/skills/definitions/e-commerce/SKILL.md firecrawl/web-agent

Trigger Scenarios

需要抓取电商网站商品列表或详情监控商品价格变动或库存状态分析竞品电商数据

Install

npx skills add firecrawl/web-agent --skill e-commerce -g -y

More Options

Non-standard path

npx skills add https://github.com/firecrawl/web-agent/tree/main/agent-core/src/skills/definitions/e-commerce -g -y

Use without installing

npx skills use firecrawl/web-agent@e-commerce

指定 Agent (Claude Code)

npx skills add firecrawl/web-agent --skill e-commerce -a claude-code -g -y

安装 repo 全部 skill

npx skills add firecrawl/web-agent --all -g -y

预览 repo 内 skill

npx skills add firecrawl/web-agent --list

SKILL.md

Frontmatter

{
    "name": "e-commerce",
    "category": "E-commerce",
    "description": "Navigate e-commerce sites to extract products, pricing, categories, and inventory. Handles pagination, variants, and JS-heavy storefronts."
}

E-commerce Extraction

General Patterns

Check for sitemap.xml first -- many stores list all product URLs
Look for /products.json, /api/products, or similar API endpoints before scraping HTML
Product listing pages usually paginate: look for ?page=N, ?offset=N, or "Load more" buttons
Always check the total count shown on the page vs what you've extracted

Product Data Checklist

Name, brand, SKU/ID
Price (current, original/compare-at, currency)
Variants (size, color, etc.) with per-variant pricing and availability
Images (primary + gallery)
Description (short + long)
Category/breadcrumb path
Availability/stock status
Ratings and review count

Pagination

Check for: next/prev links, page numbers, "showing X of Y" text
For infinite scroll: use interact to scroll and load more
Keep count: if page says "200 products" and you have 24, keep going

JS-Heavy Sites

Use interact for sites that require JavaScript rendering
Look for XHR/API calls in the page that return JSON -- often easier than parsing HTML
Single-page apps often have internal APIs at predictable paths