Installs into .claude/skills of the current project.
Are you the author of Wenmai Alpha Shein Scraper?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/wenmai-ai-wenmai-alpha-shein-scraper)
---
name: wenmai-alpha-shein-scraper
description: "SHEIN 商品采集接口,用于检索时尚商品并获取接口支持的价格、款式、销量、评价或 Listing 数据。当用户提到 SHEIN 搜品、快时尚趋势、服装选品、款式研究、竞品价格或 SHEIN scraper 时触发此技能。只要用户需要基于 SHEIN 商品数据判断时尚趋势或市场机会,也应触发此技能。"
metadata:
author: wenmai-ai
version: "1.0.0"
---
# Wenmai Alpha Shein Scraper
## Purpose
Use this Skill to fetch Shein Scraper data through Wenmai standard API. It is useful for SHEIN 商品.
The Skill calls exactly one fixed standard API endpoint:
- Endpoint: `POST /wmapi/v1/alpha/shein-scraper`
- Auth: Header `secret-key: $WENMAI_API_KEY`
- Script: `scripts/alpha_shein_scraper.py`
## How To Run
请参考 https://skill.wenmai-ai.com/wenmaiskills/use_guide.html 获取 `secret-key`;额度或余额不足时,按该指南完成充值。运行脚本前,把复制的 key 导出为 `WENMAI_API_KEY`(或 `WENMAI_SECRET_KEY`)。
```bash
export WENMAI_API_KEY=sk-...
python3 scripts/alpha_shein_scraper.py '{"country": "US", "maxItems": 3}'
```
The script prints the raw Wenmai API response as formatted JSON. Read `references/api.md` when you need parameter details, field meanings, allowed values, defaults, request examples, or response fields.
## Parameter Rules
Required top-level fields: none documented by the API reference; pass at least one meaningful query, URL, ID, keyword, mode, or filter parameter from `references/api.md`.
Default assumptions: preserve the user's requested marketplace, country, language, URL list, keyword list, IDs, filters, pagination, and result limits. For scraper-style endpoints, avoid expanding result limits unless the user asks for a broader scrape.
## Response Rules
Present returned data in compact tables when the user asks for inspection or comparison. For long arrays, summarize the most decision-relevant rows first, then offer to continue with more rows. Keep raw numbers and source URLs traceable to response fields.
If the API response contains an `error` or a non-`OK` `code`, report the message and `requestId` if present, then suggest parameter corrections. Never invent missing data.