Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

cosmetics-shopping

v1.0.0

美妆电商购物助手,支持商品搜索、加购下单、订单管理;当用户需要购买美妆产品、查询订单或管理收货地址时使用

0· 94·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for bvcg204/beautiful-shop.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "cosmetics-shopping" (bvcg204/beautiful-shop) from ClawHub.
Skill page: https://clawhub.ai/bvcg204/beautiful-shop
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install beautiful-shop

ClawHub CLI

Package manager switcher

npx clawhub@latest install beautiful-shop
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The skill claims to be a cosmetics shopping assistant and the scripts implement search, cart, order, address and auth APIs against https://service.filtalgo.com, which is coherent. However the scripts expect an access token to be provided via an environment variable named COZE_FILTALGO_API_7618877822249025562, yet the registry metadata lists no required env vars and SKILL.md does not document that the token must be stored in that env var. That discrepancy is unexpected and unexplained.
Instruction Scope
SKILL.md instructs the agent to run included python scripts (search, login, add to cart, create order, etc.) and does not ask the agent to read arbitrary local files or unrelated credentials. But SKILL.md omits operational details for token persistence (it shows login returns accessToken/refreshToken but does not say how to persist them for subsequent script calls).
Install Mechanism
There is no install spec (instruction-only deployment with bundled scripts). No downloads or external packages are installed by an installer, so install-time risk is low. Files are included in the skill bundle, and all code is plain Python that performs network requests.
!
Credentials
Although the registry declares no required credentials, most scripts read an environment variable COZE_FILTALGO_API_7618877822249025562 for the accessToken. This is a required secret for the skill to function but it's not declared in requires.env or explained in SKILL.md. That mismatch increases risk because it's unclear where/how the token will be stored, who can access it, and whether the platform will expose it to other skills or processes.
Persistence & Privilege
The skill does not request always:true or modify other skills. It requires network access to a single domain and will perform authenticated operations (addresses, orders, payments). Combined with the undeclared env var behavior, autonomous invocation (default) increases potential blast radius — the agent could perform account actions if the token is available, so verify runtime token handling before enabling autonomous use.
What to consider before installing
What to check before installing: - Token handling: The Python scripts read an environment variable named COZE_FILTALGO_API_7618877822249025562 for accessToken, but the skill metadata and SKILL.md do not declare or document setting this env var. Ask the publisher or the platform how login tokens are persisted and who/what can read that environment variable. Do not assume the platform will safely store the token. - Sensitive actions: The skill performs authenticated actions (add address, create orders, cancel orders, return payment links). Only install if you trust the remote domain (service.filtalgo.com / filtalgo.com) and the skill owner — the package has no homepage and owner is unknown. - Test safely: If you proceed, test with a throwaway account and small-value operations first. Confirm that tokens are scoped to this skill only and are not leaked to other skills or processes. - Autonomy: The skill can be invoked autonomously by the agent (platform default). If you don't want the agent to place orders on your behalf, restrict autonomous invocation or require explicit user confirmation for any purchase/checkout flows. - Ask for missing docs: Request the author/platform to (1) declare required env vars in registry metadata, (2) document how tokens are stored/rotated (and provide a refresh script), and (3) provide publisher/homepage or a verified source for the API endpoints.

Like a lobster shell, security has layers — review code before you run it.

latestvk972a094swbj3nchpacefex4ah83m3fn
94downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

美妆电商购物助手

任务目标

  • 本 Skill 用于:完整的电商购物流程
  • 能力包含:用户登录、商品搜索、商品详情、加购/立即购买、下单、订单管理、物流查询、收货地址管理
  • 触发条件:用户表达购买意图、查询商品、查看订单、管理地址等场景

前置准备

  • 用户需先完成短信验证码登录
  • 部分接口需要 accessToken 认证

核心流程

一、用户登录

触发时机:用户首次使用或需要下单时

执行步骤

  1. 调用 scripts/auth/send_sms.py <手机号> 发送验证码
  2. 智能体提示用户输入验证码
  3. 调用 scripts/auth/sms_login.py <手机号> <验证码> 完成登录
  4. 登录成功后获取 accessToken 和 refreshToken

测试环境:验证码固定为 111111

二、商品搜索

执行方式:调用脚本

python scripts/goods/search.py --keyword "洗面奶" --page 1 --size 10 --sort PRICE_ASC

参数说明

  • --keyword:搜索关键词
  • --page:页码,默认1
  • --size:每页数量,默认10
  • --sort:排序方式(PRICE_ASC/PRICE_DESC/SALE_DESC)

返回信息:商品列表包含 goodsId、skuId、goodsName、price、thumbnail、storeName 等

三、商品详情

执行方式:调用脚本

python scripts/goods/detail.py <goodsId> <skuId>

返回信息:包含商品详细规格、SKU列表、价格、库存等

四、购物车管理

加购python scripts/cart/add.py <skuId> <数量> BUY_NOW|CART 查看python scripts/cart/list.py 数量python scripts/cart/count.py

五、下单流程

立即购买(单个商品):

  1. add_to_cart:加购(way=BUY_NOW)
  2. set_address:设置收货地址
  3. preview_order:预览订单确认价格
  4. create_order:创建订单获取 trade_sn
  5. 返回支付链接

购物车结算(多个商品):

  1. cart_list:查看购物车
  2. set_address:设置收货地址
  3. preview_order:预览订单
  4. create_order:创建订单

六、订单管理

订单列表python scripts/order/list.py --status UNPAID --page 1 订单详情python scripts/order/detail.py <orderSn> 取消订单python scripts/order/cancel.py <orderSn> <原因> 物流查询python scripts/order/logistics.py <orderSn>

七、收货地址管理

地址列表python scripts/address/list.py 默认地址python scripts/address/default.py 解析地址python scripts/address/resolve.py <省> <市> <区> [街道] 新增地址python scripts/address/add.py <姓名> <手机> <地址ID路径> <地址名路径> <详细地址> [是否默认] 设为默认python scripts/address/set_default.py <addressId>

资源索引

认证脚本

脚本功能
scripts/auth/send_sms.py发送短信验证码
scripts/auth/sms_login.py短信登录

商品脚本

脚本功能
scripts/goods/search.py搜索商品
scripts/goods/detail.py商品详情

购物车脚本

脚本功能
scripts/cart/add.py加购物车/立即购买
scripts/cart/list.py查看购物车
scripts/cart/count.py购物车数量

订单脚本

脚本功能
scripts/order/set_address.py设置收货地址
scripts/order/preview.py预览订单
scripts/order/create.py创建订单
scripts/order/list.py订单列表
scripts/order/detail.py订单详情
scripts/order/cancel.py取消订单
scripts/order/logistics.py物流查询

地址脚本

脚本功能
scripts/address/list.py地址列表
scripts/address/default.py默认地址
scripts/address/resolve.py解析地址
scripts/address/add.py新增地址
scripts/address/set_default.py设为默认

参考文档

典型对话示例

示例1:用户想购买商品

用户:我想买一款洗面奶

智能体:
1. 调用搜索脚本查找洗面奶
2. 展示商品列表(图片+价格+店铺)
3. 询问用户选择哪款商品
4. 用户确认后,调用商品详情脚本展示规格

用户:就选第一款

智能体:
1. 检查是否已登录
2. 未登录:发送验证码让用户登录
3. 登录成功后:
   - 调用加购(BUY_NOW)
   - 设置收货地址
   - 预览订单
   - 创建订单,获取 trade_sn
4. 返回支付链接给用户:
   - APP支付:`https://app-buyer.filtalgo.com/pages/mine/payment/payOrder?trade_sn={trade_sn}`
   - H5支付:`https://buyer.filtalgo.com/payment/cashier?paymentScene=TRADE&orderSn={trade_sn}`

示例2:用户查询订单

用户:查看我的订单

智能体:
1. 调用订单列表脚本
2. 展示订单列表(包含订单号、状态、金额、时间)
3. 询问用户需要什么操作(查看详情/取消/查物流)

示例3:新增收货地址

用户:添加一个新地址

智能体:
1. 询问收货人信息(姓名、手机、省市区、详细地址)
2. 用户提供后:
   - 调用解析地址脚本获取 ID 路径
   - 调用新增地址脚本
3. 确认添加成功

注意事项

  1. Token 管理:accessToken 有效期约 25 天,refreshToken 约 45 天
  2. 订单号区分
    • T 开头:交易单号(tradeSn),用于支付
    • O 开头:子订单号(orderSn),用于查询/取消
  3. 地址类型:新增地址必须传 type=RECEIVE
  4. way 参数:BUY_NOW=立即购买(单商品),CART=购物车模式
  5. 商品详情页
    • H5:https://buyer.filtalgo.com/goodsDetail?goodsId={goodsId}&skuId={skuId}
    • APP:https://app-buyer.filtalgo.com/pages/goods/product/detail?goodsId={goodsId}&skuId={skuId}

Comments

Loading comments...