商品比价

v1.0.0

跨平台商品比价智能体技能。支持淘宝、京东、拼多多等主流电商平台的价格自动抓取、优惠规则解析与等效到手价计算。使用场景:用户需要比较同款商品在不同平台的价格时,如'比较 iPhone 16 Pro 在京东、天猫、拼多多的价格'。自动输出结构化比价报告,包含平台名称、当前价格、历史最低价、优惠信息及推荐购买建议。

0· 124·0 current·0 all-time
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
Name/description (跨平台商品比价) match what the files and SKILL.md do: locate products, run browser automation, parse promotions, normalize prices and output reports. Declared dependent skills (web-search-lite, automation-browser, doc-parser, file-manager) are appropriate for the described workflow.
Instruction Scope
SKILL.md instructs the agent to call browser automation and parsing skills and to run scrapers in an isolated ClawSpace container. It does not ask to read unrelated local files or external credentials. Note: the skill relies on active web scraping (anti‑scrape evasion techniques such as simulated delays/user‑agent rotation are mentioned) — this is within scope but has legal/TOS and ethical implications and requires login handling for some platforms, which is only mentioned as a caveat (not implemented).
Install Mechanism
No install spec is provided (instruction-only style). Scripts are included in the bundle but there is no external download or third‑party install step. No URLs for arbitrary code retrieval or archive extraction are present.
Credentials
The skill declares no required environment variables, no credentials, and no config paths. The code likewise does not reference external API keys or hidden endpoints. This is proportionate to the described scraping task (which typically doesn't require additional secrets).
Persistence & Privilege
always is false and model invocation is allowed (platform default). The skill does not declare writes to other skills or system‑wide settings. The SKILL.md claims execution inside ClawSpace containers, but that sandboxing claim is an operational guarantee of the platform, not enforced by the skill itself.
Assessment
This skill appears to do what it says: browser automation to scrape multiple Chinese e‑commerce platforms, parse promotions, normalize prices, and produce reports — and it does not request unrelated secrets or contact unknown servers. Before installing or enabling it, consider: 1) Compliance and TOS: active scraping with anti‑bot evasion can violate site terms or local law — confirm acceptable usage. 2) Sandboxing: SKILL.md asserts ClawSpace isolation; verify your platform actually enforces that isolation before running untrusted scraping code. 3) Missing pieces: many scraper functions are TODO stubs (login handling, detailed DOM extraction) — expect further implementation or debugging. 4) Rate limiting & privacy: run on non‑sensitive data, avoid high request rates, and do not provide platform login credentials unless you trust both this skill and the automation skill; the skill currently does not request credentials but may require them for some platforms. 5) Review dependent skills (automation-browser, doc-parser) and test in a controlled environment. If you need higher assurance, request the author to remove TODOs, provide tests, and document sandbox guarantees.

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

latestvk97aw2rszbqk1wt2s1mp3tx5c983t69y
124downloads
0stars
1versions
Updated 2w ago
v1.0.0
MIT-0

商品比价智能体技能

核心功能

本技能提供端到端的跨平台商品比价能力,实现从"人工比价"到"智能择优"的自动化流程。

解决的问题

痛点解决方案
信息碎片化统一采集多平台数据,标准化展示
操作繁琐一句话指令自动完成全流程
优惠规则复杂自动解析满减、赠品等非标优惠
实时性不足实时抓取最新价格
缺乏决策依据提供历史价格趋势与推荐建议

适用场景

  • 数码产品选购:手机、耳机等高单价电子产品比价
  • 日用品批量采购:洗衣液、纸巾等高频消费品价格追踪
  • 电商大促:618、双 11 期间实时锁定最低价
  • 企业采购:办公设备、耗材多源比价
  • 海淘分析:海外购与国内产品性价比对比

依赖技能

以下技能必须可用:

技能用途
web-search-lite商品定位与关键词提取
automation-browser浏览器自动化,访问电商平台
doc-parser非结构化优惠信息语义解析
file-manager比价报告管理
skill-vetter安全校验

模型配置

推荐配置:

{
  "model": {
    "provider": "alibaba-cloud",
    "defaultModel": "gateway/qwen3.5-plus",
    "parameters": {
      "temperature": 0.5,
      "maxTokens": 4096,
      "topP": 0.7
    }
  }
}

快速响应配置:

{
  "model": {
    "provider": "alibaba-cloud",
    "defaultModel": "bailian/qwen-turbo",
    "parameters": {
      "temperature": 0.5,
      "maxTokens": 4096,
      "topP": 0.7
    }
  }
}

工作流程

1. 意图解析

用户输入自然语言指令,如:

  • "比较 iPhone 16 Pro 在京东、天猫、拼多多的价格"
  • "帮我看看这款洗衣机哪个平台最便宜"

2. 商品定位

调用 web-search-lite 提取:

  • 商品名称/型号
  • 目标平台列表
  • 关键规格参数

3. 价格抓取

启动 automation-browser 在 ClawSpace 环境中:

  • 依次打开目标电商平台
  • 搜索目标商品
  • 提取关键字段:
    • 商品标题
    • 基础价格
    • 促销信息(满减、折扣)
    • 运费政策
    • 配送时效

4. 优惠解析

使用 doc-parser + Qwen 模型解析:

  • "买一送一" → 单价 ÷ 2
  • "前 N 名半价" → 条件性折扣
  • "满减券" → 反推单品实际支付金额
  • 赠品价值 → 按市场均价折算

5. 价格归一化

执行三步归一化:

  1. 统一货币单位为人民币
  2. 根据满减规则计算等效到手价
  3. 剔除运费干扰,计入赠品价值

6. 商品去重匹配

通过多维特征哈希实现跨平台商品对齐:

  • 品牌 + 型号 + SKU 编码
  • 主图视觉相似度
  • 标题关键词权重
  • 匹配度 > 阈值 → 判定为同一商品

7. 报告生成

输出结构化比价报告:

## 商品比价报告

**商品**: iPhone 16 Pro 256GB 黑色

| 平台 | 当前价格 | 历史最低 | 优惠信息 | 配送时效 |
|------|----------|----------|----------|----------|
| 京东 | ¥7,999 | ¥7,699 | 白条减 200 | 次日达 |
| 天猫 | ¥8,199 | ¥7,899 | 88VIP 95 折 | 2-3 天 |
| 拼多多 | ¥7,599 | ¥7,399 | 百亿补贴 | 3-5 天 |

**推荐**: 拼多多(当前最低价,省¥400)

反爬策略

采用模拟真实用户行为策略:

  • 随机延迟(1-3 秒)
  • 鼠标轨迹模拟
  • 滚动加载触发
  • User-Agent 轮换

安全隔离

所有操作在 ClawSpace 独立容器中执行:

  • 6 核 CPU + 12GB 内存
  • 与用户本地系统完全隔离
  • 敏感数据沙箱处理
  • 异常不影响终端安全

触发示例

以下用户指令应触发本技能:

  • "比较 [商品名] 在 [平台 A]、[平台 B] 的价格"
  • "[商品] 哪个平台最便宜"
  • "帮我看看 [商品] 的跨平台比价"
  • "[商品] 京东和拼多多哪个更划算"
  • "查一下 [商品] 的历史价格走势"

输出格式

标准比价表格

字段说明
平台名称京东/天猫/拼多多/苏宁等
当前价格等效到手价(人民币)
历史最低近 90 天最低价
优惠信息满减、折扣、赠品等
配送时效预计送达时间
推荐标识标注最低价平台

购买建议

根据以下维度给出推荐:

  1. 价格优先:等效到手价最低
  2. 时效优先:配送最快
  3. 服务优先:售后保障最好
  4. 综合推荐:性价比最优

错误处理

错误类型处理方式
商品未找到建议用户提供更详细的型号信息
平台无法访问跳过该平台,继续其他平台抓取
价格解析失败使用原始标价,标注"优惠待确认"
匹配度不足提示用户手动确认是否为同一商品

参考资料

  • 电商平台 DOM 结构:见 references/platform-dom.md
  • 优惠规则解析模板:见 references/promotion-rules.md
  • 价格归一化算法:见 references/normalization.md

Comments

Loading comments...