阿里各种版本的龙虾都来了
钉钉 的 悟空
Accio 的Accio Work
看了下关于 阿里国际站经营分析 SKILL技能 ，
逻辑是从 https://crm.alibaba.com/crmlogin/aisales/dingwukong/diagnoseData.json
接口中获取 店铺诊断信息
也就是后台首页的 https://i.alibaba.com/
看到的AI客户经理 的 店铺诊断建议 以及服务报告的中内容
交给模型去分析，输出结构化内容 ，包含四个板块报告：
-
店铺数据概览：以表格形式展示点击、询盘、优品率等（含环比和同行对比）。
-
诊断总结：一段感性的文字，告诉你哪里做得好，哪里拖了后腿。
-
商家任务：具体的行动建议（如：发布 5 个潜力品）。
-
完整周报链接：一个点击即可跳转到阿里后台详情页的 URL。
阿里国际站经营周报数据分析 SKILL [中文版]
# 阿里国际站经营周报数据分析 SKILL
## 概述
通过浏览器访问阿里国际站，自动获取并分析店铺经营周报数据，输出结构化的经营诊断。
## 简易流程
```
Step 1: 判断登录状态
│ 访问 i.alibaba.com，检测是否跳转登录页
│ ├─ 未登录 → 引导用户登录，等待完成
│ └─ 已登录 ↓
│
Step 2: 获取经营数据
│ 调用 diagnoseData.json 接口
│ ↓
Step 3: 数据校验（静默）
│ 校验 data / encryptedReportId / values.receipt
│ ├─ 校验失败 → 输出兜底文案，终止
│ └─ 校验通过 ↓
│
Step 4: 结构化展示
│ 一次性输出：数据总览 → 诊断总结 → 商家任务 → 周报链接
│ ↓
Step 5: 获取完整报告数据
│ 调用 queryWeekReportAllData.json 接口
│ 提示"全量数据解读完毕"，等待用户提问后按需解读
```
---
## 环境与前置要求
| 项目 | 说明 |
| ---------- | ----------------------------------------------------------------------------- |
| 浏览器工具 | 使用 `browser` 工具，用户指定 profile 则用指定的，否则默认 `profile=openclaw` |
| 目标域名 | `https://i.alibaba.com/` |
| 登录状态 | 用户需已登录，未登录则引导完成登录 |
---
## 核心流程
### Step 1：判断登录状态
访问 `https://i.alibaba.com/`：
- 跳转至 `https://login.alibaba.com/newlogin/icbuLogin.htm...` → 未登录，提示用户完成登录后告知
- 未跳转 → 已登录，继续后续流程
### Step 2：获取经营数据
在浏览器 console 中执行：
```javascript
(async () => {
const url = 'https://crm.alibaba.com/crmlogin/aisales/dingwukong/diagnoseData.json';
const r = await fetch(url, {
method: 'POST',
credentials: 'include'
});
if (!r.ok) throw new Error(`HTTP 错误：${r.status}`);
return await r.json();
})()
```
### Step 3：数据校验（静默执行，不向用户展示）
按以下顺序校验，任一不通过则输出兜底文案并终止：
1. 接口返回为空 / `null` / `data` 为空对象或空数组 → 终止
2. `encryptedReportId` 字段不存在或为空 → 终止
3. `values.receipt` 字段不存在或为空 → 终止（该字段在 Step 5 中作为请求参数使用，如果有值需记录下来供后续使用）
兜底文案：
> 抱歉，当前账号暂无经营数据，无法进行分析。
容错处理：
- `weekDiagnose` 缺失 → 注明"本期数据未包含经营诊断内容"，继续输出其他数据，不终止
- `diagnoseTitle` / `diagnoseSummary` 均为空 → 该诊断项跳过，继续处理其他内容
### Step 4：结构化数据展示
从 `values.weekDiagnose` 数组中筛选 `scope = "近 7 天"` 的数据对象，结合 `values.diagnoseSummary`，按以下四部分输出。
#### 第一部分：店铺数据总览
从 `indicatorList` 数组提取核心指标，以 Markdown 表格展示。
字段映射：
| 字段 | 含义 |
| ------------ | ------------ |
| `name` | 数据指标名称 |
| `value` | 本周数据 |
| `cycleCRC` | 环比上周 |
| `valueVsAvg` | 较同行平均 |
示例输出：
> 老板您好！您的店铺近 7 日数据如下：
| 数据指标 | 本周数据 | 环比上周 | 较同行平均 |
| ---------- | -------- | -------- | ---------- |
| 点击量 | 6.0 | +0.0% | -96.4% |
| 商机量 | 10.0 | +0.0% | -65.5% |
| 优爆品占比 | 0.0% | - | - |
> 当 `cycleCRC` 或 `valueVsAvg` 缺失时，展示为 `-`
#### 第二部分：店铺诊断总结
直接展示 `values.diagnoseSummary` 字段内容。
#### 第三部分：商家任务
遍历 `maTaskList` 数组，提取 `title` 和 `desc`（需清理 HTML 标签）。
#### 第四部分：周报链接
使用 `encryptedReportId` 拼接链接：
```
https://crm.alibaba.com/crmagent/crm-grow/luyou/report-render.htmlid={encryptedReportId}&dateScope=week&isDownload=false
```
#### 展示要求
- 四部分内容收集完成后，一次性展示给用户
- 不要将内容放到 markdown 文件中
- 返回完内容后，继续执行 Step 5
### Step 5：获取商家报告的完整数据
Step 3 数据校验通过后，使用 Step 2 返回数据中的 `values.receipt` 调用报告详情接口，获取完整的商家报告数据
在浏览器 console 中执行：
```javascript
(async () => {
const receipt = '<Step 2 返回数据中获取的 values.receipt>';
const url = 'https://crm.alibaba.com/crmlogin/aisales/dingwukong/queryWeekReportAllData.json';
const r = await fetch(url, {
method: 'POST',
credentials: 'include',
headers: { 'Content-Type': 'application/x-www-form-urlencoded' },
body: `receipt=${receipt}`
});
if (!r.ok) throw new Error(`HTTP 错误{r.status}`);
return await r.json();
})()
```
#### 数据校验（静默执行，不向用户展示
- 接口返回为空 / `null` / `data` 为空对象 提示"抱歉，暂时无法获取完整报告数据，请稍后再试并终- 部分模块数据缺失 跳过该模块，继续处理其他模块，不终止
#### 数据理解与存
接口返回的数据按 `reportPageCode` 分模块组织，各模块编码及字段含义详见 `reference/weekly_report_data_explanation.md`
主要模块一览：
| 模块编码 | 说明 |
| ------------------------------------- | ------------------------------------------------ |
| STORE_DATA_OVERVIEW | 店铺数据总览（搜索曝光、点击、商机、转化率等） |
| STORE_INFRASTRUCTURE_SITUATION | 店铺基建情况（品量、回复率等） |
| STORE_DIAGNOSIS | 店铺诊断（星等级、能力项评分及优化建议） |
| OPPORTUNITY_STAR_LEVEL_DATA_OVERVIEW | 商机星等级数据总览 |
| TRADE_STAR_LEVEL_DATA_OVERVIEW | 交易星等级数据总览 |
| ACTION_SUGGESTION | 行动建议（每日商机与成交趋势、任务列表） |
| PRODUCT_DATA_OVERVIEW | 商品数据总览（产品分层、分Top5 |
| EXPOSURE_TOP10_PRODUCT_DATA | 曝光 Top10 商品数据 |
| CATEGORY_EXPANSION_SUGGESTION | 品类扩充建议 |
| BUYER_DISTRIBUTION_DATA | 买家分布数据（曝进店/询单维度 |
| P4P_PLAN_OPTIMIZ_SUGGESTION | 直通车计划优化建议 |
| P4P_SEARCH_WORD_OPTIMIZ_SUGGESTION | 直通车搜索词优化建 |
| FLOW_SOURCE_CHANNEL_ANALYSIS | 流量来源渠道与分 |
| STORE_DETAIL_12_MONTHS | 12 个月店铺效果明细 |
| STORE_COMMUNICATION_CONVERSION_OVERVIEW | 沟通转化数据总览 |
| STORE_ACCOUNT_DATA_OVERVIEW | 店铺分账号数据总览 |
| EMPLOYEE_MANAGEMENT | 员工管理 |
> 完整字段映射及含义参考：`reference/weekly_report_data_explanation.md`
#### 数据获取后的行为
获取完整报告数据后，仅提示：
> 全量数据解读完毕。您可以问我有关店铺数据的相关问题~
等待用户提出具体问题后再基于完整数据进行针对性回答。不主动向用户展示全量数据内容，将数据保留在上下文中按需解读
#### 数据应用示例
| 用户提问 | 对应数据模块 |
| ---------------------- | ------------------------------------------------------------------------------------------------ |
| 我的店铺流量来源是什么？ | `FLOW_SOURCE_CHANNEL_ANALYSIS` |
| 直通车效果怎么样？ | `P4P_DATA_OVERVIEW_1` / `P4P_DATA_OVERVIEW_2` + `P4P_PLAN_OPTIMIZ_SUGGESTION` |
| 商品表现如何 | `PRODUCT_DATA_OVERVIEW` + `EXPOSURE_TOP10_PRODUCT_DATA` |
| 买家都来自哪里？ | `BUYER_DISTRIBUTION_DATA` |
| 星等级情况？ | `STORE_DIAGNOSIS` + `OPPORTUNITY_STAR_LEVEL_DATA_OVERVIEW` + `TRADE_STAR_LEVEL_DATA_OVERVIEW` |
回答时结`reference/weekly_report_data_explanation.md` 中的字段说明，将原始数据翻译为用户可理解的业务语言并给出分析建议。数据展示使Markdown 表格格式，保持与 Step 4 一致的风格
---
## 异常处理
| 异常场景 | 检测特 | 处理方式 |
| ---------------------- | ------------------------------------------- | ------------------------------------------ |
| 未登录（重定向） | URL 跳转`login.alibaba.com` | 提示"检测到未登录，请完成登，轮询等 |
| 未登录（接口报错 | 接口响应异常或返回错误码 | 引导用户打开登录页，10 秒重 |
| 系统繁忙 | 页面包含"系统繁忙"提示 | 提示用户稍后再试 |
| 周报加载超时 | 轮询超过 30 次（5 分钟 | 停止轮询，提周报获取超时，请稍后重试" |
| 数据为空 / 字段缺失 | 返回空数据或 `encryptedReportId` 缺失 | 输出兜底文案，终止流 |
| `weekDiagnose` 缺失 | 字段不存在或为空 | 注明"未包含经营诊断内，继续输出其他数|
| 诊断标题/摘要均为 | `diagnoseTitle` `diagnoseSummary` 均无| 跳过该诊断项，继续处 |
---
## 操作规范
1. **隐私安全**：仅获取用户明确要求的经营数据，不主动抓取或泄露店铺敏感配置
2. **内容格式**：所有分析结果以 Markdown 格式输出，保留国际站相关链接
3. **请求频率**：避免短时间内高频重复请求同一接口，防止触发安全风
---
## 技能元信息
- **技能名*: alibabacom-store-analysis
- **功能描述**: 阿里国际站（Alibaba.com）经营周报数据分析工具。通过浏览器会话自动获取店铺经营周报数据并进行结构化分析- **适用场景**:
- 查看店铺7 天的经营表现
- 了解店铺在同行中的竞争力
- 获取 AI 生成的经营诊断和优化建议
- 查询流量来源、商品表现、买家分布等详细数据
- 分析直通车效果、星等级情况等专项数
阿里国际站经营分析 SKILL技能 完整说明「英文版」
name: alibaba-store-analysis
description: Alibaba.com (Alibaba International Station) weekly business report data analysis tool. Automatically retrieves store weekly report data via browser sessions and performs structured analysis.
Alibaba International Station Weekly Business Report Analysis SKILL
Overview
Access Alibaba International Station via browser, automatically retrieve and analyze store weekly business report data, and output structured business diagnostics.
Simplified Flow
Step 1: Check Login Status
│ Visit i.alibaba.com, detect if redirected to login page
│ ├─ Not logged in → Guide user to log in, wait for completion
│ └─ Logged in ↓
│
Step 2: Retrieve Business Data
│ Call diagnoseData.json API
│ ↓
Step 3: Data Validation (Silent)
│ Validate data / encryptedReportId / values.receipt
│ ├─ Validation failed → Output fallback message, terminate
│ └─ Validation passed ↓
│
Step 4: Structured Display
│ Output all at once: Data Overview → Diagnostic Summary → Merchant Tasks → Report Link
│ ↓
Step 5: Retrieve Full Report Data
│ Call queryWeekReportAllData.json API
│ Display "Full data analysis complete", wait for user questions before providing on-demand analysis
Environment & Prerequisites
Item
Description
Browser Tool
Use browser tool; use user-specified profile if provided, otherwise default to profile=openclaw
Target Domain
https://i.alibaba.com/
Login Status
User must be logged in; guide to complete login if not
Core Workflow
Step 1: Check Login Status
Visit https://i.alibaba.com/ :
Redirected to https://login.alibaba.com/newlogin/icbuLogin.htm?... → Not logged in, prompt user to complete login and notify when done
No redirect → Logged in, proceed with workflow
Step 2: Retrieve Business Data
Execute in browser console:
(async () => {
const url = 'https://crm.alibaba.com/crmlogin/aisales/dingwukong/diagnoseData.json';
const r = await fetch(url, {
method: 'POST',
credentials: 'include'
});
if (!r.ok) throw new Error(`HTTP error: ${r.status}`);
return await r.json();
})()
Step 3: Data Validation (Silent — do not display to user)
Validate in the following order; terminate with fallback message if any check fails:
API response is empty / null / data is empty object or empty array → Terminate
encryptedReportId field does not exist or is empty → Terminate
values.receipt field does not exist or is empty → Terminate (this field is used as a request parameter in Step 5; if it has a value, record it for later use)
Fallback message:
Sorry, no business data is available for the current account. Analysis cannot be performed.
Error tolerance:
weekDiagnose missing → Note “This period’s data does not include business diagnostics”, continue outputting other data, do not terminate
Both diagnoseTitle and diagnoseSummary are empty → Skip this diagnostic item, continue processing others
Step 4: Structured Data Display
From the values.weekDiagnose array, filter for the data object where scope = "近7天" (Last 7 Days), combine with values.diagnoseSummary, and output in the following four sections.
Section 1: Store Data Overview
Extract core metrics from the indicatorList array and display as a Markdown table.
Field mapping:
Field
Meaning
name
Metric name
value
This week’s data
cycleCRC
Week-over-week change
valueVsAvg
vs. Industry average
Example output:
Here is your store’s data for the last 7 days:
Metric
This Week
WoW Change
vs. Industry Avg
Clicks
6.0
+0.0%
-96.4%
Inquiries
10.0
+0.0%
-65.5%
Premium Product Ratio
0.0%
-
-
When cycleCRC or valueVsAvg is missing, display as -
Section 2: Store Diagnostic Summary
Directly display the content of values.diagnoseSummary field.
Example output:
Your store’s average daily search exposure over the last 7 days is only 100, far below the industry average of 1048.6 — you are in catch-up mode. Notably, this metric declined 41.35% compared to last week and needs urgent attention. Recommended immediate actions: publish 5 potential winning products and build 10 premium/trending products — this can boost overall store exposure by 40%-50% within 30 days!
Section 3: Merchant Tasks
Iterate through the maTaskList array, extracting title and desc (HTML tags must be cleaned).
Example output:
Recommended key tasks:
Task 1: Publish 5 potential winning products
Estimated impact: Completing this task could boost overall store exposure by 40%-50%
Task 2: Build 10 premium/trending products
Estimated impact: Completing this task could boost overall store exposure by 20%-30%
Section 4: Weekly Report Link
Construct the link using encryptedReportId:
https://crm.alibaba.com/crmagent/crm-grow/luyou/report-render.html?id={encryptedReportId}&dateScope=week&isDownload=false
Example output:
View full weekly business report: https://crm.alibaba.com/crmagent/crm-grow/luyou/report-render.html?id={encryptedReportId}&dateScope=week&isDownload=false
Display Requirements
Output all four sections to the user at once after data collection is complete
Do not save content to a markdown file
After returning content, proceed to Step 5
Step 5: Retrieve Full Merchant Report Data
After Step 3 data validation passes, use values.receipt from Step 2 response to call the report detail API for the complete merchant report data.
Execute in browser console:
(async () => {
const receipt = '<values.receipt obtained from Step 2 response>';
const url = 'https://crm.alibaba.com/crmlogin/aisales/dingwukong/queryWeekReportAllData.json';
const r = await fetch(url, {
method: 'POST',
credentials: 'include',
headers: { 'Content-Type': 'application/x-www-form-urlencoded' },
body: `receipt=${receipt}`
});
if (!r.ok) throw new Error(`HTTP error: ${r.status}`);
return await r.json();
})()
Data Validation (Silent — do not display to user)
API response is empty / null / data is empty object → Display “Sorry, unable to retrieve full report data at this time. Please try again later.” and terminate
Partial module data missing → Skip that module, continue processing others, do not terminate
Data Understanding & Storage
The API response is organized by reportPageCode into modules. See reference/weekly_report_data_explanation.md for module codes and field descriptions.
Module overview:
Module Code
Description
STORE_DATA_OVERVIEW
Store data overview (search exposure, clicks, inquiries, conversion rate, etc.)
STORE_INFRASTRUCTURE_SITUATION
Store infrastructure status (product quantity, reply rate, etc.)
STORE_DIAGNOSIS
Store diagnostics (star rating, capability scores and optimization suggestions)
OPPORTUNITY_STAR_LEVEL_DATA_OVERVIEW
Inquiry star rating data overview
TRADE_STAR_LEVEL_DATA_OVERVIEW
Transaction star rating data overview
ACTION_SUGGESTION
Action suggestions (daily inquiry & transaction trends, task list)
PRODUCT_DATA_OVERVIEW
Product data overview (product tiers, category Top 5)
EXPOSURE_TOP10_PRODUCT_DATA
Top 10 products by exposure
CATEGORY_EXPANSION_SUGGESTION
Category expansion suggestions
BUYER_DISTRIBUTION_DATA
Buyer distribution data (by exposure/visits/inquiries)
P4P_PLAN_OPTIMIZ_SUGGESTION
P4P (Pay-for-Performance) campaign optimization suggestions
P4P_SEARCH_WORD_OPTIMIZ_SUGGESTION
P4P search keyword optimization suggestions
FLOW_SOURCE_CHANNEL_ANALYSIS
Traffic source channel analysis
STORE_DETAIL_12_MONTHS
Store performance details for the last 12 months
STORE_COMMUNICATION_CONVERSION_OVERVIEW
Communication conversion data overview
STORE_ACCOUNT_DATA_OVERVIEW
Store sub-account data overview
EMPLOYEE_MANAGEMENT
Employee management
For complete field mappings and definitions, see: reference/weekly_report_data_explanation.md
Post-retrieval Behavior
After retrieving the full report data, only display:
Full data analysis complete. Feel free to ask me any questions about your store data!
Wait for the user to ask specific questions before providing targeted answers based on the complete data. Do not proactively display full data content; keep the data in context for on-demand analysis.
Data Application Examples
User Question
Corresponding Data Module
What are my store traffic sources?
FLOW_SOURCE_CHANNEL_ANALYSIS
How is P4P performing?
P4P_DATA_OVERVIEW_1 / P4P_DATA_OVERVIEW_2 + P4P_PLAN_OPTIMIZ_SUGGESTION
How are my products performing?
PRODUCT_DATA_OVERVIEW + EXPOSURE_TOP10_PRODUCT_DATA
Where are my buyers from?
BUYER_DISTRIBUTION_DATA
What’s my star rating status?
STORE_DIAGNOSIS + OPPORTUNITY_STAR_LEVEL_DATA_OVERVIEW + TRADE_STAR_LEVEL_DATA_OVERVIEW
When answering, combine field descriptions from reference/weekly_report_data_explanation.md to translate raw data into business-friendly language with analysis and recommendations. Use Markdown table format for data display, maintaining a consistent style with Step 4.
Exception Handling
Exception Scenario
Detection Signal
Resolution
Not logged in (redirect)
URL redirects to login.alibaba.com
Prompt “Login not detected, please complete login”, poll and wait
Not logged in (API error)
Abnormal API response or error code
Guide user to open login page, retry every 10 seconds
System busy
Page contains “System busy” message
Prompt user to try again later
Weekly report load timeout
Polling exceeds 30 iterations (5 minutes)
Stop polling, prompt “Weekly report retrieval timed out, please try again later”
Empty data / missing fields
Empty response or missing encryptedReportId
Output fallback message, terminate workflow
weekDiagnose missing
Field does not exist or is empty
Note “Business diagnostics not included”, continue outputting other data
Diagnostic title/summary both empty
Both diagnoseTitle and diagnoseSummary have no value
Skip this diagnostic item, continue processing
Operational Standards
Privacy & Security: Only retrieve business data explicitly requested by the user; do not proactively scrape or disclose sensitive store configurations
Content Format: All analysis results output in Markdown format; preserve Alibaba International Station related links
Request Frequency: Avoid high-frequency repeated requests to the same API in a short period to prevent triggering security controls