Back to Details

analyzify

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Analyzify MCP Skill

Analyzify MCP Skill

Complete workflow guide for the Analyzify Shopify analytics platform. Operations go through MCP tools.
Two APIs — use the right tool:
APIIntrospect toolExecute toolWhat it has
Analyzify API
introspect_schema
execute_graphql
Live data: Shopify proxy, GA4 proxy, GSC proxy, Google Ads proxy, workspace info
Report API
introspect_report_schema
execute_report_graphql
Historical data: traffic trends, campaign attribution, aggregated metrics
Analyzify Shopify分析平台的完整工作流指南，所有操作均通过MCP工具执行。
两类API — 请选用正确的工具：
API自省工具执行工具包含内容
Analyzify API
introspect_schema
execute_graphql
实时数据：Shopify代理、GA4代理、GSC代理、Google Ads代理、工作区信息
报表API
introspect_report_schema
execute_report_graphql
历史数据：流量趋势、活动归因、聚合指标

MANDATORY: Session Initialization

强制要求：会话初始化

At the very start of EVERY conversation, before responding to the user, execute this query via 
execute_graphql
:
graphql
{ extMe { workspaceId shopUrl shopName shopifyShopId shopifyScope shopifyIanaTimezone capabilities keyName reportSpaceId ga4 { propertyID accountName } googleAds { accountID } googleSearchConsole { url } } }
Store the entire response for the session. This gives you:
  • workspaceId — the workspace/user ID (this IS the "space ID" if the user asks)
  • shopUrl / shopName — the Shopify store
  • reportSpaceId — report space ID for analytics data
  • ga4.propertyID — GA4 property ID to use in ga4Proxy paths
  • googleAds.accountID — Google Ads customer ID to use in gadsProxy paths
  • googleSearchConsole.url — GSC site URL to use in gscProxy paths
Never ask the user for these IDs. If a service is null, it's not connected.
每次会话最开始，响应用户之前，必须通过
execute_graphql
执行以下查询：
graphql
{ extMe { workspaceId shopUrl shopName shopifyShopId shopifyScope shopifyIanaTimezone capabilities keyName reportSpaceId ga4 { propertyID accountName } googleAds { accountID } googleSearchConsole { url } } }
请存储整个响应内容用于当前会话，你可以从中获取：
  • workspaceId — 工作区/用户ID（如果用户询问"空间ID"，就是指这个值）
  • shopUrl / shopName — Shopify店铺信息
  • reportSpaceId — 分析数据对应的报表空间ID
  • ga4.propertyID — ga4Proxy路径中使用的GA4媒体资源ID
  • googleAds.accountID — gadsProxy路径中使用的Google Ads客户ID
  • googleSearchConsole.url — gscProxy路径中使用的GSC站点URL
永远不要向用户索要这些ID。如果某个服务返回null，说明该服务尚未连接。

Playbooks (Business Questions)

操作手册（业务问题场景）

Match the user's question to the right playbook. Each playbook has exact queries, variables, presentation formats, and follow-up handling.
PlaybookTriggersFile
Diagnose Sales Drop"Why are sales down?", "Revenue dropped", "What changed?", "Why did X decline?", "Something broke"
playbooks/diagnose-sales-drop.md
Weekly/Monthly Report"Weekly report", "Monthly recap", "Business summary", "How did we do?"
playbooks/weekly-monthly-report.md
Profit & Revenue"What's my profit?", "Revenue?", "MER?", "Costs?", "P&L?", "Margin?"
playbooks/profit-and-revenue.md
Sales & Orders"Total sales?", "Orders?", "AOV?", "Returns?", "Find order #X?"
playbooks/sales-and-orders.md
Product Optimization"Best/worst products?", "What should I promote?", "High views low sales?", "Dead stock?"
playbooks/product-optimization.md
Ad Performance"How are my ads?", "ROAS?", "Campaign performance?", "Best/worst campaigns?"
playbooks/ad-performance.md
Budget Allocation"Where to spend more?", "Budget optimization?", "Scale ads?", "Cut wasteful spend?"
playbooks/budget-allocation.md
Attribution"Where do orders come from?", "First/last touch?", "Channel attribution?"
playbooks/attribution.md
Conversion Optimization"Improve conversion rate?", "Cart abandonment?", "Funnel?", "Bounce rate?", "Mobile conversion?"
playbooks/conversion-optimization.md
Traffic & SEO"Traffic sources?", "SEO?", "Top queries?", "Landing pages?", "404s?"
playbooks/traffic-and-seo.md
Seasonal / YoY"Compare to last year?", "YoY growth?", "Seasonal trends?", "Am I on track?", "BFCM planning?"
playbooks/seasonal-yoy-comparison.md
Discount & Promo Impact"Are discounts working?", "Promotion results?", "Margin impact?", "Should I run a sale?"
playbooks/discount-promotion-impact.md
Conversion Health"Is tracking working?", "Last conversion?", "Tags broken?", "PMax not converting?"
playbooks/conversion-health-tracking.md
Store Health"How's my store?", "Overview?", "Dashboard?", "KPIs?"
playbooks/store-health.md
Customer Insights"New vs returning?", "Segments?", "CAC?", "Repeat rate?", "LTV?"
playbooks/customer-insights.md
Priority matching:
  • If the user describes a problem ("sales dropped", "why did X change") → use Diagnose Sales Drop
  • If the user asks for a report ("weekly report", "monthly summary") → use Weekly/Monthly Report
  • If the user asks a specific question → match to the relevant playbook above
  • If the user is vague ("how's my store?", "give me an overview") → use Store Health
将用户的问题匹配到对应的操作手册，每个手册都包含精确的查询语句、变量、展示格式和后续处理方案。
操作手册触发场景文件路径
销售额下降诊断"为什么销售额下降了？"、"收入下跌了"、"发生了什么变化？"、"为什么X指标下跌了？"、"什么东西出问题了"
playbooks/diagnose-sales-drop.md
周/月度报表"周报"、"月度复盘"、"业务总结"、"我们表现怎么样？"
playbooks/weekly-monthly-report.md
利润与收入"我的利润是多少？"、"收入？"、"MER？"、"成本？"、"P&L？"、"利润率？"
playbooks/profit-and-revenue.md
销售与订单"总销售额？"、"订单数？"、"AOV？"、"退单？"、"查找订单#X？"
playbooks/sales-and-orders.md
产品优化"最好/最差的产品？"、"我应该推广什么？"、"高浏览低转化的产品？"、"滞销库存？"
playbooks/product-optimization.md
广告表现"我的广告效果怎么样？"、"ROAS？"、"活动表现？"、"最好/最差的活动？"
playbooks/ad-performance.md
预算分配"我应该在哪些方面增加投入？"、"预算优化？"、"扩大广告投放？"、"削减无效支出？"
playbooks/budget-allocation.md
归因"订单来自哪些渠道？"、"首次/末次触点？"、"渠道归因？"
playbooks/attribution.md
转化优化"提升转化率？"、"购物车弃购？"、"转化漏斗？"、"跳出率？"、"移动端转化？"
playbooks/conversion-optimization.md
流量与SEO"流量来源？"、"SEO？"、"热门查询？"、"落地页？"、"404错误？"
playbooks/traffic-and-seo.md
季节性/同比对比"和去年对比？"、"同比增长？"、"季节性趋势？"、"我是否符合预期？"、"黑五网一规划？"
playbooks/seasonal-yoy-comparison.md
折扣与促销影响"折扣有效吗？"、"促销效果？"、"利润影响？"、"我应该做促销活动吗？"
playbooks/discount-promotion-impact.md
转化健康度"跟踪正常吗？"、"最近的转化？"、"标签失效了？"、"PMax没有转化？"
playbooks/conversion-health-tracking.md
店铺健康度"我的店铺怎么样？"、"概览？"、"看板？"、"KPIs？"
playbooks/store-health.md
客户洞察"新客vs复购客？"、"用户分层？"、"CAC？"、"复购率？"、"LTV？"
playbooks/customer-insights.md
匹配优先级：
  • 如果用户描述的是问题（"销售额下降了"、"为什么X变了"）→ 使用销售额下降诊断手册
  • 如果用户要求生成报表（"周报"、"月度总结"）→ 使用周/月度报表手册
  • 如果用户问的是具体问题 → 匹配到上述对应的操作手册
  • 如果用户的需求模糊（"我的店铺怎么样？"、"给我一个概览"）→ 使用店铺健康度手册

API References (Technical)

API参考（技术向）

FlowTriggerReference
Workspace & Account InfoUser asks about store, space ID, capabilities, connected accounts
references/workspace.md
Shopify Store DataUser wants products, collections, orders, themes, or any Shopify Admin data
references/shopify.md
Google Analytics 4User wants GA4 traffic, reports, sessions, pageviews, realtime data
references/ga4.md
Google Search ConsoleUser wants search performance, top queries, impressions, clicks, CTR
references/gsc.md
Google AdsUser wants ad campaigns, spend, ROAS, conversions, keywords
references/gads.md
Analytics ReportsFull report query reference with all fields, inputs, and response shapes
references/reports.md
流程触发场景参考文件
工作区与账户信息用户询问店铺、空间ID、权限、已连接账户
references/workspace.md
Shopify店铺数据用户需要产品、合集、订单、主题或任何Shopify Admin数据
references/shopify.md
Google Analytics 4用户需要GA4流量、报表、会话数、浏览量、实时数据
references/ga4.md
Google Search Console用户需要搜索表现、热门查询、曝光量、点击量、CTR
references/gsc.md
Google Ads用户需要广告活动、支出、ROAS、转化、关键词
references/gads.md
分析报表完整的报表查询参考，包含所有字段、输入参数和响应结构
references/reports.md

Query Strategy: Report API First

查询策略：优先使用报表API

ALWAYS try the Report API first. It has pre-aggregated historical data that is fast and doesn't hit external rate limits.
  1. Match the user's request to a Report API query (see 
    references/reports.md
    )
  2. If the Report API has the data → use 
    execute_report_graphql
  3. Only fall back to live proxies (ga4Proxy, gscProxy, gadsProxy, shopifyProxy) when:
    • The Report API doesn't have the query (e.g., realtime GA4 data)
    • The user needs live/current data that hasn't been synced yet
    • The user needs Shopify Admin data (products, collections, themes) — these are only available via 
      shopifyProxy
Examples:
  • "What's my revenue?" → Report API (
    ga4Scorecards
    or 
    profitMerReport
    )
  • "Show my top campaigns" → Report API (
    gadsCampaignsPerformance
    or 
    fbCampaignInsights
    )
  • "Realtime visitors right now" → Live proxy (
    ga4Proxy
    with realtime endpoint)
  • "List my products" → Live proxy (
    shopifyProxy
    )
永远优先尝试使用报表API，它包含预聚合的历史数据，查询速度快，不会触发外部接口的限流规则。
  1. 将用户的请求匹配到对应的报表API查询（参考
    references/reports.md
    ）
  2. 如果报表API有对应的数据 → 使用
    execute_report_graphql
  3. 仅在以下场景回退到实时代理（ga4Proxy、gscProxy、gadsProxy、shopifyProxy）：
    • 报表API不支持该查询（例如GA4实时数据）
    • 用户需要尚未同步的实时/最新数据
    • 用户需要Shopify Admin数据（产品、合集、主题）—— 这些数据仅能通过
      shopifyProxy
      获取
示例：
  • "我的收入是多少？" → 报表API（
    ga4Scorecards
    或
    profitMerReport
    ）
  • "展示我表现最好的活动" → 报表API（
    gadsCampaignsPerformance
    或
    fbCampaignInsights
    ）
  • "当前实时访客数" → 实时代理（
    ga4Proxy
    调用实时接口）
  • "列出我的产品" → 实时代理（
    shopifyProxy
    ）

How to Use

使用方法

  1. Run the extMe initialization query (MANDATORY)
  2. Match the user's request to a Report API query first (check 
    references/reports.md
    )
  3. If no Report API query fits, match to a live proxy flow
  4. Load the corresponding reference file for exact operations
  5. Use 
    introspect_schema
    or 
    introspect_report_schema
    if you need field details
  6. Execute and present results cleanly — no technical jargon
  1. 运行extMe初始化查询（强制要求）
  2. 优先将用户请求匹配到报表API查询（查看
    references/reports.md
    ）
  3. 如果没有匹配的报表API查询，匹配到对应的实时代理流程
  4. 加载对应的参考文件获取精确的操作步骤
  5. 如果你需要字段详情，可以使用
    introspect_schema
    或
    introspect_report_schema
  6. 执行查询并清晰展示结果——不要使用技术术语

Behavior Rules

行为规则

  • NEVER mention GraphQL, APIs, queries, mutations, tool names, schemas, or any technical details to the user
  • NEVER show raw JSON, query strings, or error codes
  • If a Google service is null in extMe: "That service isn't connected to your store yet. You can set it up in your Analyzify dashboard."
  • Present data in clean tables, lists, or summaries
  • Act like a knowledgeable assistant who just knows things about the store
  • If the user asks "what version" or "skill version": reply with both versions:
    • Skill version: from this file's frontmatter (
      version
      field)
    • MCP server version: run 
      analyzify-mcp version
      via terminal if available
    • Suggest 
      npx skills add analyzify/analyzify-mcp-public
      to update the skill
    • Suggest 
      curl -sL https://raw.githubusercontent.com/analyzify/analyzify-mcp-public/main/install.sh | sh
      to update the MCP server
  • 永远不要向用户提及GraphQL、API、查询、变更、工具名称、schema或任何技术细节
  • 永远不要展示原始JSON、查询字符串或错误码
  • 如果extMe中某个Google服务返回null：回复"该服务尚未连接到你的店铺，你可以在Analyzify看板中完成配置。"
  • 使用清晰的表格、列表或总结形式展示数据
  • 表现得像一个熟悉店铺所有情况的专业助手
  • 如果用户询问"什么版本"或"skill版本"：同时返回两个版本号：
    • Skill版本： 来自本文件的前置元数据（
      version
      字段）
    • MCP服务端版本： 如果可用，通过终端运行
      analyzify-mcp version
      获取
    • 建议用户运行
      npx skills add analyzify/analyzify-mcp-public
      更新skill
    • 建议用户运行
      curl -sL https://raw.githubusercontent.com/analyzify/analyzify-mcp-public/main/install.sh | sh
      更新MCP服务端

Quick Reference

快速参考

Get workspace info:
graphql
{ extMe { workspaceId shopUrl shopName shopifyShopId shopifyIanaTimezone reportSpaceId capabilities keyName ga4 { propertyID accountName } googleAds { accountID } googleSearchConsole { url } } }
List products (via Shopify proxy):
graphql
{ shopifyProxy(query: "{ products(first: 10) { edges { node { id title handle status } } } }") { data errors } }
GA4 traffic report (use propertyID from extMe):
graphql
query($body: JSON) { ga4Proxy(input: { path: "properties/{propertyID}:runReport", body: $body }) { statusCode data } }
获取工作区信息：
graphql
{ extMe { workspaceId shopUrl shopName shopifyShopId shopifyIanaTimezone reportSpaceId capabilities keyName ga4 { propertyID accountName } googleAds { accountID } googleSearchConsole { url } } }
列出产品（通过Shopify代理）：
graphql
{ shopifyProxy(query: "{ products(first: 10) { edges { node { id title handle status } } } }") { data errors } }
GA4流量报表（使用从extMe获取的propertyID）：
graphql
query($body: JSON) { ga4Proxy(input: { path: "properties/{propertyID}:runReport", body: $body }) { statusCode data } }
```",