# Amazon Sorftime MCP Skills: Nine Agent Skills for Amazon Seller Research > A Claude Code skill collection that connects AI agents to Sorftime, Sif, Xiyou Insight, and SellerSprite via Model Context Protocol, automating nine research workflows for Amazon cross-border sellers. The skills handle single-listing deep analysis, full-category selection, keyword mining, review analysis, and Listing construction, each calling the relevant paid APIs and rendering structured reports. **liangdabiao/amazon-sorftime-research-MCP-skill** — 亚马逊选品 之 Listing全维度穿透分析报告 加上 全品类分析 ，关键词分析，差评分析 ，市场调研 等等。codex/claude code agent skill, amazon sorftime MCP/西柚mcp/sif mcp/卖家精灵sellersprite 智能体skill. 亚马逊跨境电商skill工具集。 - Repository: https://github.com/liangdabiao/amazon-sorftime-research-MCP-skill - Website: https://skillhub.cn/skills/user_8bb4b4f5/amazon-sorftime-research-mcp-skill - Stars: 918 · Forks: 133 - Language: Python - License: not declared - Published: 2026-09-10 · Updated: 2026-09-10 · Language: en - Canonical page: https://hysenlabs.com/projects/liangdabiao-amazon-sorftime-research-mcp-skill ## Who Uses This and What Problem It Solves Amazon cross-border sellers spend hours each day moving data between commercial research tools and their own analysis context. A seller who wants to evaluate a category runs searches in Sorftime, downloads keyword data into a spreadsheet, copies negative reviews from the product page, and then tries to draw conclusions across three or four different data sources. Amazon Sorftime MCP Skills eliminates the manual handoff. By wiring Sorftime, Sif, Xiyou Insight, and SellerSprite into Model Context Protocol servers and packaging each research workflow as a named skill, the project lets a seller type a single agent command and receive a structured report without leaving the Claude Code environment. The nine skills cover the main research tasks a seller faces before entering a new category: evaluating a competitor's listing, surveying an entire product category, mining keywords, reading reviews for pain points, researching a product from scratch, and building a new Listing from the ground up. ## Nine Skills and the Data Services Behind Each One Each skill connects to one or more paid API services. The amazon-analyse skill calls Sorftime to fetch product details, keyword traffic data, review sentiment, and cross-platform signals including TikTok video data and sourcing cost from 1688. The category-selection skill pulls Top 100 products for a category and scores the market across five dimensions: market size, growth potential, competition intensity, entry barriers, and profit margin. The keyword-research skill collects up to 1,500 keywords through the Sorftime API and classifies them across eight dimensions including negative terms, brand terms, scene terms, and core terms. The review-analysis skill performs six-dimension pain point analysis covering electronic module failures, structural issues, design defects, appearance problems, description mismatch, and service issues. The sif-amazon-research skill uses the Sif MCP service for market validation, competitor path analysis, keyword structure diagnosis, and launch evaluation with a confidence score and evidence chain. The xiyou-insight skill covers seven scenarios including hourly ad monitoring, traffic gap analysis, and competitive ad strategy via 17 Xiyou API endpoints. The sellersprite-amazon-research skill gives access to 43 SellerSprite API tools across 10 workflow commands. The amazon-listing-builder skill runs an eight-step process from keyword layering to Alexa QA content, built on SellerSprite data. ## Report Formats and Output Structure Each skill writes its output to a directory under the current working path. The category-selection skill produces three formats: a Markdown report, an Excel file, and an HTML dashboard. The xiyou-insight skill writes to a directory named xiyou-insight-reports followed by the scenario, ASIN, site, and date, containing a Markdown report, a structured JSON data file, an ECharts HTML visualization dashboard, and a raw MCP response directory. The sellersprite-amazon-research skill produces a Markdown report, an HTML visualization, and a JSON file containing the raw API data. All skills use a consistent output pattern: Markdown for human-readable findings, JSON for downstream processing, and HTML for visual inspection. This structure means a seller can share the Markdown report in a team channel, feed the JSON into a spreadsheet, and open the HTML dashboard in a browser without any additional processing. ## MCP Server Configuration and the .mcp.json File The repository includes a .mcp.json configuration file at the top level. This file wires the Sorftime, Sif, Xiyou, and SellerSprite APIs into the Claude Code MCP layer so the agent can call them as tools. Each MCP server entry specifies the server command, arguments, and environment variables that hold the API keys. The skills themselves reference these MCP tools by name; the skills will not function without valid API credentials for the services they call. The repository also includes a CLAUDE.md file that provides the agent with context about the skill collection and how to route requests to the correct skill. This architecture means installation involves two steps: adding the MCP server credentials to the environment and placing the skills in the Claude Code skills directory. ## What the Skills Cannot Do: Limitations and Dependencies The skills are orchestration layers over paid third-party APIs. Every analysis workflow depends entirely on the availability and accuracy of the upstream data service. If a Sorftime API call returns stale data or an error, the skill produces an incomplete report without any graceful degradation mechanism documented in the README. The skills are written in Chinese and designed for a Chinese cross-border seller workflow; the commands, parameter names, and report section headings are in Simplified Chinese. An English-speaking seller can run the commands but will receive Chinese-language reports. The amazon-listing-builder skill requires Cosmo semantic algorithm access via SellerSprite; the README does not document which SellerSprite subscription tier includes this. The repository has no GitHub releases and no version tags, so there is no stable reference point to pin for a production environment. ## Compared to Running Platform APIs Directly and Licence Status The main alternative to this skill collection is querying each API directly through its own web interface or a custom script. A seller using Sorftime's web UI can run most of the same analyses manually, but each query produces results in a separate session with no cross-platform synthesis. The value of the skill collection is the aggregation: a single agent command can pull data from Sorftime, classify it with LLM reasoning, cross-reference review sentiment from SellerSprite, and produce a combined report. The trade-off is complexity: the seller must configure four separate API credentials, maintain the MCP server connections, and keep the skills updated when any upstream API changes its schema. The licence field in the repository is listed as unknown. No open-source licence is stated in the README or repository root. The last push was on 2026-07-08. ## Conclusion Amazon Sorftime MCP Skills is the right tool for cross-border sellers who already pay for Sorftime, SellerSprite, Sif, or Xiyou Insight subscriptions and want to automate repetitive research steps inside Claude Code. The skills remove the need to copy and paste between data platforms and the agent window; the agent calls the APIs directly and formats the results into Markdown, CSV, and HTML reports. Teams that do not hold paid API subscriptions to these services will find the skills non-functional. The licence is not stated in the repository. Before deploying in a commercial setting, verify the licensing terms with the author. The last push was on 2026-07-08. ## FAQ ### What API subscriptions are needed to use Amazon Sorftime MCP Skills? Each skill calls one or more paid services. The core skills use Sorftime. The sif-amazon-research skill requires a Sif MCP subscription. The xiyou-insight skill requires Xiyou Insight API access. The sellersprite-amazon-research and amazon-listing-builder skills require a SellerSprite subscription. ### What output formats do the Amazon Sorftime MCP Skills produce? Most skills produce three formats: a Markdown report for reading, a JSON file containing the raw API data, and an HTML dashboard with ECharts visualizations. The category-selection skill also produces an Excel file. ### Can Amazon Sorftime MCP Skills be used with agents other than Claude Code? The README describes the skills as designed for Codex and Claude Code agent skill formats. The repository includes a CLAUDE.md file and a .mcp.json configuration, both of which are Claude Code conventions. The README does not document compatibility with other agents. ## Sources - [Issues](https://github.com/liangdabiao/amazon-sorftime-research-MCP-skill/issues) - [liangdabiao/amazon-sorftime-research-MCP-skill on GitHub](https://github.com/liangdabiao/amazon-sorftime-research-MCP-skill) - [Project website](https://skillhub.cn/skills/user_8bb4b4f5/amazon-sorftime-research-mcp-skill) - [README](https://github.com/liangdabiao/amazon-sorftime-research-MCP-skill/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/liangdabiao-amazon-sorftime-research-mcp-skill