# K-Data Gate MCP Benchmark Pack Generated: 2026-06-13T10:29:40.686Z Status: mcp_benchmark_ready_not_submitted Positioning: First-mover Korea sourcing MCP for AI agents and ecommerce sellers, not a generic ecommerce scraper. ## Executive Answer - Overseas built: Overseas seller MCPs cluster around Amazon/Shopify account analytics, ecommerce web-data extraction, price monitoring, lead generation, and agentic commerce checkout context. - K-Data wedge: K-Data Gate should own the upstream Korea-local sourcing layer: what Korean product to test, what local trend/source signals say, what ingredient/source warnings exist, and what proof pack a seller can buy before inventory risk. - Immediate benchmark: Do not copy scraper language. Copy the packaging: task-named tools, clear schemas, JSON/CSV/Markdown exports, public proof URLs, pricing boundaries, install notes, and honest source limits. ## Overseas Patterns ### Amazon seller account MCP Examples: DataDoe Amazon Seller MCP, Amazon Data Kiosk MCP sample Buyer job: Let a seller or agency ask about orders, inventory, fees, ads, vendor POs, chargebacks, and profit without building SP-API glue. Tool shape: get_orders, get_fba_inventory, brand_analytics, profit_by_sku, tacos_by_campaign Trust mechanic: Official OAuth/API access, scoped private endpoint, token rotation, rate-limit handling, schema auto-discovery. Monetization: Subscription or private account connector. Value is post-sale operations and true-profit analytics. K-Data move: Show K-Data Gate as the pre-inventory Korea sourcing layer, then integrate with Amazon/Shopify tools later through exported CSV/JSON and stable category IDs. Avoid: Do not pretend to access Amazon Seller Central or customer private store data. ### Shopify storefront/admin MCP Examples: Shopify Storefront MCP, Shopify Admin MCP servers on Glama/Smithery Buyer job: Expose product catalogs, carts, customers, orders, discounts, collections, and store metadata to AI agents. Tool shape: search_products, get_cart, query_orders, manage_discounts, list_collections Trust mechanic: GraphQL API, store-specific permissions, install flow, and clean tool schema. Monetization: Agency app, store AI assistant, analytics connector, or hosted MCP deployment. K-Data move: Make K-Data Gate outputs store-ingestible: title, category, local price, USD estimate, trend score, ingredient/source warning, and proof URL. Avoid: Do not crawl Shopify stores as a substitute for permissioned store APIs. ### Apify ecommerce intelligence MCP Examples: Shopify store analysis Actors, Amazon seller scrapers, price trackers, Google Maps lead finders Buyer job: Run web-data actors for competitive analysis, product extraction, price monitoring, seller leads, and lead quality scoring. Tool shape: shopify_analyze, shopify_products, amazon_search, amazon_product, maps_leads, price_monitor Trust mechanic: Ready MCP config, pay-per-result or pay-per-run, actor input schema, JSON datasets, API token. Monetization: Usage-based data extraction marketplace; many small tools with clear pricing and sample outputs. K-Data move: Keep our derived tools narrow and paid-scope friendly: Korea Seller Signal, Compliance Scout, Buyer Discovery, Delivery Pack, Benchmark Pack. Avoid: Do not use anti-blocking or proxy language as a selling point for restricted Korean retail sources. ### Agentic commerce protocols Examples: Google UCP, Stripe agentic commerce guide, MCP+x402 payment guides Buyer job: Give agents structured product, inventory, price, checkout, and approval context without browser scraping. Tool shape: discover_product, compare_options, quote_scope, request_approval, record_order_intent Trust mechanic: Machine-readable inventory/pricing, clear human approval loop, auditable payment or HTTP 402 flow. Monetization: Agent-readable storefronts first, automated checkout later. K-Data move: Paddle can wait. The immediate win is agent-readable discovery, fixed-scope offers, order intent, invoice draft, and delivery acceptance criteria. Avoid: Do not add autonomous payment or x402 before demand and support workflow are proven. ### MCP registry distribution Examples: MCP.Directory, MCP.so, Glama, Smithery Buyer job: Help agents and builders discover installable MCP servers with clear README, tools, tags, source, and connection instructions. Tool shape: tools/list, README tool table, OpenAPI link, llms.txt link, sample call Trust mechanic: Repository metadata, install notes, public HTTPS endpoint, schema consistency, working examples. Monetization: Discovery surface, not the checkout itself. Conversion happens through pilot, listing, or invoice path. K-Data move: Position as the first Korea sourcing and MFDS source-map MCP, not as a generic ecommerce scraper. Avoid: Do not submit vague claims like "all Korean data"; use exact tools and honest live/blocked status. ## Gap Map ### korea_sourcing_gap Market gap: Overseas tools are strong on Amazon/Shopify account analytics and generic web extraction, but weak on Korea-local product/trend/source context in English. K-Data move: Lead with "Korea sourcing intelligence for AI agents and sellers": K-beauty, K-food, K-pop merch, local Korean demand, source warnings, and MFDS ingredient source mapping. Proof: https://kdata-gate.vercel.app/seller-signals?category=korean_sunscreen&target_market=Amazon+US&utm_source=korea_sourcing_gap&utm_campaign=mcp_benchmark ### pre_inventory_gap Market gap: Amazon MCPs answer what happened after products are already listed. Sellers also need what to source before buying inventory. K-Data move: Sell Signal Sprint and Growth Feed as pre-inventory decision packets with CSV/JSON outputs the seller can paste into Amazon/Shopify planning tools. Proof: https://kdata-gate.vercel.app/briefs/sourcing?category=korean_sunscreen&utm_source=pre_inventory_gap&utm_campaign=mcp_benchmark ### compliance_source_gap Market gap: Generic ecommerce MCPs rarely package official Korean source rows for importer review. K-Data move: Use Compliance Scout as the premium wedge: official-source MFDS mapping, blocked claims, risk questions, and legal-review handoff notes. Proof: https://kdata-gate.vercel.app/compliance-scout?category=korean_sunscreen&target_market=Amazon+US&utm_source=compliance_source_gap&utm_campaign=mcp_benchmark ### agent_discovery_gap Market gap: Registries reward clear tool contracts, examples, and install text more than broad dashboards. K-Data move: Expose one benchmark page, one submission desk, one llms.txt, and tool names that map to seller jobs. Proof: https://kdata-gate.vercel.app/submission-desk?utm_source=agent_discovery_gap&utm_campaign=mcp_benchmark ## Implementation Moves - 1. [ready_to_apply] Rename the public story from generic Korean data to Korea sourcing MCP for agents and ecommerce sellers. Proof: Use website title, MCP description, llms.txt, and submission copy to repeat the same buyer job. - 2. [implemented] Keep derived MCP tools as work packets, not raw tables only. Proof: Signal pack, Compliance Scout, Buyer Discovery, Invoice Draft, Delivery Pack, Freelance Listings, Operator Playbook, and this Benchmark Pack are agent-readable closeout packets. - 3. [next_patch] Add Shopify/Amazon handoff fields to seller-facing outputs over time. Proof: Use stable category, local price, USD estimate, trend score, compliance/source warning, proof URL, and next test query. - 4. [implemented] Treat Apify-style pay-per-result pricing as a benchmark, but keep first sale as fixed-scope B2B. Proof: $500/$1,500/$3,000/$5,000 scopes are easier to sell before self-serve checkout. - 5. [guardrail] Do not compete as a scraper. Compete as licensed/official Korea context plus buyer-ready exports. Proof: Olive Young/private endpoint bypass, marketplace login scraping, and restricted-source resale stay blocked. ## Listing Copy Korea sourcing intelligence MCP for AI agents and ecommerce sellers: Korean product signals, K-beauty MFDS ingredient source maps, trends, buyer discovery, and fixed-scope sales packets. - 14 Korean dataset tools plus seller-ready derived tools - K-beauty/K-food sourcing signal packs with CSV/JSON/Markdown exports - MFDS ingredient source-map samples for importer review, not legal advice - Buyer discovery and delivery/invoice packets are approval-gated - Honest live/fixture source status and no private endpoint bypass Agent first question: What Korean product category should this seller validate before buying inventory, and which K-Data Gate proof asset should they open first? ## Proof Links - sellerSignals: https://kdata-gate.vercel.app/seller-signals?category=korean_sunscreen&target_market=Amazon+US&utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - sellerSignalsApi: https://kdata-gate.vercel.app/api/seller/signals?category=korean_sunscreen&target_market=Amazon+US&format=md&utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - complianceScout: https://kdata-gate.vercel.app/compliance-scout?category=korean_sunscreen&target_market=Amazon+US&utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - buyerDiscovery: https://kdata-gate.vercel.app/buyer-discovery?target_market=Amazon+US&utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - freelanceListings: https://kdata-gate.vercel.app/freelance-listings?offer=signal_sprint&category=korean_sunscreen&target_market=Amazon+US&utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - operatorPlaybook: https://kdata-gate.vercel.app/operator-playbook?utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - submissionDesk: https://kdata-gate.vercel.app/submission-desk?utm_source=mcp_benchmark&utm_campaign=mcp_benchmark - llms: https://kdata-gate.vercel.app/llms.txt - openapi: https://kdata-gate.vercel.app/openapi.json ## Guardrails - No claim that K-Data Gate can access Amazon Seller Central, Shopify admin, Olive Young private endpoints, or private marketplace dashboards. - No restricted-source scraping, login scraping, cookie reuse, proxy bypass, fake review, rank manipulation, or compliance-clearance claim. - No external MCP hub submission, public post, DM, email, invoice, payment collection, account change, or Paddle activation without action-time approval. - No fixture-backed or blocked source is described as live data. - Benchmark examples are used for packaging strategy, not copied code, copied branding, or implied partnership. ## Sources - DataDoe Amazon Seller MCP: https://www.datadoe.com/connect/amazon/mcp (Benchmarked for official OAuth, private endpoint, self-describing schema, 40+ Amazon tools, true-profit framing, and scheduled/webhook workflows.) - Amazon Data Kiosk MCP discussion: https://github.com/amzn/selling-partner-api-samples/discussions/131 (Benchmarked for domain-specific seller/vendor analytics tools and query-builder pattern against an official Amazon API.) - Shopify Storefront MCP docs: https://shopify.dev/docs/apps/build/storefront-mcp (Benchmarked for product/cart/customer structured commerce data through a standard MCP client-server contract.) - Apify MCP server documentation: https://docs.apify.com/platform/integrations/mcp (Benchmarked for actor discovery, ready MCP configuration, result storage, and usage-based data extraction workflows.) - E-Commerce Intelligence MCP Server: https://mcpservers.org/servers/apify-com-nexgendata-ecommerce-intelligence-mcp-server (Benchmarked for Shopify store analysis, product extraction, competitive analysis, pricing disclosure, and low-friction MCP marketplace packaging.) - MCP.Directory submit: https://mcp.directory/submit (Benchmarked for listing requirements: repository, description, language/license/tools/README auto-pull, and review flow.) - Smithery publish docs: https://smithery.ai/docs/build/publish (Benchmarked for public HTTPS server distribution, analytics, configuration UI, and bot/WAF allowance requirements.) - Google agentic commerce/UCP announcement: https://blog.google/products/ads-commerce/agentic-commerce-ai-tools-protocol-retailers-platforms/ (Benchmarked for agentic discovery, shopping journey standardization, and merchant discoverability in conversational commerce.) - Stripe agentic commerce guide: https://stripe.com/guides/agentic-commerce (Benchmarked for agent preference for structured data, fast responses, inventory/pricing context, and human approval before purchase.) ## Boundary Benchmark and positioning packet only. It does not submit to MCP hubs, scrape private sources, change accounts, invoice, collect payment, or activate Paddle.