DEV Community: curatedmcp

MCP Ecosystem Week 37: The Governance Challenge of Nested AI and Cross-IDE Control

curatedmcp — Mon, 07 Sep 2026 10:39:38 +0000

MCP Ecosystem Week 37: The Governance Challenge of Nested AI and Cross-IDE Control

As AI coding tools proliferate across your engineering org, a critical governance gap is emerging: developers are spinning up MCP servers faster than platform teams can audit them. Week 37 reveals which integrations your team is already using—and what security decisions you need to make about them.

This Week in MCP

No new servers entered the CuratedMCP catalog this week, but that's not a slowdown—it's a sign of maturity. The existing 72 risk-classified servers are seeing accelerating adoption. This is the moment to ensure your allowlist policies are enforced consistently across Claude Code, Cursor, Windsurf, and GitHub Copilot. If you haven't conducted a live audit of which servers are running on developer machines right now, Week 37 is your reminder: shadow MCP usage is real, and it's growing.

On the Radar

The five most-viewed servers this week tell you exactly what your developers want to connect to their AI agents:

GitHub Copilot MCP (98K views) and GitHub MCP (76K views) dominate. Both grant AI agents deep access to your source code, pull requests, and workflows. Governance consideration: Does your team require mutual TLS or IP allowlisting for these integrations? Are audit logs being retained?

OpenAI MCP (87K views) lets developers route requests through OpenAI's APIs from within Claude or other clients. Supply-chain risk: You're now dependent on two LLM vendors' uptime and data handling policies. Token spend visibility becomes essential when requests can flow through multiple providers.

Figma MCP (82K views) exposes design tokens and component libraries to AI agents. AppSec angle: Who owns the Figma workspace credentials being stored locally? Is there a credential rotation policy?

Anthropic Claude MCP (76K views) allows nesting Claude as a sub-agent. This is architectural—it adds latency, multiplies token consumption, and creates audit visibility challenges across multiple inference calls.

Governance Take

Here's what keeps platform leaders awake: you cannot govern what you cannot see. This week's view counts reflect organic developer demand, but they likely undercount actual deployments. Developers are running MCP servers across four separate IDE integrations (Claude Code, Cursor, Windsurf, GitHub Copilot), each with its own local configuration. Your RBAC and allowlist are only as strong as your ability to enforce them per-machine.

Start here: audit which MCP servers are actually running in your org right now. Use CuratedMCP's per-machine enforcement to lock down policy—don't rely on developer best practices. Then, layer in visibility: when developers connect OpenAI MCP or Figma MCP, you should know about it the same way you monitor SaaS app usage. If you're seeing multiple Claude integrations and nested Claude calls, you're also seeing token spend sprawl. TokenShield gives you the visibility to measure that spend and understand which optimization patterns actually work in your environment—before you commit to cost-saving approaches that might degrade your team's velocity.

The policy library at 72 risk-classified servers means you have a foundation. Now enforce it.

Govern MCP usage across your team with CuratedMCP — or scan your own stack free at https://www.curatedmcp.com/auditor.

AWS MCP Suite: Query and Manage Your Entire AWS Cloud via Natural Language

curatedmcp — Mon, 07 Sep 2026 10:39:30 +0000

AWS MCP Suite: Query and Manage Your Entire AWS Cloud via Natural Language

What It Does

The AWS MCP Suite is the official bridge between AI agents and your AWS infrastructure. Instead of wrestling with SDK boilerplate, IAM policy syntax, and credential management, you can now ask Claude or Cursor to inspect EC2 instances, query DynamoDB tables, upload to S3, invoke Lambda functions, and manage RDS databases—all through conversation.

The suite wraps seven core AWS services (DynamoDB, S3, Lambda, EC2, RDS/Aurora, CloudWatch, and IAM) into a permission-aware protocol. Your agent inherits your existing AWS credentials—IAM roles, profiles, or environment variables—so there's no extra auth layer to configure. Critically, scoped permissions mean your agent only touches what you've explicitly allowed via your IAM setup. You stay in control.

Real wins: You can ask your agent to "list all EC2 instances in us-east-1 and stop the ones tagged 'dev'," or "create a snapshot of my production RDS database and show me the latest CloudWatch metrics," or "upload this CSV to S3 and update the bucket lifecycle policy." The agent handles the API calls, error handling, and response parsing. You get structured, natural-language infrastructure management.

How to Install

npx -y @aws/mcp

Add it to your Claude Desktop config:

{
  "mcpServers": {
    "aws-mcp-suite": {
      "command": "npx -y @aws/mcp"
    }
  }
}

Same process for Cursor or Windsurf. Your AWS credentials are picked up automatically from ~/.aws/credentials, environment variables, or IAM roles (if running on EC2/Lambda).

Real-World Use Cases

Infrastructure auditing: Ask Claude to scan all your EC2 security groups and report which ones allow unrestricted inbound traffic on port 22, then remediate them in one shot.
Database operations: Query RDS logs, create snapshots, and check Aurora performance metrics without leaving your editor—great for debugging production issues on the fly.
CI/CD helpers: Invoke Lambda functions directly from Cursor to trigger deployments, check job status, or pull CloudWatch logs—no terminal tab-switching needed.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Linear MCP: Stop Context-Switching Between Claude and Linear

curatedmcp — Sun, 06 Sep 2026 10:39:34 +0000

Linear MCP: Stop Context-Switching Between Claude and Linear

Engineering teams switch between their AI agent and Linear dozens of times a day. You ask Claude for help debugging, then manually hop over to Linear to create a ticket. You're in a sprint planning call, need to check capacity, and alt-tab away from your IDE. Linear MCP eliminates that friction by giving your AI agent direct access to your entire issue tracker.

This is the official Linear integration—built and maintained by Linear themselves. It connects Claude, Cursor, Windsurf, and other MCP-compatible agents to your Linear workspace with full OAuth support, so each team member's permissions stay intact. Your AI can now create issues, manage sprints, update statuses, query workload, and pull metrics without you leaving your flow.

What It Does

Linear MCP unlocks five core capabilities:

Issue Management: Create fully-formed bugs and features with title, description, assignee, labels, and priority in a single prompt. Search and filter across projects, assignees, statuses, or labels to find exactly what you need.

Workflow Automation: Move issues through your team's workflow states, add comments with context (like PR links), and link related tickets—all without touching the Linear UI.

Sprint & Cycle Visibility: Query what's in the current sprint, check team capacity, and get cycle completion metrics to answer "what did we ship last week?" in seconds.

Project Strategy: Access your roadmap, explore project hierarchy, and check milestone status for planning conversations with your AI agent.

Per-User Security: OAuth authentication means each team member's access is scoped to what Linear already allows them to see.

The result: your AI agent becomes a genuine productivity multiplier for engineering teams, not just a chatbot you paste code into.

How to Install

Install via npm:

npx -y @linear/mcp

Then add Linear MCP to your Claude Desktop configuration:

{
  "mcpServers": {
    "linear-mcp": {
      "command": "npx -y @linear/mcp"
    }
  }
}

Authenticate with your Linear workspace when prompted, and you're live.

Real-World Use Cases

Instant Bug Reports: Say "Create a high-priority bug: login timeout on mobile, affects production" and Claude generates the ticket with full details and assigns it to you—no manual typing.
Sprint Triage in Chat: Ask Claude "Show me all blocked issues in the current sprint" and get a filtered list with context, ready to discuss in standup.
Link PR to Issue: After code review, tell Claude "Move ENG-431 to In Review and add a comment with [PR link]"—one command, no context switch.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Vercel MCP: Debug Production Errors Without Leaving Your Editor

curatedmcp — Sat, 05 Sep 2026 10:39:32 +0000

Vercel MCP: Debug Production Errors Without Leaving Your Editor

The worst part of Next.js development isn't writing code—it's the "works locally but not on Vercel" nightmare. You deploy, something breaks in production, and you're stuck copy-pasting logs into ChatGPT while Claude waits helplessly outside your deployment dashboard. Vercel MCP fixes that.

This official Vercel integration brings your entire deployment workflow directly into Claude, Cursor, or Windsurf. Your AI agent can now inspect deployments, pull build logs from failing builds, manage environment variables, check edge function analytics, and even create new projects—all without you manually hunting through the Vercel dashboard.

What It Does

Vercel MCP unlocks these capabilities for your AI agent:

List and inspect deployments across all your projects with real-time status
Pull build logs from failing deployments so Claude can actually see what went wrong
Manage environment variables (create, update, delete) for any deployment environment
Create and configure projects from scratch when you need to spin up new ones
Monitor deployment health including build times and performance metrics
Access edge function logs and analytics for debugging serverless issues
Handle custom domains and DNS configuration programmatically
Manage team members and permissions across your Vercel organization

The key difference: instead of describing errors in words, Claude reads the actual logs. It traces the stack trace, identifies the root cause, and suggests a fix based on what actually happened in your build.

How to Install

Install the Vercel MCP server with a single command:

npx -y @vercel/mcp

Then add it to your Claude Desktop config:

{
  "mcpServers": {
    "vercel-mcp": {
      "command": "npx -y @vercel/mcp"
    }
  }
}

You'll need a Vercel API token set as the VERCEL_API_TOKEN environment variable.

Real-World Use Cases

Debug a failed production deployment: Paste your project URL and ask Claude to fetch the latest build logs, identify the error, and suggest a fix—all in one turn.
Sync environment variables across environments: Ask Claude to read your production env vars and replicate them to staging, ensuring consistency without manual copy-paste.
Investigate DNS or domain issues: Have Claude check your custom domain configuration, DNS records, and SSL certificate status when deployments aren't accessible.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Figma MCP: Read Design Tokens and Components Directly Into Your AI Coding Session

curatedmcp — Fri, 04 Sep 2026 10:39:35 +0000

Figma MCP: Read Design Tokens and Components Directly Into Your AI Coding Session

Design handoff friction is real. You're building a component, context-switch to Figma to grab a hex value, squint at the typography panel, guess at spacing. Figma MCP eliminates that loop entirely.

This official Figma integration lets Claude, Cursor, and Windsurf read your design files directly—no screenshots, no manual specs. Your AI agent can extract exact token values, component definitions, Auto Layout constraints, and frame measurements. It's the design-to-code gap closed.

What It Does

Figma MCP gives AI agents programmatic access to your Figma workspace. Read file contents, component variants, and design tokens in real time. Extract precise hex colors, font weights, line heights, spacing scales, border radius values, and responsive layout constraints without leaving your editor.

The server handles the structural stuff too: navigate page hierarchies, read frame coordinates and dimensions, access component properties, pull frame comments and annotations. You can even export specific frames as images for reference.

The impact is immediate. Before: "Make the button padding match the design." After: AI reads padding: 12px, background: #2563EB, font-weight: 600 directly from your Figma file and implements it. Pixel-perfect, first try.

How to Install

Add Figma MCP to your local environment:

npx -y figma-developer-mcp

Configure it in Claude Desktop (or Cursor/Windsurf—docs linked below):

{
  "mcpServers": {
    "figma-mcp": {
      "command": "npx -y figma-developer-mcp"
    }
  }
}

You'll need a Figma personal access token in your environment. See the docs for setup details.

Real-World Use Cases

Token-driven component building: Ask Claude to build a button component and specify "use tokens from my Figma file." AI reads your exact color palette, typography scale, and spacing tokens, then generates components that match your design system without manual copying.
Design system documentation: Extract all component variants, their properties, and constraints in one go. Feed this to Claude to auto-generate component stubs, prop types, or Storybook documentation.
Responsive layout implementation: Have Claude read your Auto Layout rules and responsive breakpoints directly from Figma frames, then generate CSS or React layouts that respect your design constraints.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Cloudflare MCP: Manage Your Edge Infrastructure Through AI

curatedmcp — Thu, 03 Sep 2026 10:39:32 +0000

Cloudflare MCP: Manage Your Edge Infrastructure Through AI

Cloudflare MCP is the official integration that lets Claude, Cursor, and other AI agents control your entire Cloudflare ecosystem via natural language. No manual dashboard clicks. No copy-pasting API calls. Just tell your AI what you need, and it deploys Workers, manages databases, updates DNS records, and monitors analytics in real time.

What It Does

This server bridges AI agents directly to Cloudflare's infrastructure layer. You get full programmatic access to:

Cloudflare Workers — deploy, update, and manage serverless functions
KV & D1 — read/write configuration data and query SQLite databases
R2 Object Storage — upload, list, and manage files at edge scale
DNS & Pages — configure zones, records, and static site deployments
Analytics & WAF — monitor traffic patterns and security rules

The key unlock: instead of context-switching between your IDE, the Cloudflare dashboard, and API docs, you describe what you want in plain English. "Deploy a new Worker that handles CORS preflight requests" or "Show me my zone's request volume for the last 24 hours"—and it just works.

Authentication uses OAuth, so your credentials stay secure and no API keys are stored locally.

How to Install

npx -y @cloudflare/mcp-server-cloudflare

Add to your Claude Desktop config:

{
  "mcpServers": {
    "cloudflare-mcp": {
      "command": "npx -y @cloudflare/mcp-server-cloudflare"
    }
  }
}

Then restart Claude and authenticate via the OAuth flow when prompted.

Real-World Use Cases

Deploy a CORS-handling Worker on demand — "Create a new Worker that proxies requests to my API and adds the correct CORS headers for my domain"
Automate configuration storage — "Store this feature flag configuration in KV namespace my-config with a 24-hour TTL"
Debug infrastructure at scale — "Show me analytics for the last 7 days, broken down by country, and identify any traffic spikes"

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Stripe MCP: Let Your AI Agent Manage Your Entire Revenue Stack

curatedmcp — Wed, 02 Sep 2026 10:39:32 +0000

Stripe MCP: Let Your AI Agent Manage Your Entire Revenue Stack

Stop context-switching between your IDE and the Stripe dashboard. Stripe MCP is the official Model Context Protocol server that gives Claude, Cursor, and other AI agents safe, structured access to your entire Stripe API—so you can manage payments, customers, subscriptions, and billing through natural language.

What It Does

Stripe MCP unlocks a new workflow: instead of manually navigating dashboards or writing API calls, you ask your AI agent to handle revenue operations. Create and manage customers, payment intents, and invoices. Query subscription statuses and billing cycles. Generate payment links. Access financial reports and balance data. Handle refunds, inspect webhooks, and replay events—all without leaving your agent conversation.

The server runs in test mode by default, so you can safely experiment without touching real charges. It's built and maintained by Stripe, meaning it's kept in sync with API updates and follows Stripe's security best practices. Perfect for developers who want their agents to autonomously handle billing workflows, respond to customer payment issues, or generate revenue reports on demand.

How to Install

Install via npm:

npx -y @stripe/mcp

Then add it to your Claude Desktop config:

{
  "mcpServers": {
    "stripe-mcp": {
      "command": "npx -y @stripe/mcp"
    }
  }
}

Authenticate using your Stripe API keys (restrict to test keys initially). Once connected, your agent has access to the full Stripe toolkit.

Real-World Use Cases

Instant subscription checks: Ask Claude "What's the subscription status for john@example.com?" and get live billing data without opening Stripe.
Self-service payment links: "Create a payment link for the Pro plan at $49/month" — your agent generates a link, you share it.
Billing triage: "Show me all failed payments in the last 7 days" — your agent pulls the data so you can prioritize customer outreach.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

GitHub Copilot MCP: Bring Code Intelligence Into Your AI Agent

curatedmcp — Tue, 01 Sep 2026 10:39:31 +0000

GitHub Copilot MCP: Bring Code Intelligence Into Your AI Agent

GitHub Copilot MCP connects GitHub's trained code AI directly to Claude, Cursor, or any MCP-compatible agent—no more context-switching between tools. While the GitHub MCP server gives you raw repository access, this server layers in Copilot's code understanding: completions, explanations, security checks, and PR reviews. Think of it as adding a senior engineer's code judgment to your AI workflows.

What It Does

This MCP server unlocks Copilot's capabilities within your agent environment. You get code completions grounded in your repository context, natural language-to-code generation, automated vulnerability detection, and PR review suggestions—all callable from Claude Desktop or Cursor without leaving your workflow.

The real power emerges when paired with GitHub MCP: one gives you repo access, the other gives you code intelligence. Together, they enable autonomous code review, intelligent refactoring with explanations, test generation with edge cases, and even security-focused Autofix suggestions. Multi-file context awareness means the agent understands your full codebase, not isolated snippets.

Features include code completions, natural language generation, security vulnerability detection, PR review suggestions, test generation, code refactoring, and Copilot Chat integration for interactive guidance.

How to Install

npx -y @github/copilot-mcp

Add to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "github-copilot-mcp": {
      "command": "npx -y @github/copilot-mcp"
    }
  }
}

You'll need an active GitHub Copilot subscription ($10/month individual, $19/seat/month for teams).

Real-World Use Cases

Autonomous PR Review: Point Claude at a pull request; it scans the diff, identifies style violations, security issues, and missing tests, then generates a detailed review—all without manual context-switching.
Bulk Test Generation: Feed a module to Claude with Copilot MCP enabled; it generates comprehensive tests with edge cases, boundary conditions, and security considerations.
Security-First Refactoring: Paste legacy code; Claude explains the refactoring, applies Copilot's Autofix for known vulnerabilities, and generates the improved version with inline documentation.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

MCP Ecosystem Week 36: When Official Integrations Become Your Biggest Governance Challenge

curatedmcp — Mon, 31 Aug 2026 10:39:41 +0000

MCP Ecosystem Week 36: When Official Integrations Become Your Biggest Governance Challenge

The MCP ecosystem is consolidating around official vendor integrations. This week's usage data shows developers gravitating hard toward GitHub Copilot, OpenAI, Figma, and Anthropic's own MCP servers—a pattern that surfaces a critical governance decision for platform teams: how to manage the blast radius when first-party tools become MCP clients.

This Week in MCP

No new servers were added to the CuratedMCP catalog this week. The catalog remains at 73 risk-classified servers, all free to review. The flat week on new submissions isn't a slowdown—it reflects consolidation. Developers are asking for deeper integrations with tools they already own (GitHub, OpenAI, Figma, Anthropic), not more point solutions. For platform teams, this is a clarifying signal: the governance battle isn't about blocking niche tools anymore. It's about controlling access to the integrations your developers will demand.

On the Radar

GitHub Copilot MCP (98,010 views) and GitHub MCP (76,020 views) lead adoption this week. Both grant Claude and other MCP clients direct access to GitHub repositories, issues, pull requests, and workflows. Before allowlisting: verify your authentication model (token scope, expiration, revocation). If one developer's GitHub token is embedded in an MCP server config, a single compromised machine leaks access to all your repos. Enforce token rotation and consider machine-level enforcement through CuratedMCP's per-device policies.

OpenAI MCP (87,008 views) bridges Claude to GPT-4o, DALL-E, Whisper, and Embeddings. The governance question: are you accidentally creating a dual-model spend pattern? Developers might route work to OpenAI endpoints through Claude without visibility. This is where TokenShield's spend ledger becomes essential—you need a live view of which models are running where before you can govern them.

Figma MCP (82,012 views) pulls design files and component tokens directly into AI workflows. The risk: design tokens and component structures can leak proprietary design patterns. Allowlist this only after confirming your Figma workspace permissions are segregated by team, and that token-level audit logs are enabled.

Anthropic Claude MCP (76,006 views) enables nesting Claude calls within Claude. This multiplies token spend and latency, but also creates reasoning chains that are harder to audit. If you allowlist this, require that nested Claude calls are logged separately and that developers document why they're using sub-agents instead of a single, direct call.

Governance Take

Here's what keeps most platform leaders up at night: developers are already using these servers in Cursor and Windsurf outside your allowlist, and you don't know it yet. Shadow MCP usage is real. The GitHub Copilot and OpenAI numbers this week aren't coincidental—they reflect what developers want to do, not what your policy says they can do.

Start a two-week audit. Query your IDE telemetry, GitHub Actions logs, and your Claude API spend. Where are MCP calls happening? Which authentication tokens are in play? Then decide: do you allowlist the official integrations under tight auth controls, or do you ban them and accept that you're now managing workarounds?

One more thing: if you're running TokenShield, this week is the moment to pull spend data by MCP server and model. You'll likely find that OpenAI MCP and Claude MCP usage is adding up faster than expected. Visibility first, then policy.

Govern MCP usage across your team with CuratedMCP — or scan your own stack free at https://www.curatedmcp.com/auditor.

Zoom MCP: Automate Meeting Management and Transcript Processing in Your AI Workflows

curatedmcp — Mon, 31 Aug 2026 10:39:31 +0000

Zoom MCP: Automate Meeting Management and Transcript Processing in Your AI Workflows

Zoom MCP is the official integration that connects Claude, Cursor, and other AI agents directly to your Zoom workspace. Instead of manually juggling meeting links, transcripts, and recordings, your AI can now schedule meetings, retrieve transcripts, manage recordings, and pull attendance data—all through natural language commands.

Think of it as giving Claude access to your entire Zoom calendar and meeting history. It works with Zoom Workplace Business and above, and it's free.

What It Does

Zoom MCP unlocks several capabilities for AI agents:

Meeting scheduling: Create meetings with participants, set agendas, configure settings, and generate shareable links
Transcript access: Retrieve full meeting transcripts and AI-generated summaries from Zoom AI Companion
Recording management: Download, list, and organize cloud recordings programmatically
Analytics: Pull attendance reports, participation metrics, and engagement data
Webinars: Create and manage webinar registrations
Team chat: Access Zoom Team Chat history for context
Zoom Apps: Build and manage custom integrations within Zoom

The killer feature is post-meeting automation. After a call ends, Claude can fetch the transcript, extract action items, update your Linear or Jira board (via other MCPs), and draft follow-up emails—no manual work required.

How to Install

Install Zoom MCP with a single command:

npx -y @zoom/mcp

Then add it to your Claude Desktop configuration:

{
  "mcpServers": {
    "zoom-mcp": {
      "command": "npx -y @zoom/mcp"
    }
  }
}

You'll need valid Zoom API credentials (OAuth token or JWT). See the official docs for credential setup.

Real-World Use Cases

Automated meeting summaries: Schedule a meeting, let Claude sit in on it (via transcript access), extract decisions and TODOs, and sync them to your project management tool instantly
Meeting prep at scale: Have Claude pull transcripts from your last 5 standups, identify recurring blockers, and draft a summary for your engineering lead
Recording library: Ask Claude to find all recordings from Q4, organize them by project, and generate searchable summaries for knowledge sharing

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Airtable MCP: Give Claude Full Read-Write Access to Your No-Code Databases

curatedmcp — Sun, 30 Aug 2026 10:39:31 +0000

Airtable MCP: Give Claude Full Read-Write Access to Your No-Code Databases

What It Does

Airtable MCP bridges the gap between your AI agent and the databases powering your team's operations. Instead of copying data in and out manually, Claude (or Cursor) can now query records, update fields, create entries, and manage views directly within Airtable — all through natural language.

The server unlocks the full Airtable API surface: list and search across bases, create/update/delete records, work with linked records and lookups, apply filters using Airtable's formula syntax, and batch operations for high-volume changes. You get access to automation history too, so Claude can understand what's already been triggered.

This is especially powerful because Airtable is where many teams have already built their operational spine — CRMs, project trackers, content calendars, inventory systems. Now your AI can actually work inside those systems instead of just reading static exports.

How to Install

Run this single command:

npx -y @airtable/mcp-server

Then add it to your Claude Desktop config:

{
  "mcpServers": {
    "airtable-mcp": {
      "command": "npx -y @airtable/mcp-server"
    }
  }
}

You'll need an Airtable API token (generate one in your account settings). The server handles auth automatically once configured.

Real-World Use Cases

CRM automation: Tell Claude "update all deals in the lost column to closed, add a note explaining why, and flag for follow-up next quarter." It queries the view, filters records, updates multiple fields, and logs the changes — no spreadsheet juggling.
Content calendar triage: "Review my editorial calendar, identify pieces missing reviewer assignments, and auto-assign them based on topic." Claude reads your linked records, understands your tagging system, and updates assignments in bulk.
Bug triage with linking: "Create a critical bug ticket in Airtable, link it to this GitHub PR, set priority to P1, and notify the on-call engineer via the automation." The server creates the record, manages relationships, and triggers your existing Airtable automations.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Shopify MCP: Let Claude Manage Your Store Like a Command Line

curatedmcp — Sat, 29 Aug 2026 10:39:33 +0000

Shopify MCP: Let Claude Manage Your Store Like a Command Line

Managing a Shopify store through an AI agent just became possible. Shopify MCP is the official Model Context Protocol server that gives Claude, Cursor, and other AI agents direct access to your store's Admin API—letting you automate inventory, orders, customers, and pricing through natural language commands.

What It Does

Shopify MCP bridges the gap between AI agents and your Shopify backend. Instead of toggling between dashboards and spreadsheets, you can ask Claude to fetch order data, update product prices, manage inventory, or process refunds—all in conversation.

The server unlocks granular control: create and update products with variants, process orders (fulfill, refund, cancel), access customer profiles and purchase history, manage collections, update theme content, run direct GraphQL queries, handle promotions, and pull analytics. It works across all Shopify plans and uses the Admin API with scoped permissions, so you control exactly what your AI agent can touch.

Think of it as a programmatic storefront manager. No APIs to debug. No boilerplate. Just natural language.

How to Install

Install globally via npm:

npx -y @shopify/mcp

Then configure in your Claude Desktop config file (claude_desktop_config.json):

{
  "mcpServers": {
    "shopify-mcp": {
      "command": "npx -y @shopify/mcp",
      "env": {
        "SHOPIFY_ACCESS_TOKEN": "your_admin_api_token",
        "SHOPIFY_SHOP_URL": "https://your-store.myshopify.com"
      }
    }
  }
}

Get your access token from the Shopify Admin → Apps and integrations → Develop apps → Custom apps.

Real-World Use Cases

Bulk pricing updates: "Increase all Summer collection products by 15%, except clearance items" — no manual edits, no mistakes.
Smart order filtering: "Show me orders over $200 that haven't shipped in 3 days and flag them for priority fulfillment" — real-time inventory insights without leaving your chat.
Rapid product launches: "Create a new sneaker product with sizes 6–14, add images, set pricing, and assign to the New Arrivals collection" — from brief to live in seconds.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.