チームマーケットプレイスでのMCPと組織管理機能 MCPs and Organizations in Team Marketplaces

Cursor has expanded its Team Marketplace to support the Model Context Protocol (MCP) and organization-level management, giving companies a central way to distribute and standardize AI tooling across every developer on their teams. For organizations that have struggled to keep configurations consistent as their use of AI coding assistants grows, this change addresses a practical pain point: ensuring that everyone works with the same approved tools, context sources, and settings rather than assembling them individually.

The Team Marketplace is Cursor's mechanism for sharing reusable assets within a workspace. With MCP support added, administrators and teams can now publish and adopt MCP servers through the same channel. MCP, an open standard introduced by Anthropic in late 2024, defines a common way for AI applications to connect to external data sources, tools, and services. Instead of building bespoke integrations for each connection, developers can point a model at MCP servers that expose resources such as documentation, databases, ticketing systems, internal APIs, or version control. Bringing MCP into the marketplace means those connections can be curated once and shared, rather than configured repeatedly by each engineer.

The organization-level management piece is the second half of the update. Where team settings typically apply to a single workspace, organization controls appear to operate one tier higher, spanning multiple teams under a single corporate account. This structure is common in enterprise software, and it allows a central administrator to define defaults, approve or restrict specific MCP servers, and roll out standardized configurations to everyone at once. The stated goal is to make it easier for large organizations to unify their development environments, which is often difficult when tooling decisions are left entirely to individual developers or small teams.

For readers less familiar with the product, Cursor is an AI-native code editor built as a fork of Visual Studio Code, developed by Anysphere. It integrates large language models directly into the editing experience for tasks such as code completion, chat-based assistance, and multi-file edits. Because it inherits the VS Code foundation, it supports many of the same extensions and workflows developers already know, while layering AI features on top. As adoption has moved from individual users to entire engineering organizations, features aimed at governance, consistency, and administrative control have become increasingly relevant, and this marketplace update fits that trajectory.

The emphasis on MCP also reflects a broader industry shift. Since its release, MCP has been adopted or supported by a growing number of tools and vendors, including AI assistants and development platforms that want a vendor-neutral way to connect models to context. Supporting MCP through a shared marketplace lowers the friction of that connection for teams and aligns Cursor with an emerging standard rather than a proprietary integration approach. This is likely to appeal to organizations that want to avoid lock-in and to reuse the same MCP servers across different AI clients where possible.

Standardization matters for reasons beyond convenience. When developers configure their own tools and data connections, organizations can face inconsistent results, security gaps, and difficulty auditing what external systems an AI assistant can reach. Centralized management gives administrators a clearer view of which MCP servers are in use and the ability to enforce policy. For regulated industries or teams handling sensitive code, that visibility is often a prerequisite for approving AI tools at all. The update appears designed to make Cursor more defensible in those settings by turning ad hoc setups into managed, shareable defaults.

Specific details such as how permissions are scoped, whether MCP servers can be marked as required versus optional, and how the feature interacts with existing team billing or privacy modes were not fully detailed in the changelog summary, so organizations evaluating it will want to confirm those points in the official documentation. As with any centralized configuration system, the practical value will depend on how granular the controls are and how smoothly they integrate with existing identity and access management. Still, the direction is clear: Cursor is positioning its Team Marketplace as a governance layer for AI-assisted development at scale, combining a widely supported connection standard with enterprise-oriented administrative controls.