mcp-management

name	mcp-management
description	Configure and manage Model Context Protocol servers for external tool access
compatibility	>=1.4

MCP Management

Skill metadata: version "1.1"; license MIT; tags [mcp, servers, configuration, integration]; compatibility ">=1.4"; recommended tools [codebase, editFiles, fetch].

MCP (Model Context Protocol) is GA in VS Code as of v1.102. MCP servers provide tools, resources, and prompts beyond built-in capabilities. Configuration lives in .vscode/mcp.json (workspace-scoped) or profile-level mcp.json (user-scoped).

When to use

• The user asks to configure, add, list, or check MCP servers
• You need to determine which MCP servers are available
• A task would benefit from an external tool not yet configured

Configuration locations

Location	Scope	When to use
.vscode/mcp.json	Workspace	Project-specific servers shared via version control
Profile-level mcp.json	User	Personal servers available across all workspaces
settings.json "mcp" key	User/Workspace	Alternative to standalone mcp.json
Dev container customizations.vscode.mcp	Container	Per-container MCP servers

VS Code commands:

• MCP: Open Workspace Configuration — edit .vscode/mcp.json
• MCP: Open User Configuration — edit profile-level mcp.json

Server tiers

Tier	Default servers	When to enable	Configuration
Always-on	filesystem, git	Every project — core development tools	Enabled by default in .vscode/mcp.json
External access	github, fetch	When GitHub or web access is needed	github uses VS Code OAuth (HTTP remote, no PAT required); fetch needs no credentials
Documentation	context7	Any project using third-party libraries	HTTP remote, free tier requires no auth; optional API key for higher rate limits

Available servers

Server	Tier	Transport	Purpose
@modelcontextprotocol/server-filesystem	Always-on	npx (stdio)	File operations within the workspace; supports OS-level sandboxing
mcp-server-git	Always-on	uvx (stdio, Python)	Git history, diffs, and branch operations
github/github-mcp-server	Credentials	HTTP remote (https://api.githubcopilot.com/mcp/)	GitHub API — issues, PRs, repos, Actions, CI/CD, security alerts, Dependabot
mcp-server-fetch	Credentials	uvx (stdio, Python)	HTTP fetch for web content and APIs
@upstash/context7-mcp	Documentation	HTTP remote (https://mcp.context7.com/mcp)	Live, version-specific library documentation — prevents hallucinated or outdated APIs

Removed (v3.2.0): @modelcontextprotocol/server-memory — replaced by VS Code's built-in memory tool (/memories/), which provides persistent storage with three scopes: user (cross-workspace), session (conversation), and repository.

Archived (deprecated): @modelcontextprotocol/server-github (npm) — replaced by github/github-mcp-server remote HTTP server. Do not add new configurations using the archived npm package.

Stack-specific servers

The servers below are not included in the base consumer template. Add them to .vscode/mcp.json based on your project's technology stack.

Stack	Server	Transport	Notes
Browser / UI testing	@playwright/mcp (Microsoft)	npx -y @playwright/mcp@latest	Accessibility-tree-based; no vision model needed
PostgreSQL	Search MCP Marketplace for postgres	varies	Official reference server is archived; use marketplace for maintained replacement
SQLite	Search MCP Marketplace for sqlite	varies	Same — official archived; find active replacement
Redis	Search MCP Marketplace for redis	varies	Same
Docker / containers	Search MCP Marketplace for docker	varies	Several options; evaluate trust and permissions carefully
AWS	Search MCP Marketplace for aws	varies	Use fine-grained IAM credentials via ${env:}, never hardcode

Discover servers: code.visualstudio.com/mcp (gallery) · registry.modelcontextprotocol.io (official registry) · glama.ai · smithery.ai

Optional: Sequential Thinking

@modelcontextprotocol/server-sequential-thinking (npx) — adds a structured step-by-step reasoning tool. Useful for complex planning or debugging sessions that benefit from explicit thought chains. Not project-specific; consider adding to your user-level mcp.json rather than the workspace config.

Configuration:

"sequentialThinking": {
  "type": "stdio",
  "command": "npx",
  "args": ["-y", "@modelcontextprotocol/server-sequential-thinking"]
}

MCP capabilities (GA since v1.102)

MCP servers can expose four capability types:

Capability	Description	Agent interaction
Tools	Functions the agent can invoke (e.g., query database, call API)	Agent calls tools directly
Resources	Data sources the agent can read (e.g., database schemas, config files)	Agent reads from # context menu
Prompts	Reusable prompt templates provided by the server	Available via / slash commands
MCP Apps	Interactive UI components (forms, visualisations, drag-and-drop)	Rendered inline in chat responses
Sampling	Server requests the agent to generate text on its behalf	Agent responds to server requests

Additional features: elicitations (server requests user input via the agent), MCP auth (OAuth/token flows for secure server connections).

Server discovery

• MCP gallery: In the Extensions view, search @mcp to browse and install servers directly (installs to user profile or workspace)
• MCP Marketplace: Browse and install servers from code.visualstudio.com/mcp
• Official registry: github.com/modelcontextprotocol/servers
• Community registries: mcp.so, glama.ai, smithery.ai
• Agent plugins: MCP servers can be bundled inside agent plugins (@agentPlugins in Extensions view)

Adding a new server

Before adding any MCP server:

1. Check if a built-in tool or existing MCP server already covers the need
2. Search the MCP Marketplace (code.visualstudio.com/mcp) and official registry
3. Check for npx vs uvx vs HTTP remote transport — prefer HTTP remote for officially hosted servers (no local process, OAuth-managed auth)
4. Add to .vscode/mcp.json (workspace) or profile mcp.json (user) with appropriate tier
5. For credentials-required stdio servers, use ${input:} or ${env:} variable syntax — never hardcode secrets; HTTP remote servers use VS Code's built-in OAuth where supported
6. For stdio servers on Linux/macOS, add sandboxEnabled: true with sandbox.filesystem.denyRead rules for credential directories (~/.ssh, ~/.gnupg, ~/.aws) as a defence-in-depth measure against prompt injection. Optionally add sandbox.network.allowedDomains to restrict outbound network access
7. Agent files can restrict MCP access using the mcp-servers frontmatter field

Sandbox compatibility (Linux)

On immutable Linux distros (Fedora Atomic/Bazzite/Silverblue, NixOS) where /home is a symlink to /var/home, the bwrap sandbox rejects allowWrite paths because symlink resolution points outside the expected location. Detect at setup time:

[[ "$(readlink -f /home)" != "/home" ]] && echo "immutable" || echo "standard"

• standard: use sandboxed config (sandboxEnabled: true + allowWrite/denyRead rules)
• immutable: omit sandboxEnabled, sandbox, and the top-level sandbox block entirely

Auto-start

Set "chat.mcp.autostart": "newAndOutdated" in .vscode/settings.json so MCP servers start automatically when a chat message is sent. This eliminates the need to manually click the refresh/start button each session. VS Code will show a trust dialog the first time a new or changed server auto-starts.

CLI and external agent access

As of VS Code 1.113, MCP servers configured in .vscode/mcp.json are automatically bridged to Copilot CLI and Claude agents. No additional configuration is required — servers registered in the workspace or user profile are available to all agent runtimes.

Settings Sync

With Settings Sync enabled, MCP server configurations can be synchronised across devices. Run Settings Sync: Configure and enable the MCP Servers option to maintain a consistent development environment across machines.

Subagent MCP use

Subagents inherit access to all configured MCP servers. A subagent may invoke any server already in .vscode/mcp.json. To add a new server, the subagent must flag the proposal to the parent agent, which confirms before modifying .vscode/mcp.json.