Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Code MCP Server open_project my-web-app and show me the main files"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
This monorepo contains the VSCode MCP Server and its companion VSCode Extension, which together enable AI agents and assistants, like Goose or Claude, to interact with VSCode through the Model Context Protocol.
vscode-mcp/
├── server/ # MCP server implementation
└── extension/ # VS Code extensionRelated MCP server: LSP-MCP
Install the MCP Server
npx vscode-mcp-server installInstall the MCP Extension
ID: code-mcp
Name: VS Code
Description: Allows interaction with VS Code through the Model Context Protocol
Command: npx vscode-mcp-server
Add this to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"vscode-mcp-server": {
"command": "npx",
"args": ["vscode-mcp-server"],
"env": {}
}
}
}The Code MCP server provides the following tools for AI agents to interact with VS Code:
create_diffCreates and shows a diff for modifying existing files:
Shows changes preview before applying
Requires user approval
Only works with existing files
open_fileOpens files in the VS Code editor:
Used for viewing new or modified files
open_projectOpens a project folder in VS Code:
Sets up working directory for AI agent
check_extension_statusChecks if extension is installed and responding
get_extension_portGets the port number for VS Code MCP Extension
list_available_projectsShows projects from port registry file
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
Copyright 2025 Block, Inc.
This product includes software developed at Block, Inc.
Add a command to the project's allowed commands list
Inputs: command (required).
Analyze code structure and dependencies
Inputs: path (required), language.
Analyze differences between two files with detailed statistics
Inputs: file1 (required), file2 (required).
Apply a patch file to the workspace
Inputs: patchFile (required), dryRun, reverse, stripPaths, +1 more.
Automatically commit AI-made changes with tracking
Inputs: message (required), files, amendSession.
Automatically select the most appropriate VS Code workspace
Inputs: no named parameters.
Compare two commits with detailed analysis
Inputs: commit1 (required), commit2 (required), filePattern, format, +1 more.
Compare two files and show differences with multiple format options
Inputs: file1 (required), file2 (required), format, contextLines, +4 more.
Create a new directory
Inputs: path (required).
Create a patch file from differences between files or directories
Inputs: source (required), target (required), outputFile.
Create a new project with boilerplate
Inputs: name (required), type (required), path.
Delete a file or directory
Inputs: path (required).
Automatically detect open VS Code instances and their workspaces
Inputs: includeRecent, includeRunning, maxResults.
Build a Docker image from a Dockerfile
Inputs: context, dockerfile, tag, build_args, +4 more.
Check if Docker is installed and available
Inputs: no named parameters.
Clean up tracked Docker containers and resources
Inputs: no named parameters.
Manage Docker Compose services
Inputs: action (required), service, file, detach, +7 more.
Manage Docker containers (list, start, stop, remove, etc.)
Inputs: action (required), container, command, follow, +4 more.
Manage Docker images (list, pull, push, remove, etc.)
Inputs: action (required), image, tag, force, +2 more.
Manage Docker networks
Inputs: action (required), network, container, driver, +2 more.
Run a Docker container
Inputs: image (required), name, command, args, +13 more.
Docker system operations and information
Inputs: action (required), all, volumes, force.
Manage Docker volumes
Inputs: action (required), volume, driver, force, +1 more.
End the current coding session
Inputs: no named parameters.
Show git diff with multiple format options (unified, side-by-side, stat, word-diff)
Inputs: staged, file, format, contextLines, +5 more.
Find and replace text across multiple files with preview option
Inputs: searchPattern (required), replacement (required), files, filePattern, +5 more.
Generate a Docker Compose file template
Inputs: services (required), include_database.
Generate a Dockerfile template for a specific language/framework
Inputs: language (required), framework.
Generate a sample .vscode-mcp.toml configuration file
Inputs: saveToFile.
Get list of commands that are allowed to run in this project
Inputs: no named parameters.
Get information about the current coding session
Inputs: no named parameters.
Get diff statistics (files changed, lines added/removed)
Inputs: staged, file, commit1, commit2, +1 more.
Get history of all coding sessions and their commits
Inputs: limit.
Get information about the current coding session
Inputs: no named parameters.
Get current workspace path
Inputs: no named parameters.
Stage files for commit
Inputs: files, all, cwd.
List, create, or switch branches
Inputs: action (required), name, cwd.
Commit staged changes
Inputs: message (required), cwd.
Show git diff
Inputs: staged, file, cwd.
Show commit history
Inputs: limit, oneline, cwd.
Pull changes from remote
Inputs: remote, branch, cwd.
Push commits to remote
Inputs: remote, branch, cwd.
Get git repository status
Inputs: cwd.
Install all project dependencies
Inputs: path.
List contents of a directory with details
Inputs: path (required), recursive.
List all running processes started by the agent
Inputs: no named parameters.
List recently used workspaces
Inputs: no named parameters.
Load project-specific configuration from .vscode-mcp.toml
Inputs: path.
Move or rename a file or directory
Inputs: source (required), destination (required).
Run npm commands (install, test, build, etc)
Inputs: command (required), args, cwd.
Install Python packages using pip
Inputs: packages, requirements_file, venv.
Present detected workspaces to user for selection
Inputs: no named parameters.
Preview all changes before committing
Inputs: no named parameters.
Enhanced change preview with multiple format options
Inputs: format, contextLines, ignoreWhitespace, filePattern, +1 more.
Read the contents of a file
Inputs: path (required).
Remove a command from the project's allowed commands list
Inputs: command (required).
Rollback all changes made in the current AI session
Inputs: confirm (required), preserveUnstaged.
Execute a shell command
Inputs: command (required), cwd, env.
Execute a custom tool defined in project configuration
Inputs: toolName (required), args.
Execute JavaScript code or script with Node.js
Inputs: code, script_path, args.
Execute Python code or script
Inputs: code, script_path, args, venv.
Run tests using appropriate test runner
Inputs: language (required), framework, path, pattern.
Save project-specific configuration to .vscode-mcp.toml
Inputs: config (required), path.
Search for patterns in code
Inputs: query (required), path (required), file_pattern, regex.
Execute a command with security restrictions (only allowed commands)
Inputs: command (required), cwd, timeout, env, +2 more.
Execute multiple commands in sequence with security restrictions
Inputs: commands (required), cwd, timeout, commitResult, +2 more.
Set the active workspace directory
Inputs: path (required).
Intelligently initialize workspace with VS Code detection and user choice
Inputs: no named parameters.
Start a new coding session for change tracking
Inputs: description (required), branch.
Start a development server
Inputs: command (required), name (required), port, cwd.
Stop a running server
Inputs: name (required).
Update specific project configuration settings
Inputs: allowedCommands, formatOnSave, gitAutoCommit, formatCommand, +1 more.
Write content to a file
Inputs: path (required), content (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 73 tools
Multiple tools have overlapping purposes, causing ambiguity. For example, 'git_diff' and 'enhanced_git_diff' both show git differences, 'preview_changes' and 'preview_changes_enhanced' both preview changes, and 'get_current_session' and 'get_session_info' appear to retrieve the same information. Tools like 'run_command', 'secure_run_command', and 'secure_run_command_sequence' also have unclear boundaries, making it difficult for an agent to choose the right one without trial and error.
Most tools follow a consistent verb_noun or verb_adjective_noun pattern (e.g., 'create_directory', 'analyze_code', 'docker_build'), with clear and descriptive names. However, there are minor deviations, such as 'npm_command' and 'pip_install' using different naming styles, and some redundancy in names like 'git_diff' vs. 'enhanced_git_diff', but overall the naming is readable and predictable.
With 73 tools, the count is excessive for a code-focused server, leading to bloat and potential confusion. Many tools could be consolidated (e.g., multiple Docker tools, redundant git and preview tools) without losing functionality. This large number suggests poor scoping and makes the toolset overwhelming for agents to navigate effectively.
The toolset provides comprehensive coverage for code development, including file operations, git management, Docker containerization, dependency installation, and session tracking. It supports full CRUD/lifecycle operations across these domains, with no obvious gaps. Agents can perform end-to-end workflows from project creation to deployment and cleanup.
Enable secure connectivity between Sentry issues and debugging data, and LLM clients, using a Model Context Protocol (MCP) server.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
A Model Context Protocol server for Wix AI tools
Connect AI agents to Replynodes over the Model Context Protocol.
