TL;DR: A powerful MCP server that gives AI assistants full directory analysis and file editing capabilities.

• 📁 Analyze directories - Get project structure, find files, search content
• ✏️ Edit files - Create, update, delete, rename files and folders
• 🔍 Smart search - Find code/text across multiple files with regex support
• 🔄 Git integration - View commits, branches, and changes
• 🛡️ Safe operations - Auto-backups and transaction rollback

1. Install dependencies: npm install
2. Add to your Claude Desktop config:

{
  "mcpServers": {
    "directory-context": {
      "command": "node",
      "args": ["path/to/server.js"]
    }
  }
}

A comprehensive Model Context Protocol (MCP) server that provides advanced directory analysis and file management capabilities for AI assistants like Claude. This server combines directory context extraction with full file editing capabilities, making it a powerful tool for code analysis, project management, and file operations.

• Complete Directory Structure: Get detailed tree views of any directory with metadata
• Intelligent Project Analysis: Automatic project type detection and context summary
• Smart File Search: Regex and text-based search across multiple file types
• Git Integration: Extract repository context, commit history, and working directory status
• File Statistics: Comprehensive analysis of file types, sizes, and modification dates

• Create Files: Generate new files with specified content
• Update Files: Sophisticated search-and-replace operations with regex support
• Append Content: Add content to existing files with formatting options
• Delete Files: Remove files with optional backup creation
• Rename/Move Files: Relocate and rename files with collision detection
• Directory Management: Create directory structures recursively

• Batch Operations: Execute multiple file operations in a single transaction
• Automatic Backups: Optional backup creation for destructive operations
• Transaction Rollback: Rollback capability for failed batch operations
• Resource Exposure: Automatic exposure of important project files as MCP resources
• Error Handling: Comprehensive error handling with detailed feedback

• Node.js 18.0.0 or higher
• npm or yarn package manager

1. Clone or download the MCP server files
2. Install dependencies:

npm install

3. Configure your MCP client (Claude Desktop, etc.) to use this server:

Claude Desktop Configuration: Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "directory-context": {
      "command": "node",
      "args": ["path/to/your/server.js"],
      "cwd": "your-default-working-directory"
    }
  }
}

Alternative Configuration: Use the provided mcp-json-config.json as a template and update the paths accordingly.

Before using any features, set your working directory:

set_working_directory(directory="/path/to/your/project")

get_directory_structure(
  max_depth=10,           // Maximum traversal depth
  include_hidden=false,   // Include hidden files/directories
  file_types=[".js", ".py", ".md"]  // Filter by extensions
)

analyze_project_context(
  include_code_analysis=true,    // Include complexity analysis
  focus_files=["specific_file.js"]  // Focus on specific files
)

search_files(
  query="function myFunction",  // Search query
  file_types=[".js", ".ts"],   // File types to search
  is_regex=false,              // Use regex matching
  max_results=50               // Maximum results
)

get_git_context(
  include_diff=true,     // Include working directory changes
  commit_count=10        // Number of recent commits
)

create_file(
  path="src/new-component.js",
  content="const MyComponent = () => { ... }",
  encoding="utf8",
  overwrite=false
)

update_file(
  path="src/config.js",
  updates=[
    {
      search: "oldValue: 'old'",
      replace: "newValue: 'new'",
      regex: false,
      all: true
    }
  ],
  backup=true
)

append_to_file(
  path="README.md",
  content="## New Section\nContent here...",
  newline_before=true
)

delete_file(
  path="old-file.js",
  backup=true  // Creates .deleted backup
)

rename_file(
  old_path="old-name.js",
  new_path="new-name.js",
  overwrite=false
)

create_directory(
  path="src/components/new-feature",
  recursive=true
)

Execute multiple operations atomically:

batch_file_operations(
  operations=[
    {
      operation: "create",
      params: {
        path: "src/component.js",
        content: "export default () => {}"
      }
    },
    {
      operation: "update",
      params: {
        path: "src/index.js",
        updates: [...]
      }
    }
  ],
  rollback_on_error=true
)

The server automatically detects project types based on configuration files:

• Node.js/JavaScript: package.json
• Python: requirements.txt
• Rust: Cargo.toml
• Go: go.mod
• Java: pom.xml, build.gradle
• PHP: composer.json
• Ruby: Gemfile
• Docker: Dockerfile, docker-compose.yml
• And more...

The server identifies important files automatically:

• Configuration files (package.json, config.js, etc.)
• Entry points (index.js, main.py, app.js)
• Documentation (README.md, docs)
• Docker files
• Environment files (.env)

Important files are automatically exposed as MCP resources, making them easily accessible to AI assistants without explicit file reading requests.

• Backup Creation: Automatic backups for destructive operations
• Collision Detection: Prevents accidental overwrites
• Transaction Rollback: Batch operations can be fully rolled back on error
• Path Validation: Ensures operations stay within working directory bounds
• Detailed Error Messages: Comprehensive error reporting for debugging

• Working Directory: Set via set_working_directory or during initialization
• File Type Filters: Configurable file extension filtering
• Search Limits: Configurable maximum search results
• Backup Policies: Optional backup creation for destructive operations

Configure the server in your MCP client with appropriate command and arguments. The server communicates via stdio and requires no additional network configuration.

enhanced-directory-context-mcp/
├── server.js                 # Main MCP server implementation
├── package.json             # Node.js package configuration
├── mcp-json-config.json     # Example MCP client configuration
└── README.md               # This documentation

• @modelcontextprotocol/sdk: Core MCP functionality
• Node.js built-in modules: fs, path, child_process

npm run dev

This starts the server with Node.js inspector enabled for debugging.

MIT License - See package.json for details.

For issues, feature requests, or contributions, please refer to the repository's issue tracker.

Note: This MCP server provides powerful file system access. Ensure you trust the AI assistant and understand the operations being performed, especially destructive operations like file deletion or modification.