MCP Enhanced Server
Provides file system tools, code analysis prompts, async tasks, and secure HTTP transport following the 2025-03-26 MCP specification.
README
MCP Enhanced Server v2.0.0
Enhanced MCP (Model Context Protocol) Server with full 2025-03-26 specification support:
- Tools - With annotations, outputSchema, structuredContent, resource_link
- Resources - URI-identified data with subscriptions and list change notifications
- Prompts - Templated messages and workflows with argument schemas
- Elicitation - Server-initiated user input requests
- Tasks - Async long-running operations with progress tracking
- Streamable HTTP - Production-ready HTTP transport with session management
- Auth - Bearer token authentication for remote deployments
- Origin Validation - DNS rebinding attack prevention
Quick Start
npm install
npm run build
npm start
Transport Modes
Stdio (default)
npm start
Streamable HTTP
MCP_TRANSPORT=http MCP_PORT=3000 npm start
Streamable HTTP with Auth
MCP_TRANSPORT=http MCP_PORT=3000 MCP_AUTH_TOKEN=your-secret-token npm start
Features
Resources
| Resource | URI | Description |
|---|---|---|
| Server Info | mcp://server/info |
Server name, version, features |
| Config Info | mcp://server/config |
Transport, auth, environment |
| Project Files | file:///{path} |
Template-based file access |
Prompts
| Prompt | Description | Arguments |
|---|---|---|
code_review |
Review code quality | code, language |
explain_code |
Explain code behavior | code, detail_level |
generate_tests |
Generate unit tests | code, framework |
debug_error |
Debug code errors | code, error |
Tools
| Tool | readOnly | destructive | idempotent | Description |
|---|---|---|---|---|
read_file |
✓ | ✗ | ✓ | Read file contents |
write_file |
✗ | ✓ | ✓ | Write file (overwrites) |
list_directory |
✓ | ✗ | ✓ | List directory contents |
delete_file |
✗ | ✓ | ✗ | Delete a file |
search_code |
✓ | ✗ | ✓ | Regex search in code |
run_command |
✗ | ✓ | ✗ | Execute shell command |
confirm_action |
✓ | ✗ | ✓ | Request user confirmation |
request_input |
✓ | ✗ | ✓ | Request user input |
task_status |
✓ | ✗ | ✓ | Get async task status |
task_cancel |
✗ | ✓ | ✓ | Cancel a running task |
task_list |
✓ | ✗ | ✓ | List all tasks |
batch_process |
✗ | ✗ | ✗ | Start batch processing |
get_weather |
✓ | ✗ | ✓ | Weather data (structured output) |
system_info |
✓ | ✗ | ✓ | System information (structured output) |
list_available_resources |
✓ | ✗ | ✓ | List resources as resource links |
Environment Variables
| Variable | Default | Description |
|---|---|---|
MCP_TRANSPORT |
stdio |
Transport: stdio or http |
MCP_PORT |
3000 |
HTTP port |
MCP_HOST |
127.0.0.1 |
HTTP host |
MCP_AUTH_TOKEN |
- | Bearer token for auth (optional) |
Cloudflare Workers Deployment
Live deployment: https://mcp-enhanced-server.catnip-haze.workers.dev
- Health:
GET /health - MCP endpoint:
POST /mcp(JSON-RPC 2.0) - MCP info:
GET /mcp
See cloudflare/ directory for Workers deployment with D1 database.
cd cloudflare
npm install
wrangler d1 create mcp-enhanced-tasks
# Update database_id in wrangler.toml
wrangler d1 execute mcp-enhanced-tasks --file=src/schema.sql
wrangler deploy
Workers Environment Variables
Set via Cloudflare Dashboard or wrangler secret put:
MCP_AUTH_TOKEN- Authentication tokenMCP_ALLOWED_ORIGINS- Comma-separated allowed origins
Architecture
src/
├── index.ts # Main entry, server setup, all features
├── tools/
│ ├── annotations.ts # Tool Annotation type definitions
│ └── registry.ts # Tool registry with Zod validation
├── resources/
│ ├── manager.ts # Resource lifecycle manager
│ └── tools.ts # Resource tools
├── prompts/
│ ├── manager.ts # Prompt lifecycle manager
│ └── builtin.ts # Built-in prompt templates
├── elicitation/
│ ├── manager.ts # Elicitation request/response manager
│ └── tools.ts # Elicitation tools
└── tasks/
├── manager.ts # Async task lifecycle manager
└── tools.ts # Task tools
Debug
npx @modelcontextprotocol/inspector node dist/index.js
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。