Open MCP Server
A modular productivity automation server providing reusable prompt templates, composable skills, and multi-step workflows for tasks like daily planning, code review, document summarization, and project management.
README
Open MCP Server
A modular Model Context Protocol (MCP) server with Prompts, Skills, and Workflows for personal productivity automation.
Features
Core Components
- 6 Prompt Templates - Reusable, parameterizable prompts for common tasks
- 11 Skills - Pre-defined task sequences that compose multiple tools
- 9 Workflows - Multi-step automation pipelines with conditionals and loops
- 15+ Tools - File operations, web scraping, git operations, system info, and AI summarization
Categories
- Productivity: Daily planning, task prioritization, briefings
- Development: Code review, project setup, refactoring guidance
- Research: Topic exploration, document summarization
- Document: File analysis, text summarization, word counting
Quick Start
# Install dependencies
npm install
# Build the project
npm run build
# Start the server
npm start
Server Resources
| Resource | Description |
|---|---|
prompts:// |
List all prompt templates |
prompt://{id} |
Get specific prompt template |
skills:// |
List all available skills |
skill://{id} |
Get specific skill details |
workflows:// |
List all workflows |
workflow://{id} |
Get specific workflow details |
server-info:// |
Server statistics and capabilities |
Available Tools
Prompt Management
list_prompts- List all prompts (optional category filter)search_prompts- Search prompts by query stringget_prompt- Get prompt template with parametersrender_prompt- Render a prompt with parametersvalidate_prompt- Validate prompt parametersget_prompt_categories- List all prompt categories
Skills (as Tools)
summarize_document- Read and summarize a fileanalyze_text- Read and analyze textsetup_project- Initialize a new projectdaily_briefing- Get daily productivity briefingproject_status- Get project status report- And 6 more...
Workflow Execution
execute_workflow- Execute a workflow by ID with variables
Original Tools
- File:
read_file,write_file,list_directory,search_files - Web:
fetch_url,scrape_html - Dev:
git_status,git_log,git_diff,system_info,get_time - AI:
summarize
Configuration
Command-line Arguments
node dist/index.js /path/to/workspace /home/user/documents
Environment Variables
GOOGLE_GENERATIVE_AI_API_KEY- Required for AI summarization features
Usage with Conductor
Add this MCP server to Conductor:
claude mcp add open-mcp -s user -- node /Users/sdluffy/conductor/workspaces/playground/san-jose/open-mcp/dist/index.js
Or add to your conductor.json:
{
"mcpServers": {
"open-mcp": {
"command": "node",
"args": ["/Users/sdluffy/conductor/workspaces/playground/san-jose/open-mcp/dist/index.js"]
}
}
}
Project Structure
open-mcp/
├── src/
│ ├── index.ts # Main entry point
│ ├── core/ # Core engines
│ │ ├── prompt-manager.ts # Prompt template management
│ │ ├── skill-executor.ts # Skill execution engine
│ │ ├── workflow-engine.ts # Workflow engine with conditionals
│ │ └── registry.ts # Central component registry
│ ├── prompts/ # Prompt templates (YAML)
│ │ ├── productivity/
│ │ ├── code/
│ │ └── research/
│ ├── skills/ # Skill definitions
│ │ ├── categories/
│ │ │ ├── document.ts
│ │ │ ├── development.ts
│ │ │ └── productivity.ts
│ │ └── skills.ts # Skill registry
│ ├── workflows/ # Workflow definitions
│ │ ├── definitions/
│ │ │ ├── daily-routine.ts
│ │ │ ├── code-review.ts
│ │ │ └── project-setup.ts
│ │ └── workflows.ts # Workflow registry
│ ├── tools/ # Original tools
│ ├── types/ # TypeScript definitions
│ └── utils/ # Utilities
└── prompts/ # YAML prompt templates
Forked Dependencies
We maintain forks of key dependencies for customization:
| Repository | Fork | Purpose |
|---|---|---|
| @modelcontextprotocol/typescript-sdk | ishuru/typescript-sdk | MCP SDK modifications |
| openai/openai-openapi | ishuru/openai-openapi | OpenAI API spec |
Contributing to Forks
- Make changes in your fork
- Open a PR to the upstream repository
- Reference the open-mcp issue you're solving
Development
# Watch mode
npm run dev
# Build
npm run build
# Run with output
npm run dev:full
Adding New Prompts
Create a YAML file in prompts/{category}/:
id: my_prompt
name: My Prompt
description: Description
category: productivity
template: |
Your template here with {{variables}}
parameters:
- name: variable
type: string
required: true
Adding New Skills
Create a skill in src/skills/categories/{category}.ts:
export const mySkill: Skill = {
id: "my_skill",
name: "My Skill",
description: "Description",
category: "my_category",
tools: [
{ tool: "tool_name", parameters: {...} }
],
inputSchema: { type: "object", properties: {...} },
outputSchema: { type: "object", properties: {...} }
};
Adding New Workflows
Create a workflow in src/workflows/definitions/{name}.ts:
export const myWorkflow: Workflow = {
id: "my_workflow",
name: "My Workflow",
description: "Description",
steps: [
{ id: "step1", type: "tool", name: "Step 1", config: {...} }
]
};
Architecture
┌─────────────────────────────────────────────────────────────────┐
│ MCP Client (Claude) │
└─────────────────────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────────────┐
│ MCP Server (stdio) │
│ ┌───────────────────────────────────────────────────────────┐ │
│ │ Tool Registry │ │
│ │ ┌───────────┐ ┌───────────┐ ┌─────────────────────┐ │ │
│ │ │ Prompts │ │ Skills │ │ Workflows │ │ │
│ │ │ (Resource)│ │ (Tools) │ │ (Tools) │ │ │
│ │ └─────┬─────┘ └─────┬─────┘ └──────────┬──────────┘ │ │
│ └────────┼─────────────┼──────────────────┼─────────────────┘ │
│ │ │ │ │
│ ┌────────┼─────────────┼──────────────────┼─────────────────┐ │
│ │ ▼ ▼ ▼ │ │
│ │ ┌─────────┐ ┌─────────────┐ ┌──────────────────┐ │ │
│ │ │ Prompt │ │ Skill │ │ Workflow │ │ │
│ │ │ Manager │ │ Executor │ │ Engine │ │ │
│ │ └────┬────┘ └──────┬──────┘ └────────┬─────────┘ │ │
│ │ │ │ │ │ │
│ │ ▼ ▼ ▼ │ │
│ │ ┌─────────────────────────────────────────────────┐ │ │
│ │ │ Existing Tools │ │ │
│ │ │ file-tools | web-tools | dev-tools | ai-tools │ │ │
│ │ └─────────────────────────────────────────────────┘ │ │
│ └───────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
Security
- Path validation with allowed directories whitelist
- Command injection prevention
- Timeout protection on HTTP requests
- Directory traversal attack prevention
License
MIT License - see LICENSE for details
Links
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。