cor-mcp
Enables AI assistants to manage COR project management platform resources like projects, tasks, hours, clients, teams, and users through natural language.
README
cor-mcp
MCP (Model Context Protocol) server for the COR project management platform.
Exposes COR's REST API as MCP tools so any compatible AI assistant (Claude Desktop, Claude Code, Cursor, Windsurf, etc.) can manage projects, tasks, hours, clients, teams, and users using natural language.
Setup
1. Install dependencies
npm install
2. Configure authentication
Copy the example env file:
cp env.example .env
Option A — Client credentials (recommended for CI/automation):
Edit .env and add your API key and client secret:
COR_API_KEY=your_api_key_here
COR_CLIENT_SECRET=your_client_secret_here
Get these from COR: Settings → Integrations → API. The server will automatically obtain a Bearer token on startup using the OAuth2 client_credentials flow and refresh it when it expires.
Option B — Interactive login:
Leave COR_API_KEY / COR_CLIENT_SECRET unset. After connecting the MCP server, call cor_login with your email and password:
cor_login(email: "you@company.com", password: "...")
The access token is stored in memory for the current session. You'll need to call cor_login again each time the server restarts.
3. Build
npm run build
4. Configure your MCP client
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"cor": {
"command": "node",
"args": ["/absolute/path/to/cor-mcp/dist/index.js"]
}
}
}
Claude Code
claude mcp add cor -- node /absolute/path/to/cor-mcp/dist/index.js
Cursor / Windsurf / VS Code (Cline)
Add to your MCP config:
{
"mcpServers": {
"cor": {
"command": "node",
"args": ["/absolute/path/to/cor-mcp/dist/index.js"]
}
}
}
Available Tools
Authentication
| Tool | Description |
|---|---|
cor_login |
Authenticate with email + password (stores token for the session) |
Projects
| Tool | Description |
|---|---|
cor_list_projects |
List projects (paginated, optional name search) |
cor_get_project |
Get project details by ID |
cor_create_project |
Create a new project |
cor_update_project |
Update an existing project |
Tasks
| Tool | Description |
|---|---|
cor_list_tasks |
List tasks (paginated, optional project filter) |
cor_get_task |
Get task details by ID |
cor_create_task |
Create a task within a project |
cor_update_task |
Update an existing task |
Hours
| Tool | Description |
|---|---|
cor_list_hours |
List logged hours (optional date filter) |
cor_log_hours |
Log worked hours to a task |
Clients
| Tool | Description |
|---|---|
cor_list_clients |
List clients (optional name search) |
cor_get_client |
Get client details by ID |
Teams
| Tool | Description |
|---|---|
cor_list_teams |
List all teams |
cor_get_team |
Get team details by ID |
Users
| Tool | Description |
|---|---|
cor_list_users |
List users (optional name search) |
cor_get_user |
Get user profile by ID |
Reference Values
Task status: nueva · en_proceso · estancada · finalizada
Task priority: 0 = Low · 1 = Medium · 2 = High · 3 = Urgent
Project status: in_process · finished · suspended
Project health: 1 = On track · 2 = At risk · 3 = Delayed · 4 = Critical
Development
Run without building (uses tsx):
npm run dev
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。