Agileday MCP Server

A Model Context Protocol (MCP) server for Agileday.

This server bridges Agileday's competence and employee data with LLMs (like Claude Desktop or Cursor). It allows you to ask natural language questions about your organization's skills, find experts, and explore competence profiles.

Capabilities

This server implements the following tools:

• find_experts_with_skill
  • Query: "Who knows React?"
  • Action: Searches for employees who possess a specific skill and returns their proficiency levels.
• get_employee_competence_profile
  • Query: "What are Jane Doe's top skills?"
  • Action: Retrieves a full list of skills, proficiency levels, and motivation for a specific employee.
• list_organization_skills
  • Query: "What skills do we have in the organization?"
  • Action: Retrieves the full taxonomy of skills available in your Agileday workspace, grouped by category. This helps the AI understand exactly which terms to search for and avoids guessing.

Configuration

You need the following environment variables to run the server:

Variable	Description	Example
AGILEDAY_TENANT_ID	Your company slug (subdomain).	acme (for acme.agileday.io)
AGILEDAY_API_TOKEN	Your API Bearer token.	eyJhbGciOi...

Quick Start (Docker)

The easiest way to run the server is using the pre-built Docker image from the GitHub Container Registry.

1. Run with Claude Desktop

Add this configuration to your claude_desktop_config.json:

{
  "mcpServers": {
    "agileday": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e", "AGILEDAY_TENANT_ID=your_tenant_id",
        "-e", "AGILEDAY_API_TOKEN=your_api_token",
        "ghcr.io/eficode/mcp-agileday:latest"
      ]
    }
  }
}

Option 2: Docker

1. Clone the repository:

   git clone [https://github.com/yourusername/agileday-mcp-server.git](https://github.com/yourusername/agileday-mcp-server.git)
   cd agileday-mcp-server
   

2. Build the image:

   docker build -t agileday-mcp-server .
   

3. Configure Claude Desktop: Add the following to your claude_desktop_config.json (usually located at ~/Library/Application Support/Claude/ on macOS or %APPDATA%\Claude\ on Windows).

   {
     "mcpServers": {
       "agileday": {
         "command": "docker",
         "args": [
           "run",
           "-i",
           "--rm",
           "-e", "AGILEDAY_TENANT_ID=your_tenant_id",
           "-e", "AGILEDAY_API_TOKEN=your_api_token",
           "agileday-mcp-server"
         ]
       }
     }
   }
   

4. Restart Claude Desktop.

Option 3: Local Python

1. Clone the repository and install dependencies:

   git clone [https://github.com/yourusername/agileday-mcp-server.git](https://github.com/yourusername/agileday-mcp-server.git)
   cd agileday-mcp-server
   pip install -r requirements.txt
   

2. Configure Claude Desktop: Update your claude_desktop_config.json to point to the python script:

   {
     "mcpServers": {
       "agileday": {
         "command": "python3",
         "args": ["/absolute/path/to/agileday_server.py"],
         "env": {
           "AGILEDAY_TENANT_ID": "your_tenant_id",
           "AGILEDAY_API_TOKEN": "your_api_token"
         }
       }
     }
   }
   

Option 4: Run as service

To deploy this as a standalone service (e.g., for remote access or clients that support HTTP transport), use the http transport mode.

```bash
  docker run -d \
    -p 8000:8000 \
    -e AGILEDAY_TENANT_ID=your_tenant \
    -e AGILEDAY_API_TOKEN=your_token \
    ghcr.io/eficode/mcp-agileday:latest \
    --transport http
```

The server will be available at:

Universal Endpoint: http://localhost:8000/mcp (Supports both SSE stream and POST messages)

Health Check: http://localhost:8000/health

Example Prompts

Once connected, you can ask Claude questions like:

• "Who are the experts in Python at our company?"
• "Find me someone who knows Service Design and has a high proficiency."
• "What is Jane Doe's competence profile?"
• "Do we have enough people with Kubernetes skills?"

Security Note

• API Tokens: Never commit your AGILEDAY_API_TOKEN to version control. Always pass it via environment variables or your local config file.
• ReadOnly Access: It is recommended to use an API token with read-only permissions if possible, as this tool only fetches data.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/AmazingFeature).
3. Commit your changes (git commit -m 'Add some AmazingFeature').
4. Push to the branch (git push origin feature/AmazingFeature).
5. Open a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.