dotnet-todo-mcp 0.1.2

A todo MCP server built with FastMCP

dotnet-todo-mcp

A FastMCP server that exposes the dotnet-todo CLI (a Todo.txt port) as MCP tools, making your todo list accessible to AI assistants via the Model Context Protocol.

Prerequisites

Install the todo CLI (.NET global tool):

dotnet tool install -g dotnet-todo

Running the Server

Via uvx (from PyPI):

uvx dotnet-todo-mcp

Local development (from the project directory):

uv sync
uv run python server.py

Tools

Tool	Description	Parameters
list_todos	List pending todos, optionally filtered	terms: list[str] | None — filter terms (AND logic, prefix - to exclude)
create_todo	Create a new todo item	title: str, priority: str | None — letter A–Z
complete_todo	Mark a todo as done	item_number: int
delete_todo	Delete a todo	item_number: int
set_priority	Set or remove priority	item_number: int, priority: str | None
list_todos_by_priority	List todos sorted by priority	priorities: str | None (e.g. "A" or "A-C"), terms: list[str] | None
list_all_todos	List all todos including completed	terms: list[str] | None
list_contexts	List all @context tags	(none)
list_projects	List all +project tags	(none)

Todo item format: 02 (D) 2023-07-30 Task description @context +project

MCP Client Configuration

Add to your MCP client config (e.g. Claude Desktop claude_desktop_config.json):

{
  "mcpServers": {
    "todo": {
      "command": "uvx",
      "args": [
        "--from",
        "dotnet-todo-mcp",
        "todo-mcp"
      ]
    }
  }
}

Local development — point at your local clone:

{
  "mcpServers": {
    "todo": {
      "command": "uv",
      "args": ["run", "--project", "/path/to/dotnet-todo-mcp", "todo-mcp"]
    }
  }
}

This requires that you run uv sync first.