Content
# MCP Shrimp Task Manager
[](https://smithery.ai/server/@cjo4m06/mcp-shrimp-task-manager)
> 🚀 An intelligent task management system based on Model Context Protocol (MCP), providing an efficient programming workflow framework for AI Agents.
The Shrimp Task Manager guides the Agent in systematically planning programming steps through structured workflow prompts, enhancing task memory management mechanisms, and effectively avoiding redundant and repetitive programming work.
## ✨ Features
- **Task Planning and Analysis**: In-depth understanding and analysis of complex task requirements
- **Intelligent Task Splitting**: Automatically breaks down large tasks into manageable smaller tasks
- **Dependency Management**: Precisely handles dependencies between tasks to ensure correct execution order
- **Execution Status Tracking**: Real-time monitoring of task execution progress and status
- **Task Integrity Verification**: Ensures task outcomes meet expected requirements
- **Task Complexity Assessment**: Automatically evaluates task complexity and provides optimal handling suggestions
- **Automatic Task Summary Updates**: Generates summaries upon task completion, optimizing memory efficiency
## 🔄 Task Management Workflow
This system provides a complete task management lifecycle:
1. **Start Planning** `plan_task`: Analyze task issues and define the scope of requirements
2. **In-depth Analysis** `analyze_task`: Review the existing codebase to avoid redundant work
3. **Solution Reflection** `reflect_task`: Critically review analysis results to ensure the solution is comprehensive
4. **Task Splitting** `split_tasks`: Decompose complex tasks into smaller tasks, establishing clear dependencies
5. **Task List** `list_tasks`: View all tasks and their execution statuses
6. **Execute Task** `execute_task`: Execute specific tasks while assessing complexity
7. **Result Verification** `verify_task`: Thoroughly check task completion status
8. **Task Completion** `complete_task`: Mark tasks as complete and generate reports, automatically updating summaries
9. **Task Management** `delete_task`: Manage incomplete tasks (completed tasks will be retained in the system)
## 📚 Documentation Resources
- [System Architecture](docs/architecture.md): Detailed system design and data flow explanation
## 🔧 Installation and Usage
### Installing via Smithery
To install Shrimp Task Manager for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@cjo4m06/mcp-shrimp-task-manager):
```bash
npx -y @smithery/cli install @cjo4m06/mcp-shrimp-task-manager --client claude
```
### Manual Installation
```bash
# Install dependencies
npm install
# Build and start the service
npm run build
```
## 🔌 Using with MCP-Supported Clients
The Shrimp Task Manager can be used with any client that supports Model Context Protocol, such as Cursor IDE.
### Configuring in Cursor IDE
The Shrimp Task Manager provides two configuration methods: global configuration and project-specific configuration.
#### Global Configuration
1. Open the global configuration file for Cursor IDE (usually located at `~/.cursor/mcp.json`)
2. Add the following configuration in the `mcpServers` section:
```json
{
"mcpServers": {
"shrimp-task-manager": {
"command": "node",
"args": ["/mcp-shrimp-task-manager/dist/index.js"],
"env": {
"DATA_DIR": "/mcp-shrimp-task-manager/data"
}
}
}
}
```
> ⚠️ Please replace `/mcp-shrimp-task-manager` with your actual path.
#### Project-Specific Configuration
You can also set up dedicated configurations for each project to use independent data directories:
1. Create a `.cursor` directory in the project root
2. Create a `mcp.json` file in that directory with the following content:
```json
{
"mcpServers": {
"shrimp-task-manager": {
"command": "node",
"args": ["/path/to/mcp-shrimp-task-manager/dist/index.js"],
"env": {
"DATA_DIR": "/path/to/project/data" // Must use absolute path
}
}
}
}
```
### ⚠️ Important Configuration Notes
The **DATA_DIR** parameter is the directory where the Shrimp Task Manager stores task data, conversation logs, and other information. Correctly setting this parameter is crucial for the normal operation of the system. This parameter must use an **absolute path**; using a relative path may cause the system to fail to locate the data directory correctly, leading to data loss or functionality issues.
> **Warning**: Using a relative path may lead to the following issues:
>
> - Data files cannot be found, causing system initialization failures
> - Task statuses may be lost or not saved correctly
> - Application behavior may be inconsistent across different environments
> - System crashes or fails to start
## 💡 System Prompt Guidance
### Cursor IDE Configuration
You can enable the Cursor Settings => Features => Custom modes feature and configure the following two modes:
#### TaskPlanner Mode
```
You are a professional task planning expert. You must interact with the user, analyze their needs, and gather project-related information, ultimately using mcp_shrimp_task_manager_plan_task to create tasks. Once the tasks are created, you must summarize and inform the user to use the task execution Model for task execution. You must focus on task planning and are prohibited from using mcp_shrimp_task_manager_execute_task to execute tasks. A serious warning: you are a task planning expert, and you cannot directly modify the code; you can only plan tasks.
```
#### TaskExecutor Mode
```
You are a professional task execution expert. When the user specifies a task to execute, use mcp_shrimp_task_manager_execute_task to perform the task. If there are no tasks to execute, use mcp_shrimp_task_manager_list_tasks to find tasks to execute. After execution, you must summarize and inform the user. You can only execute one task at a time, and unless the user explicitly states otherwise, you are prohibited from proceeding to the next task upon completion of a single task. If the user requests "continuous mode," execute all tasks in order.
```
> 💡 Choose the appropriate mode based on your needs:
>
> - Use **TaskPlanner** mode when planning tasks
> - Use **TaskExecutor** mode when executing tasks
### Using in Other Tools
If your tool does not support the Custom modes feature, you can:
- Manually paste the corresponding prompts at different stages
- Or directly use simple commands like `Please plan the following task: ......` or `Please start executing the task...`
## 🛠️ Available Tools Overview
Once configured, you can use the following tools:
| Function Category | Tool Name | Function Description |
| ----------------- | ---------------- | -------------------------- |
| **Task Planning** | `plan_task` | Start planning tasks |
| **Task Analysis** | `analyze_task` | In-depth analysis of task requirements |
| **Solution Evaluation** | `reflect_task` | Reflect and improve solution ideas |
| **Task Management** | `split_tasks` | Split tasks into subtasks |
| | `list_tasks` | Display all tasks and statuses |
| | `delete_task` | Delete incomplete tasks |
| **Task Execution** | `execute_task` | Execute specific tasks |
| | `verify_task` | Verify task completion status |
| | `complete_task` | Mark task as completed |
## 🔧 Technical Implementation
- **Node.js**: High-performance JavaScript runtime environment
- **TypeScript**: Provides a strongly typed development environment
- **MCP SDK**: Interface for seamless interaction with large language models
- **UUID**: Generates unique and reliable task identifiers
## 📄 License
This project is licensed under the MIT License.
