In a nutshell
task-cli leverages an LLM to handle AI project management from the command line: automatically generating task lists, interactive Gantt charts, and real-time Q&A.

🔍 Use Cases

Scenario	Traditional Method	The task-cli Advantage
Requirements Analysis	Manually breaking down tasks	AI-powered decomposition of complex requirements into executable sub-tasks
Progress Tracking	Manual updates in Excel/Notion	One-click Gantt chart generation for real-time timeline visualization
Project Q&A	Searching through documents	Natural language conversation to quickly retrieve project information

✨ Core Features

• 🤖 AI Task Decomposition – Based on high-level objectives, utilizes an LLM to automatically break down complex tasks into executable sub-tasks. Supports multi-turn dialogue to continuously refine the generated task plan.
• 📊 Interactive Gantt Charts – Automatically generates HTML Gantt charts, making task dependencies clear and intuitively tracking project progress.
• 💬 Intelligent Project Q&A – Chat with your project database using natural language queries.
• 🔧 Rich CLI Commands – init, decompose, gantt, report cover the entire project lifecycle.
• 🌍 Language Support – The agent can be configured to respond in the user's preferred language.

---

🚀 Quick Start

1. Installation

# Clone the repository
git clone https://github.com/sunjiawe/task-cli.git
cd task-cli

# Create a virtual environment (optional)
python -m venv venv && source venv/bin/activate  # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

📖 Detailed Usage Guide

1. Initialize a Project

Use the init command to start a new project:

python main.py init
? Enter project name: Cross-platform To-Do App
? Enter project goal: All software features are fully functional

2. Start the Project Assistant

Before use, you need to configure the DEEPSEEK_API_KEY environment variable to use the LLM's API.

After initialization, run the program to enter the assistant terminal:

set DEEPSEEK_API_KEY="your-api-key"
python main.py

📖 Command Details

Type / to trigger subcommand autocompletion.

Command	Description	Example
/decompose <req>	AI Task Decomposition	/decompose Develop a cross-platform to-do app
/list	View the task list	/list
/howto <id>	Get guidance on a task	/howto design-ui
/update <id>	Update task status	/update design-ui done
/gantt	Generate a Gantt chart	/gantt
/report	Generate a project summary	/report
/qa <question>	Project Q&A	/qa What tasks are due this week?
/help	Display help	/help

Debug Mode: Set the environment variable DEBUG_MODE=true to view LLM request details.

---

🎯 Examples

1. Task Decomposition

> /decompose Develop a cross-platform to-do app
✅ Decomposed into 8 sub-tasks. View with /list

2. Generate Gantt Chart

> /gantt
📊 Gantt chart saved to gantt.html

Open gantt.html to see the interactive timeline:

3. Others

# Ask the assistant for guidance on executing a single task
/howto 

# Update task status
/update

# Project summary report
/report

# Ask any questions about the project in natural language
/qa

# Display help information
/help

---

🤝 Contributing

All forms of contribution are welcome!

• Found a Bug → Submit an Issue
• Feature Suggestion → Start a Discussion
• Code Contribution → Submit a Pull Request

---

📄 License

This project is licensed under the Apache 2.0 License. Apache 2.0 © 2025 sunjiawe

🙏 Acknowledgments

• Pocket Flow - The minimalist Agent framework that drives this project.
• gemini-cli - An open-source AI agent that brings the power of Gemini directly into your terminal.

---