Lovable Prompt Creator - AI Prompt Generator

## APPLICATION OVERVIEW
The project is a web application designed as a unified AI agent platform that enables users to execute local tasks and perform real-world actions through a chat-based interface. The application aims to provide an efficient AI operator that not only reasons about tasks but also executes them seamlessly, integrating various external services to enhance user productivity.

## CORE FEATURES
1. **Chat Interface**: A ChatGPT-style UI where users can input tasks such as “Send an email” or “Summarize my Slack messages” and receive streaming responses.
   
2. **Agent Execution Engine**: Parses user intent, generates a simple plan, and executes tasks via integrated tools like Gmail and Slack, returning the results to the user.

3. **Composio Integration Layer**: Allows users to connect their accounts through OAuth and see the tools they have integrated, facilitating seamless task execution.

4. **Task Persistence + Memory**: Stores conversations and task history, featuring basic memory capabilities to retrieve the last N messages for context.

5. **Observability Panel**: Displays detailed information for each task executed, including steps taken, tools used, inputs/outputs, and errors, with a collapsible panel for user convenience.

6. **Authentication**: Requires user authentication through Google or GitHub OAuth, limiting tool access to only connected integrations.

## DESIGN SPECIFICATIONS
- **Visual Style**: minimalist - Clean, simple design with plenty of white space, minimal color palette, and focus on typography
- **Color Mode**: Light theme with dark text on light backgrounds
- **Primary Color**: #1978E5 (accent for buttons, links, highlights)
- **Typography**: Use Inter from Google Fonts for headings, Inter for body text and UI elements
- **Border Radius**: 8px (moderately rounded) for buttons, cards, and inputs
- **Layout**: The main layout consists of a header with the chat interface at the center, a message list above the input box, and an observability panel below the chat area. The design should maintain a clear visual hierarchy, with space for loading states and responses.

## TECHNICAL REQUIREMENTS
- **Framework**: React with TypeScript
- **Styling**: Tailwind CSS
- **UI Components**: shadcn/ui
- **State Management**: Using React Context or a lightweight state management library like Zustand.

## IMPLEMENTATION STEPS
1. **Set Up the Project**: Initialize a new React project with TypeScript and install Tailwind CSS and shadcn/ui.
   
2. **Create the Chat Interface**: Develop the chat interface component with a message list and input box, ensuring it supports dark mode and loading states.

3. **Implement the Agent Execution Engine**: Build the logic to parse user input, generate execution plans, and integrate with Composio tools for task handling.

4. **Develop the Composio Integration Layer**: Create components for user authentication and account management, enabling users to connect their accounts via OAuth.

5. **Implement Task Persistence and Memory**: Set up Postgres for database management and Redis for task queue handling to store and retrieve task history and context.

6. **Create the Observability Panel**: Design and implement the panel that displays execution details for each task, ensuring it can be toggled by the user.

7. **Testing and Debugging**: Conduct thorough testing of each feature, focusing on user interactions and API integrations.

8. **Deployment**: Prepare the application for deployment, ensuring all necessary configurations for production are in place.

## USER EXPERIENCE
Users will interact with a clean and minimal chat interface where they can type in tasks. The system will process their requests and provide real-time feedback through streaming responses. The observability panel will enhance transparency by allowing users to see how their tasks are executed, making it easier to understand the AI's actions. Authentication will be seamless, offering users a secure way to connect their accounts and utilize the platform effectively.