npx agentspec-cli install ba38bb8e-59d2-46f9-9a54-eeb9e18f0b79 --file-type cursorrules→ .cursor/rules/
# Project Instructions
Use these guidelines to build the Prompt Builder app efficiently.
## Overview
Prompt Builder is a single-page web app for managing prompts. It allows users to add, edit, and delete prompts with tags, select multiple prompts to create a "Master Prompt," and copy it to the clipboard.
## Tech Stack
- **Frontend:**
- React.js (via `create-react-app`)
- Chakra UI for components and styling
- DND Kit for drag-and-drop functionality
- Emotion for styled components
- Framer Motion for animations
- **Backend:**
- Node.js with Express.js
- SQLite3 for database
- CORS for cross-origin requests
- Nodemon for development
## Testing Stack
- **Frontend:**
- Jest
- React Testing Library
- **Backend:**
- Jest
- Supertest for API testing
- **Integration Tests:**
- Cypress for end-to-end testing
## Project Structure
- **Frontend:**
- `src/components/` - UI components (e.g., `PromptList`, `PromptEditor`, `MasterPrompt`)
- `src/App.js` - Main app logic and layout
- **Backend:**
- `server/routes/prompts.js` - API routes for prompts
- `server/db.js` - Database connection and queries
- `server/server.js` - Server setup
## Rules
### General Rules
- Leverage AI coding assistants (e.g., GitHub Copilot, Cursor) to accelerate development.
- Focus on core functionality; defer non-essential features (e.g., tag filtering, dark mode).
- Perform manual testing; automated tests are optional for the MVP.
- Use JavaScript for speed; TypeScript is optional.
#### Env Rules
- Use `.env` files for configuration:
- Frontend: Use `REACT_APP_` prefix (e.g., `REACT_APP_API_URL`).
- Backend: Store database paths or API settings if needed.
- Do not hardcode sensitive values like API URLs.
### Frontend Rules
- **Components:**
- Use Chakra UI components for consistent styling
- Implement drag-and-drop with @dnd-kit libraries
- Use Emotion/styled-components for custom styling when needed
- `PromptList`: Display prompts with checkboxes for selection
- `PromptEditor`: Form to add or edit prompts (text and tags)
- `MasterPrompt`: Show combined text of selected prompts with a copy button
- **State Management:**
- Use `useState` for component state (e.g., selected prompts)
- Use `useEffect` to fetch prompts on mount
- Optional: Use Context API for sharing prompts data across components
- **API Interactions:**
- Use `fetch` to call backend API endpoints.
- Handle basic loading states; skip extensive error handling for MVP.
- **UI:**
- Use Chakra UI's responsive layout components
- Implement animations with Framer Motion where appropriate
- Use toast notifications from @chakra-ui/toast for user feedback
### Backend Rules
- **Database:**
- Use SQLite with a `prompts` table:
- `id`: Integer (auto-incrementing primary key)
- `content`: Text (prompt text)
- `tags`: Text (comma-separated, e.g., "coding, AI")
- `created_at`: Timestamp (default to current time)
- **API Endpoints:**
- `POST /prompts`: Create a new prompt.
- `GET /prompts`: Retrieve all prompts (ordered by `created_at`, newest first).
- `PUT /prompts/:id`: Update a prompt by ID.
- `DELETE /prompts/:id`: Delete a prompt by ID.
- **Implementation:**
- Use Express.js for RESTful routing.
- Use `sqlite3` or `better-sqlite3` for database operations.
- Enable CORS if frontend and backend domains differ.
- Keep error handling simple (e.g., basic status codes and messages).
### Deployment Rules
Currently not deploying. App is for personal use only.
<!-- - **Frontend:**
- Deploy to Vercel or Netlify.
- Set `REACT_APP_API_URL` in `.env` to the backend API base URL.
- **Backend:**
- Deploy to Heroku.
- Note: SQLite data resets on Heroku dyno restarts; accept this for MVP or use a hosted database for persistence. -->
Blog Front Configuration
Cursor IDE rules and conventions for AI-assisted development.
by AgentHub Crawler
prompt-builder
Guide users through creating high-quality GitHub Copilot prompts with proper structure, tools, and best practices.
by AgentHub Crawler
PromptPack Project Overview
This repository contains the ecosystem for **PromptPack**, a tool designed to intelligently "pack" codebases into concise contexts for Large Language Models (LLMs). It features a multi-platform architecture sharing a common core.
by AgentHub Crawler
Sign in to leave a comment.
Project Rules
This project is a prompt manager where I store all my favorite prompts for easy usage.
by AgentHub Crawler