LLM Prompt Vault

A repository for storing and managing LLM prompts for various tasks.

Overview

This repository serves as a centralized collection of prompts for different Large Language Models (LLMs). It is organized to efficiently manage both model-specific prompts and task-oriented prompts that can be used across multiple LLM platforms.

Repository Structure

llm-prompt-vault/
├── models/              // Model-specific prompts
│   ├── claude/
│   │   ├── system-prompts/  // System prompts for Claude
│   │   └── templates/       // Claude-specific templates
│   ├── gemini/
│   │   ├── system-prompts/  // System prompts for Gemini
│   │   └── templates/       // Gemini-specific templates
│   ├── chatgpt/
│   │   ├── system-prompts/  // System prompts for ChatGPT
│   │   └── templates/       // ChatGPT-specific templates
│   └── roo-code/
│       ├── custom-instructions/ // Custom instructions for Roo Code
│       └── templates/           // RooCode-specific templates
│
└── tasks/               // Task-specific prompts (usable across models)
    ├── writing/         // Content writing prompts
    ├── coding/          // Programming assistance prompts
    ├── data-analysis/   // Data analysis prompts
    ├── translation/     // Translation prompts
    ├── prompt-engineering/ // Prompt engineering prompts
    └── summarization/   // Summarization prompts

Usage Guidelines

Model-Specific Prompts

The models directory contains prompts that are optimized for specific LLM platforms:

• system-prompts: Contains prompts that define the basic behavior and capabilities of the model
• templates: Contains model-specific templates optimized for that particular LLM
• custom-instructions (Roo Code only): Contains custom instructions for the Roo Code VSCode extension

Task-Specific Prompts

The tasks directory organizes prompts by their intended use case, regardless of which LLM they are used with:

• writing: Prompts for creative writing, content creation, etc.
• coding: Prompts for programming assistance and code generation
• data-analysis: Prompts for analyzing and interpreting data
• translation: Prompts for translating content between languages
• summarization: Prompts for summarizing content

Contributing

To add a new prompt:

1. Identify whether it's model-specific or task-specific
2. Place it in the appropriate directory
3. Use clear, descriptive filenames (e.g., code-review-prompt.md, creative-story-generator.md)
4. Include comments within the prompt file explaining its purpose and any specific instructions

Supported LLM Platforms

• Claude by Anthropic
• Gemini by Google
• ChatGPT by OpenAI
• Roo Code (prev. Roo Cline) (VSCode extension)

License

MIT License