🧠 Prompt Library

🚧 Project Under Development - Evolving project, expect changes. Feedback welcome!

Prompt Library is your personal collection of AI agent prompts that transforms the way you interact with AI models. Store, organize, and execute sophisticated prompts through an intuitive CLI - supporting both Claude (Anthropic) and GPT (OpenAI) models.

📚 Table of Contents

• 🎯 Purpose & Features
  • Key Features
• ⚡ Quick Start
• 🛠️ How It Works
• 🖥️ CLI Usage
  • Command Overview
  • Global Options
  • 🔄 Switch AI Models
  • 🚀 Execute Prompts
  • 🔎 Managing Prompts
    • Executing Prompts with Files
    • Working with Fragments
    • Sync Options
    • Environment Management
• 📂 Prompt Showcase
  • Your AI Agent Collection
• 🚀 Getting Started
  • Step 1: Set Up Your Repository
  • Step 2: Configure API Access
  • Step 3: Install and Run CLI
  • Step 4: For GitHub Actions (Optional)
• 🧩 Prompt Composition with Fragments
  • Creating Fragments
• ⚙️ Customizing Metadata Generation
• 🤝 Contributing
• 📄 License

🎯 Purpose & Features

Treat AI prompts as reusable mini-applications that transform AI models into specialized agents. This approach promotes:

• Modularity: Compose complex prompts from reusable components
• Reusability: Apply successful prompts across multiple projects
• Shareability: Exchange proven prompts with your team

In today's rapidly evolving AI landscape, maintaining a personal prompt library is becoming essential for individuals and organizations alike.

Key Features

✅ Organized Prompt Storage - Categorized structure with rich metadata
✅ Multi-Model Support - Works with both Claude and GPT family models
✅ Intelligent CLI - Interactive prompt execution with variable support
✅ Fragment System - Build prompts from reusable components
✅ Automatic Metadata - GitHub Actions maintain documentation
✅ Dev-Friendly - TypeScript codebase with comprehensive testing

⚡ Quick Start

# Clone repository
git clone https://github.com/YOUR_USERNAME/prompt-library.git

# Install dependencies
cd prompt-library && npm install

# Build and install CLI globally
npm run build && npm install -g .

# Launch interactive prompt menu
prompt-library-cli

📋 Complete setup guide

See detailed instructions in Getting Started.

🛠️ How It Works

1. Create Prompts: Write markdown prompt files in the prompts directory
2. Generate Metadata: Commit to trigger automatic metadata generation
3. Organize Automatically: README files update with categorized prompt listings
4. Execute via CLI: Run prompts through the interactive CLI or command line

💡 Pro Tip: Working with GitHub Actions

Create and commit prompt.md files individually to allow GitHub Actions to generate corresponding metadata.yml files. Both files are required for CLI prompt usage.

🖥️ CLI Usage

The CLI offers a complete prompt management and execution solution:

Command Overview

Command	Description	Example
prompt-library-cli	Interactive menu	prompt-library-cli
model	Configure AI model settings	prompt-library-cli model
prompts	List and manage prompts	prompt-library-cli prompts --list
execute	Run a specific prompt	prompt-library-cli execute -p 74
fragments	List and view fragments	prompt-library-cli fragments
config	Manage CLI configuration	prompt-library-cli config
env	Manage environment variables	prompt-library-cli env
sync	Update your library	prompt-library-cli sync
settings	Manage CLI settings	prompt-library-cli settings
flush	Reset all data (preserves config)	prompt-library-cli flush

Global Options

Usage: prompt-library-cli [options] [command]

Options:
  -V, --version               Output the version number
  -e, --execute <id_or_name>  Execute a prompt by ID or name
  -l, --list                  List all available prompts
  -s, --search <keyword>      Search prompts by keyword
  -h, --help                  Display help for command

🔄 Switch AI Models

prompt-library-cli model

The streamlined model menu lets you:

• 🤖 Change AI provider: Switch between Anthropic (Claude) or OpenAI (GPT)
• 🧠 Change model: Select from available models for your provider
• 🔋 Change max tokens: Adjust token limits for your responses

🚀 Execute Prompts

# List available prompts (to see IDs and names)
prompt-library-cli prompts --list

# Execute using prompt ID (most reliable)
prompt-library-cli execute -p 80  # ID for Software Architect Visionary

# Execute using prompt name (partial matches work too)
prompt-library-cli execute -p "git_commit_message_agent"
prompt-library-cli execute -p "commit message"

# With variables
prompt-library-cli execute -p 74 --description "Add user authentication" --files "auth.ts,users.ts"
# Or by name
prompt-library-cli execute -p "commit" --description "Add user authentication" --files "auth.ts,users.ts"

# Use file inputs
prompt-library-cli execute -p 81 -fi code=./my-code.ts

# Preview mode
prompt-library-cli execute -p "translator" -i

Tip: While both prompt IDs and names are supported, using IDs is more reliable. Directory names follow the pattern {function}_{domain}_agent and are matched using fuzzy search.

🔎 Managing Prompts

# List all prompts with their IDs and categories
prompt-library-cli prompts --list

# List all prompt categories
prompt-library-cli prompts --categories

# Search prompts by keyword (title, description, category)
prompt-library-cli prompts --search "git"

# Show favorite prompts
prompt-library-cli prompts --favorites

# Show recently executed prompts
prompt-library-cli prompts --recent

📚 Advanced CLI Usage

Executing Prompts with Files

# Execute a prompt with file input
prompt-library-cli execute -p "translator" -fi input=./document.txt

# Execute with multiple file inputs
prompt-library-cli execute -p "code_review" -fi code=./src/app.js -fi tests=./tests/app.test.js

# Execute with both file inputs and regular variables
prompt-library-cli execute -p "document_generator" -fi template=./template.md --title "Project Overview" --author "Team"

Working with Fragments

# List all available fragments
prompt-library-cli fragments

# The interactive menu lets you browse and view fragment contents

Sync Options

# Fetch latest prompts from the default repository
prompt-library-cli sync

# Specify a custom repository URL
prompt-library-cli sync -u https://github.com/your-username/your-repo.git

# Force sync without confirmation
prompt-library-cli sync --force

Environment Management

# Set environment variables for API keys
prompt-library-cli env

# The interactive menu guides you through setting API keys and other environment variables

Note: For security, your API keys are stored securely and never logged or displayed in full.

📂 Prompt Showcase

Your AI Agent Collection

Create custom agents for any task with our pre-built prompt templates

Note: These prompts are fully customizable examples. Build your own collection based on your unique needs.

🔹 Coding

Prompt	Description
Git Branch Name Generator	Generates optimized git branch names based on project context and user requirements
Git Commit Message Agent	Generates precise and informative git commit messages following Conventional Commits specification
GitHub Issue Creator	Creates comprehensive and actionable GitHub issues based on provided project information
Software Architect Code Reviewer	Generates comprehensive pull requests with architectural analysis and optimization suggestions
Software Architect Specification Creator	Creates comprehensive software specification documents based on user requirements
Software Architect Visionary	Analyzes user requirements and creates comprehensive software specification documents
Software Development Expert Agent	Provides expert, adaptive assistance across all aspects of the software development lifecycle.

🔹 Content Creation

Prompt	Description
Documentation Specialist Agent	Generates revolutionary software documentation using advanced AI techniques and industry best practices

🔹 Healthcare

Prompt	Description
Health Optimization Agent	Generates personalized, adaptive health optimization plans based on comprehensive user data analysis
Psychological Support and Therapy Agent	Provides AI-driven psychological support and therapy through digital platforms

🔹 Problem Solving

Prompt	Description
Problem Solving AI Agent	Generates expert networks and strategies to solve complex problems and achieve goals

🔹 Prompt Engineering

Prompt	Description
AI Assistant Architect	Conceptualizes innovative and feasible AI assistant designs for various domains
Prompt Engineering God	Crafts divine-tier prompts to maximize AI potential while adhering to ethical standards

🔹 Translation

Prompt	Description
Universal Translator Agent	Translates between any languages, modes of expression, or conceptual frameworks

💡 Pro Tip: Check out the Prompt Engineering category for prompts that help you create high-quality prompts

🚀 Getting Started

Step 1: Set Up Your Repository

# Fork and clone the repository
git clone https://github.com/YOUR_USERNAME/prompt-library.git
cd prompt-library

# Install dependencies
npm install

Step 2: Configure API Access

Choose your preferred AI provider:

🔵 Anthropic Claude

1. Get an API key from the Anthropic Console
2. Set the environment variable: export ANTHROPIC_API_KEY=your_key_here

🟢 OpenAI GPT

1. Get an API key from the OpenAI Platform
2. Set the environment variable: export OPENAI_API_KEY=your_key_here

Step 3: Install and Run CLI

# Build TypeScript project
npm run build

# Install CLI globally
npm install -g .

# Launch interactive CLI
prompt-library-cli

# Configure your AI model
prompt-library-cli model

Note: On first run, you'll need to configure which AI model to use. The CLI will guide you through selecting a provider (Anthropic/OpenAI), model, and token settings.

Step 4: For GitHub Actions (Optional)

If you want automatic README generation, add your API key to repository secrets:

• Go to your GitHub repository → Settings → Secrets → New repository secret
• Add ANTHROPIC_API_KEY or OPENAI_API_KEY

🧩 Prompt Composition with Fragments

Fragments are modular, reusable prompt components that help you build complex prompts efficiently:

# Example prompt using fragments
I need you to help me with {{TASK_TYPE}}.

{{BEHAVIOR_GUIDELINES}}

Context:
{{USER_CONTEXT}}

Now, please {{SPECIFIC_INSTRUCTION}}.

Creating Fragments

1. Add .md files to the fragments/ directory (categorized by type)
2. Reference them in your prompt files with {{FRAGMENT_NAME}}
3. Browse and manage fragments with prompt-library-cli fragments

🔍 Fragment Examples

Check the prompt_engineering directory for example fragments including:

• Behavior attributes
• Formatting guidelines
• Safety guidelines
• Output templates

⚙️ Customizing Metadata Generation

The system automatically analyzes prompts and generates metadata. To customize this process:

1. Modify the system prompt at src/system_prompts/prompt_analysis_agent/prompt.md
2. Test your changes with npm run update-metadata
3. Commit to trigger automatic updates

⚠️ Important Notes

Changes to the metadata system affect all future prompt analyses. Test thoroughly before committing.

🤝 Contributing

Contributions are welcome! Here's how you can help:

• Add New Prompts: Share your best prompt designs
• Improve Templates: Enhance the prompt structure
• Fix Bugs: Help improve the CLI functionality
• Add Features: Extend the system's capabilities

Please submit issues or pull requests through GitHub.

📄 License

This project is licensed under the MIT License - feel free to use, modify, and distribute as needed.