AI Prompts & Skills

The Ultimate Foundation for AI-Powered Creativity

A curated collection of expertly crafted prompts and agent skills for every AI use case

Why This Repository?

In the rapidly evolving world of AI, the right prompt can unlock extraordinary results. This repository serves as your definitive resource for:

High-quality prompts tested across multiple AI platforms
Agent Skills following the Agent Skills specification and discoverable via skills.sh
Best practices distilled from real-world usage
Foundation patterns applicable to any AI use case

Whether you’re generating stunning visuals, crafting compelling videos, synthesizing audio, or building intelligent agents — this is your toolkit.

What’s Inside

Image Generation

Prompts optimized for:

Nano Banana 2 - Fast, high-fidelity image generation
DALL-E 3 - Photorealistic & artistic outputs
Midjourney - Stylized creative imagery
Stable Diffusion - Fine-tuned control

prompts/images/

Video Creation

Master prompts for:

Veo - Google’s video AI
Runway Gen-3 - Cinematic sequences
Pika Labs - Dynamic motion
Sora - Long-form narratives

prompts/videos/

Audio & Voice

Prompts for:

Music generation - Original compositions
Voice synthesis - Natural speech patterns
Sound effects - Immersive audio design
Podcast content - Engaging scripts

prompts/audio/

Text & Content

Expert prompts for:

Creative writing - Stories, poetry, scripts
Technical documentation - Clear, concise guides
Marketing copy - Conversion-focused content
Research synthesis - Academic excellence

prompts/text/

Agentic Skills

Skills in this repository follow the Agent Skills Specification — an open format for giving AI agents new capabilities and expertise. Skills are discoverable and installable via the skills.sh registry.

What are Agent Skills?

Agent Skills are folders containing instructions, scripts, and resources that agents can discover and use to perform tasks more accurately and efficiently. Each skill is a self-contained directory with a SKILL.md file.

Available Skills

Skill	Description
`code-review`	Structured code review process with actionable feedback
`debug-and-fix`	Systematic debugging and root-cause analysis
`design-to-tasks`	Break a design (or PRD) into actionable engineering tasks
`idea-to-prd`	Create a Product Requirements Document from a feature idea
`next-step`	Report plan status and route to the right skill
`performance-review`	Measurement-driven whole-system performance sweep
`plan-retrospective`	Reflect on completed work and capture learnings
`prd-to-design`	Turn a PRD into a technical design and ADRs
`refactor`	Restructure code safely without changing behavior
`release-checklist`	Pre-release verification and deployment checklist
`security-review`	Lightweight threat model and security posture sweep
`tasks-to-code`	Implement engineering tasks from a task list
`ui-design-audit`	Audit UI components for design consistency

See the Skills README for complete documentation, recommended order of operations, and workflow details.

Skill Directory Structure

Each skill lives in its own directory inside skills/:

skills/
└── skill-name/
    ├── SKILL.md           # Required: main instructions + frontmatter
    ├── scripts/           # Optional: executable helper scripts
    ├── references/        # Optional: additional documentation
    └── assets/            # Optional: static resources

SKILL.md Format

Every SKILL.md must include YAML frontmatter with at minimum name and description:

---
name: skill-name
description: >
  What this skill does and when to use it. Include specific keywords
  that help agents identify when the skill is relevant.
license: MIT
metadata:
  author: pokanop
  version: "1.0"
---

# Skill Instructions

Detailed instructions and guidance for the agent...

Frontmatter Fields

Field	Required	Description
`name`	✅ Yes	Lowercase alphanumeric + hyphens only. Must match the directory name.
`description`	✅ Yes	What the skill does and when to use it (1–1024 chars).
`license`	No	License name or reference (e.g. `MIT`, `Apache-2.0`).
`compatibility`	No	Environment requirements, target agent products, or required tools.
`metadata`	No	A map of arbitrary string keys/values (e.g. `author`, `version`).
`allowed-tools`	No	List of tools the skill is permitted to use.

name rules: 1–64 characters, lowercase a-z and - only, must not start or end with -, no consecutive --. Must match the parent directory name exactly.

Adding a Skill

Create a directory in skills/ using lowercase letters and hyphens (e.g. skills/my-skill/)
Add a SKILL.md with valid frontmatter and instructions
Optionally include scripts/, references/, or assets/ subdirectories

Install any skill from this repo via the CLI:

bunx skills add pokanop/ai

Supported by: Claude Code, Cursor, Windsurf, OpenCode, GitHub Copilot, OpenHands, and more.

See the Agent Skills Specification for the complete format reference.

Repository Structure

ai/
├── assets/                 # SVGs, images, and visual resources
│   ├── logo.svg
│   ├── hero-banner.svg
│   ├── features-icons.svg
│   ├── icon-video.svg
│   ├── icon-audio.svg
│   ├── icon-agents.svg
│   └── icon-code.svg
├── prompts/
│   ├── images/            # Image generation prompts
│   ├── videos/            # Video creation prompts
│   ├── audio/             # Audio generation prompts
│   └── text/              # Text & content prompts
├── skills/                # Agent Skills (agentskills.io format)
│   ├── code-review/
│   ├── debug-and-fix/
│   ├── design-to-tasks/
│   ├── idea-to-prd/
│   ├── plan-retrospective/
│   ├── prd-to-design/
│   ├── release-checklist/
│   ├── tasks-to-code/
│   └── ui-design-audit/
└── README.md

Quick Start

Using Prompts

# Clone the repository
git clone https://github.com/pokanop/ai.git

# Navigate to your use case
cd ai/prompts/images

Installing Skills

Install all skills in this repository directly into your project:

bunx skills add pokanop/ai

Or browse and install individual skills from the skills.sh leaderboard.

The Possibilities

This repository is designed to be the foundation for all AI use cases:

For Creators

Generate consistent brand imagery
Create video content at scale
Produce original music and soundtracks
Write engaging content effortlessly

For Developers

Build intelligent agents with specialized skills
Automate repetitive coding tasks
Create AI-powered workflows
Integrate prompts into applications

For Businesses

Standardize AI interactions across teams
Maintain brand voice consistency
Scale content production
Reduce AI iteration time

For Researchers

Study prompt engineering patterns
Benchmark AI model performance
Contribute to collective knowledge
Share reproducible experiments

Prompt Engineering Best Practices

Structure of a Great Prompt

[CONTEXT] - Who/what is involved
[TASK] - What needs to be done
[CONSTRAINTS] - Limitations and requirements
[STYLE] - Tone, format, aesthetic
[EXAMPLES] - Reference outputs (optional)

Tips by Category

Category	Key Tips
Images	Be specific about style, lighting, composition
Videos	Describe motion, transitions, timing
Audio	Specify mood, tempo, instruments
Text	Define audience, tone, length
Agents	Set clear goals, constraints, success criteria

Contributing

We welcome contributions! Here’s how to help:

Add prompts - Share your best-performing prompts
Create skills - Build reusable agent capabilities following the specification
Improve docs - Enhance explanations and examples
Report issues - Found something that doesn’t work? Let us know

See CONTRIBUTING.md for detailed guidelines.

Roadmap

License

This project is licensed under the MIT License - see the LICENSE file for details.

Connect & Contribute

Made with care for the AI community

Star this repo if you find it useful!

Report Bug · Request Feature · Join Discussion

Building the future of AI, one prompt at a time

☕ Support This Project

If this project helped you, consider supporting its development: