╔═══════════════════════════════════════════════════════════════╗ ║ ║ ║ ███████╗██╗ ██╗██████╗ ███████╗██████╗ ║ ║ ██╔════╝██║ ██║██╔══██╗██╔════╝██╔══██╗ ║ ║ ███████╗██║ ██║██████╔╝█████╗ ██████╔╝ ║ ║ ╚════██║██║ ██║██╔══██╗██╔══╝ ██╔══██╗ ║ ║ ███████║╚██████╔╝██████╔╝███████╗██║ ██║ ║ ║ ╚══════╝ ╚═════╝ ╚═════╝ ╚══════╝╚═╝ ╚═╝ ║ ║ ║ ║ ██╗ ██████╗ ██████╗ ███████╗ ║ ║ ██║██╔═══██╗██╔══██╗██╔════╝ ║ ║ ██║██║ ██║██████╔╝███████╗ ║ ║ ██║██║ ██║██╔══██╗╚════██║ ║ ║ ██║╚██████╔╝██████╔╝███████║ ║ ║ ╚═╝ ╚═════╝ ╚═════╝ ╚══════╝ ║ ║ ║ ╚═══════════════════════════════════════════════════════════════╝
A curated collection of system prompts and configurations designed to enhance AI agent behavior. Each skill addresses specific failure modes:
| Problem | Solution |
|---|---|
| ❌ Instruction drift | ✅ Spec-lock protocol |
| ❌ Hallucination | ✅ Explicit assumptions |
| ❌ Inconsistent output | ✅ Deterministic patterns |
| ❌ Scope collapse | ✅ Step-by-step checkpoints |
| ❌ Over-refactoring | ✅ Minimal patch discipline |
git clone https://github.com/aaronaalmendarez/agentskills.git
cd agentskills
cat gemini/GEMINI.md # Load as system prompt|
🎯 Precise Prompts
|
📦 Modular Design
|
|
🔄 Regular Updates
|
📚 Well Documented
|
agentskills/
├── gemini/ 🪐 # Gemini-specific prompts
├── claude/ 🧠 # Claude-specific prompts
├── openai/ 🔵 # OpenAI-specific prompts
└── README.md 📄 # This file
# Clone the repository
git clone https://github.com/aaronaalmendarez/agentskills.git
# Navigate into the directory
cd agentskillsLoad any skill file as a system prompt when working with compatible AI models.
# Example: Load a skill
cat gemini/GEMINI.md| Metric | Value |
|---|---|
| Platforms | 3+ |
| Prompts | Growing |
| License | MIT |
| Status | Active |
- Initial repository structure
- Gemini prompts
- Claude prompts
- OpenAI prompts
- Custom configuration examples
- Integration guides
What are agent skills?
System prompts engineered to improve AI behavior and reduce common failure modes.
How do I use these?
Load the skill file as a system prompt when interacting with compatible AI models.
Can I contribute?
Absolutely! Feel free to submit issues and pull requests.
Is this free?
Yes, licensed under MIT.
Feel free to submit issues and pull requests.
Be respectful. Be constructive. Be kind.
Thank you to everyone who stars and contributes to this project!
MIT © 2026 aaronaalmendarez