Getting Started with AI Agents

What Is an AI Agent?

An AI agent is a set of instructions — typically written in Markdown — that tells an AI assistant how to behave for a specific task. Instead of typing the same context and instructions every time you start a conversation, you load an agent file once and the AI knows exactly what role to play.

Think of it like a job description for your AI: "You are a senior React developer reviewing pull requests. Focus on performance, accessibility, and maintainability. Ask clarifying questions before suggesting changes."

Why Markdown?

The .md format was chosen because:

• Universal compatibility — GitHub Copilot, Cursor, Claude Projects, and ChatGPT all support Markdown instructions
• Human-readable — You can review and understand exactly what the agent does
• Version-controllable — Store agents in your repo alongside your code
• Portable — Share agents with teammates via Git, Slack, or AgentHut

Your First Download

1. Browse the marketplace at AgentHut.ai/browse
2. Find an agent that matches your use case (e.g., "React Code Reviewer" or "Python Test Generator")
3. Click the agent card to see a full preview of its instructions
4. Click Download Agent — you'll get a .md file

Loading the Agent in Cursor

1. Open your project in Cursor
2. Go to Settings → Rules for AI (or press Cmd+Shift+P → "Open Cursor Rules")
3. Paste the agent's content or reference the file path
4. Start a new chat — Cursor will now follow the agent's instructions

Loading in GitHub Copilot

1. Place the .md file in your repo as .github/copilot-instructions.md
2. Copilot will automatically pick it up for workspace-scoped suggestions
3. Or paste the content directly into a Copilot Chat conversation

Tips for Getting the Most Out of Agents

• Be specific when you select an agent — a "Python Pytest Generator" will be much more useful than a generic "code helper"
• Combine agents — load a code-review agent AND a documentation agent for a comprehensive workflow
• Iterate — if an agent doesn't quite fit your stack, download it and customize the Markdown to match your specific framework or conventions
• Share back — improved agents can be uploaded to AgentHut so the whole community benefits

---

Ready to explore? Browse the full marketplace →