> ## Documentation Index
> Fetch the complete documentation index at: https://yieldxyz.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Skills

> Use Yield.xyz agent skills with AI coding tools like Cursor, Claude Code, Codex, and Gemini CLI

## Overview

Yield.xyz provides a set of **agent skills** following the [Agent Skills open standard](https://agentskills.io/specification) that enable AI agents to autonomously interact with the Yield.xyz API. These skills are structured, machine-readable capability files that tell AI agents exactly how to discover yields, manage positions, and track portfolios.

<Info>
  Agent skills work with any AI tool that supports the agentskills.io specification, including **Cursor**, **Claude Code**, **OpenAI Codex**, **Gemini CLI**, and others.
</Info>

***

## Quick Start

### Install via Skills CLI

Add Yield.xyz skills to your AI agent's context:

```bash theme={null}
npx skills add https://docs.yield.xyz
```

### Manual Setup

You can also reference the skill files directly. Each skill is a directory containing a `SKILL.md` file:

```
skills/
├── find-yields/
│   └── SKILL.md
├── get-yield-info/
│   └── SKILL.md
├── enter-position/
│   └── SKILL.md
├── exit-position/
│   └── SKILL.md
├── manage-position/
│   └── SKILL.md
└── check-portfolio/
    └── SKILL.md
```

***

## Available Skills

<CardGroup cols={2}>
  <Card title="find-yields" icon="magnifying-glass">
    Search and discover yield opportunities across 2,900+ options on 80+ networks. Filter by network, token, provider, type, or APY.
  </Card>

  <Card title="get-yield-info" icon="circle-info">
    Retrieve detailed metadata, APY/APR rates, token information, validator data, and protocol details for a specific yield.
  </Card>

  <Card title="enter-position" icon="arrow-right-to-bracket">
    Enter (deposit into) a yield position by generating unsigned transactions. Covers staking, lending, vaults, and all yield types.
  </Card>

  <Card title="exit-position" icon="arrow-right-from-bracket">
    Exit (withdraw from) a yield position. Handles unstaking, withdrawals, cooldown periods, and partial exits.
  </Card>

  <Card title="manage-position" icon="gear">
    Manage existing positions: claim rewards, restake, claim unstaked tokens, rebond, vote, and other lifecycle operations.
  </Card>

  <Card title="check-portfolio" icon="wallet">
    View yield portfolio balances and positions across multiple networks and protocols. Track rewards and unstaking progress.
  </Card>
</CardGroup>

***

## Skill Format

Each skill follows the [agentskills.io specification](https://agentskills.io/specification) with:

* **YAML frontmatter** containing `name`, `description`, and metadata
* **Markdown body** with step-by-step instructions, API endpoints, request/response examples, and error handling guidance

```yaml theme={null}
---
name: find-yields
description: Search and discover yield opportunities...
license: MIT
metadata:
  author: yield-xyz
  version: "1.0"
---
```

***

## How It Works

When an AI agent encounters a task related to Yield.xyz (e.g., "Find the best ETH staking yields"), it:

<Steps>
  <Step title="Skill detection">
    The agent reads the skill's `name` and `description` to determine if the skill is relevant to the user's request.
  </Step>

  <Step title="Skill activation">
    The agent loads the full `SKILL.md` content to get detailed instructions.
  </Step>

  <Step title="API execution">
    The agent follows the step-by-step instructions to make API calls, handle pagination, and process responses.
  </Step>

  <Step title="Result presentation">
    The agent formats and presents the results to the user.
  </Step>
</Steps>

***

## Example Prompts

Once skills are installed, try these prompts with your AI coding tool:

<AccordionGroup>
  <Accordion title="Discovery">
    * "What are the best yields for ETH right now?"
    * "Show me all staking options on Solana"
    * "Find USDC lending yields with the highest APY"
    * "Compare Aave vs Compound yields on Ethereum"
  </Accordion>

  <Accordion title="Position management">
    * "Stake 1 ETH with Lido"
    * "Deposit 5000 USDC into Aave on Arbitrum"
    * "Unstake my ATOM from Cosmos"
    * "Claim my staking rewards"
  </Accordion>

  <Accordion title="Portfolio">
    * "Show my yield portfolio across all networks"
    * "Check my staking balance on Ethereum"
    * "Do I have any rewards to claim?"
    * "What's the status of my DOT unstaking?"
  </Accordion>
</AccordionGroup>

***

## Prerequisites

To use these skills, you need:

1. **A Yield.xyz API key** — sign up at [dashboard.yield.xyz](https://dashboard.yield.xyz/sign-up/register-interest)
2. **An AI tool** that supports agent skills (Cursor, Claude Code, Codex, Gemini CLI)
3. **The API key set as an environment variable:**

```bash theme={null}
export YIELD_API_KEY=your_api_key_here
```

***

## Using with Cursor

If you're using Cursor IDE, you can reference the skills directly in your project:

1. Add the skills directory to your project
2. Reference skills in your Cursor rules (`.cursor/rules/`)
3. The agent will automatically detect and use relevant skills

See [Cursor Setup](/guides/ai/cursor-setup) for detailed configuration.

***

## Using with LLMs

For direct LLM integration, you can also use:

* **`/llms.txt`** — Directory of all documentation pages
* **`/llms-full.txt`** — Complete documentation content for full context
* **`/skill.md`** — Root capability file describing all skills

See [Using Docs with LLMs](/guides/ai/using-docs-with-llms) for more details.

***

## Next Steps

<CardGroup cols={2}>
  <Card title="Cursor Setup" icon="code" href="/guides/ai/cursor-setup">
    Set up Cursor IDE for Yield.xyz development
  </Card>

  <Card title="Prompting Patterns" icon="message" href="/guides/ai/prompting-patterns">
    Effective prompts for Yield.xyz with AI
  </Card>

  <Card title="API Reference" icon="book" href="/api-reference/getting-started">
    Full API documentation
  </Card>

  <Card title="Guardrails" icon="shield" href="/guides/ai/guardrails">
    Safety considerations for AI-assisted development
  </Card>
</CardGroup>