Use with AI Agents

Connect AI coding assistants to the Staking Rewards API using a ready-made skill file.

Give your AI coding assistant full knowledge of the Staking Rewards API by adding a single skill file. The agent will be able to query staking data, fetch ratings, and help you write integration code — no manual API docs reading required.

What is SKILL.md?

A SKILL.md file is a structured reference document that teaches AI agents how to use an API. It contains endpoints, authentication details, query patterns, and common pitfalls — everything the agent needs to make correct API calls on your behalf.

Supported Tools

The skill file works with any AI tool that supports custom instructions or context files:

Claude Code — native skill support
Claude Desktop / Claude.ai — add as project knowledge
Cursor — add as project rules or context
Windsurf — add as project context
GitHub Copilot — add as repository instructions
ChatGPT — paste into custom instructions or conversation

Setup

Get your API key

If you don't have one yet, get an API key.

Set it as an environment variable:

export STAKING_REWARDS_API_KEY=your_key_here

Download the skill file

Download SKILL.md from the documentation site:

curl -o SKILL.md https://api-docs.stakingrewards.com/SKILL.md

Add to your AI tool

Install as a skill so Claude Code loads it on demand when you ask about staking data:

# Create skill directory and install
mkdir -p ~/.claude/skills/staking-rewards-api
cp SKILL.md ~/.claude/skills/staking-rewards-api/SKILL.md

Alternatively, to scope it to a single project, rename it to AGENTS.md in your project root — Claude Code auto-loads CLAUDE.md and AGENTS.md files but not SKILL.md.

Add the skill as a Cursor rule:

# Add as a project rule
mkdir -p .cursor/rules
cp SKILL.md .cursor/rules/staking-rewards-api.md

Alternatively, paste the file contents under Settings > Rules for AI.

Open your project in Claude and add SKILL.md as project knowledge:

Open project settings
Click Add content
Upload SKILL.md

The skill will be available in all conversations within that project.

For GitHub Copilot, add as repository instructions:

mkdir -p .github
cp SKILL.md .github/copilot-instructions.md

For Windsurf, ChatGPT, or other tools that support custom context:

Copy the contents of SKILL.md
Add it to your tool's context, instructions, or rules configuration
The agent will use it as a reference when you ask about staking data

Start querying

Ask your AI assistant natural language questions:

"Get the current reward rate for Ethereum"
"Show me the top 10 staking providers by AUM"
"What are the best-rated DeFi protocols?"
"Fetch historical staking data for Cosmos over the last 30 days"
"Compare validator commission rates for Polkadot"

The agent will construct the correct API calls, handle authentication, and format the results.

What the Agent Can Do

With the skill file, your AI agent knows how to:

Query staking data — assets, providers, validators, reward options, and metrics via GraphQL
Fetch ratings — DeFi and infrastructure ratings via REST endpoints
Filter and sort — by metric values, date ranges, asset types, and more
Handle historical data — with date filters and interval aggregation
Write integration code — generate JavaScript, Python, or any language for your use case
Debug errors — recognizes common error codes and suggests fixes

Example Session

You: What's the current reward rate for ETH?

Agent: Let me query the Staking Rewards API for that.

The current ETH reward rate is 3.42%.

Data from Staking Rewards.

You: Show me the top 5 DeFi protocols by rating

Agent: Here are the top-rated DeFi protocols:

| Protocol          | Provider    | Rating | TVL      | APY   |
|-------------------|-------------|--------|----------|-------|
| Aave v3 WETH      | Aave        | A      | $5.99B   | 2.18% |
| Rocket Pool rETH  | Rocket Pool | BB     | $813M    | 2.13% |
| ...               | ...         | ...    | ...      | ...   |

Need help? Contact us at [email protected] or via Telegram @berlincrypto.

On this page