clicksprotocol/agent-treasury-benchmark

---
license: mit
task_categories:
- tabular-regression
- time-series-forecasting
tags:
- defi
- yield
- usdc
- base
- ai-agents
- treasury
- agent-economics
- x402
- aave
- morpho
- mcp
pretty_name: Agent Treasury Benchmark
size_categories:
- 1K<n<10K
language:
- en
---

# Agent Treasury Benchmark

Simulated treasury performance data for AI agents using Clicks Protocol on Base.

## What is this?

A benchmark dataset for evaluating **agent treasury strategies**: how AI agents manage idle USDC through automated yield protocols. Each row represents one agent's monthly treasury snapshot.

Use this to:
- Benchmark different yield allocation strategies (5-50% yield split)
- Model agent economics under varying DeFi conditions
- Train models to predict optimal yield percentages
- Evaluate referral network growth patterns
- Study x402 payment flow economics

## Dataset Structure

### agent_treasury_monthly.csv

| Column | Type | Description |
|--------|------|-------------|
| agent_id | string | Unique agent identifier (0x address) |
| month | int | Month number (1-12) |
| yield_protocol | string | Active protocol (aave_v3 or morpho) |
| apy_percent | float | Current APY at snapshot |
| total_payments_received | float | USDC received that month |
| liquid_balance | float | 80% portion, available for operations |
| yield_deposit | float | 20% portion, earning yield |
| yield_earned | float | Yield earned that month |
| protocol_fee | float | 2% fee on yield |
| net_yield | float | Yield after fee |
| cumulative_value | float | Total agent treasury value |
| referral_level | int | 0-3 (0 = no referrer) |
| referral_earnings | float | Fee share from downstream agents |
| yield_percentage | int | Agent's configured split (5-50%) |
| x402_payments | int | Number of x402-style machine payments |

### agent_profiles.csv

| Column | Type | Description |
|--------|------|-------------|
| agent_id | string | Unique agent identifier |
| agent_type | string | trading_bot, service_agent, payment_router, data_provider, infrastructure |
| registration_month | int | When the agent registered |
| referrer_id | string | Who referred this agent (null if none) |
| initial_deposit | float | First USDC deposit |
| yield_percentage | int | Configured yield split |

## Key Statistics

- 200 simulated agents over 12 months
- 5 agent types reflecting real-world use cases
- APY range: 2-15% (matching real Aave/Morpho rates on Base)
- 3-level referral chain simulation
- x402 payment frequency modeling

## How Clicks Protocol Works

1. Agent receives USDC payment
2. Auto-split: 80% liquid, 20% to DeFi yield
3. YieldRouter picks best APY (Aave V3 or Morpho Blue)
4. Withdraw anytime, no lockup
5. 2% fee on yield only, never on principal
6. 3-level referral: 40% / 20% / 10% of protocol fee

## Links

- [Website](https://clicksprotocol.xyz)
- [SDK](https://www.npmjs.com/package/@clicks-protocol/sdk) — `npm install @clicks-protocol/sdk`
- [MCP Server](https://www.npmjs.com/package/@clicks-protocol/mcp-server) — `npx @clicks-protocol/mcp-server`
- [GitHub](https://github.com/clicks-protocol)
- [Treasury Simulator](https://huggingface.co/spaces/clicksprotocol/treasury-simulator)
- [llms.txt](https://clicksprotocol.xyz/llms.txt)
- [OpenAPI](https://clicksprotocol.xyz/api/openapi.json)

## Citation

```bibtex
@dataset{clicks_protocol_treasury_benchmark_2026,
title={Agent Treasury Benchmark},
author={Clicks Protocol},
year={2026},
url={https://huggingface.co/datasets/clicksprotocol/agent-treasury-benchmark}
}
```