An autonomous crypto trading agent that checks token risk before every trade — powered by Marcus Rug Intel via x402 payments.
No API keys. No subscriptions. Just pay-per-request with USDC on Base or Solana.
Agent sees token opportunity
↓
Pays $0.04 USDC via x402 → Marcus Rug Intel risk check
↓
Risk score 85+ → SKIP (saved from rug pull)
Risk score < 40 → TRADE (with confidence)
Risk score 40-85 → Deep analysis ($0.50) → decide
This is a reference implementation showing how any AI trading agent can integrate on-chain risk intelligence using the x402 payment protocol. Fork it, adapt it, ship it.
Traditional API integration requires: API keys, account creation, billing setup, rate limit management.
With x402: hit the endpoint, pay per request, get data. Your agent handles everything autonomously.
┌──────────────────────────────────────────────────┐
│ Safe Trading Agent │
│ │
│ ┌──────────┐ ┌──────────┐ ┌───────────────┐ │
│ │ Discovery │→│ Risk Gate │→│ Trade Decision │ │
│ │ (Bazaar) │ │ (x402) │ │ (LLM or rule) │ │
│ └──────────┘ └──────────┘ └───────────────┘ │
│ ↓ ↓ ↓ │
│ Find services Pay & check Buy / Skip / Deep │
└──────────────────────────────────────────────────┘
↕ ↕
CDP Bazaar Marcus Rug Intel
(discovery) (cryptorugmunch.app)
- Python 3.10+
- A wallet with USDC on Base (even $1 is enough for ~25 risk checks)
git clone https://github.com/CryptoRugMunch/x402-trading-agent.git
cd x402-trading-agent
pip install -r requirements.txt
cp .env.example .env
# Edit .env with your wallet private key# Check a single token before trading
python agent.py check So11111111111111111111111111111111111111112
# Run the autonomous trading loop (watches for new tokens, checks risk, decides)
python agent.py watch --chain solana
# Discover available Marcus Rug Intel endpoints via Bazaar
python agent.py discoverThe agent can discover Marcus Rug Intel endpoints dynamically via the x402 Bazaar:
from x402_client import discover_services
services = discover_services(category="crypto-security")
# Returns: check-risk, holder-deepdive, marcus-quick, etc.Every token check costs $0.04 USDC, paid automatically:
from x402_client import check_risk
result = check_risk("7GCihgDB8fe6KNjn2MYtkzZcRjQy3t9GHdC8uHYmW2hr")
# Returns: {risk_score: 85, recommendation: "AVOID", risk_factors: [...]}if result["risk_score"] >= 80:
print("🚨 HIGH RISK — skipping")
elif result["risk_score"] >= 40:
# Get deeper analysis for borderline cases
deep = marcus_forensics(token_address)
decide_from_analysis(deep)
else:
print("✅ LOW RISK — proceeding with trade")
execute_trade(token_address, amount)| Endpoint | Price | Description |
|---|---|---|
check-risk |
$0.04 | Risk score (0-100), honeypot detection, recommendation |
check-batch |
$0.30 | Batch scan up to 20 tokens |
holder-deepdive |
$0.10 | Sniper detection, bundle analysis, whale tracking |
deployer |
$0.06 | Deployer history, serial rugger detection |
token-intel |
$0.06 | Full token intelligence (price, volume, LP lock) |
marcus-quick |
$0.15 | AI forensic verdict (Claude Sonnet 4) |
marcus-forensics |
$0.50 | Full AI investigation |
marcus-ultra |
$2.00 | Deep analysis (Claude Opus 4) |
watch |
$0.20 | 7-day webhook monitoring for risk changes |
# .env
WALLET_PRIVATE_KEY=0x... # Your wallet private key (Base USDC)
RISK_THRESHOLD=70 # Skip tokens above this score
DEEP_ANALYSIS_THRESHOLD=40 # Trigger deep analysis for scores 40-70
MAX_SPEND_PER_CHECK=0.50 # Max USDC per individual check
CHAIN=solana # Default chain
MRI_API_URL=https://cryptorugmunch.app # API base URL| Strategy | Checks/Day | Daily Cost |
|---|---|---|
| Conservative (scan before every trade) | 50 | $2.00 |
| Active trader (scan + deep on flags) | 200 | $10-15 |
| Portfolio monitor (batch + watch) | 10 batches | $3.00 + $2.00 watches |
Compare: One rug pull can cost $1K-$100K+. Risk checking costs pennies.
- Add your own trading logic in
strategies/ - Swap the LLM — use Marcus AI tiers or bring your own
- Multi-chain — works with Solana, Ethereum, Base, Arbitrum, Polygon, Optimism, Avalanche
- Webhook alerts — use the
watchendpoint for passive monitoring
⚠️ Note: This project was formerly known as CryptoRugMunch. The $CRM token is not affiliated with Marcus Rug Intel.
MIT — use it however you want. If it saves you from a rug pull, consider it a win.
Built by Marcus Rug Intel — the only AI-powered crypto risk intelligence on x402.