@1247/yield Optimizer

Yield Optimizer

Autonomous USDC yield optimization across DeFi lending protocols and chains.

What This Does

Scans real-time yields from DeFi Llama, ranks opportunities by risk-adjusted return, recommends allocations, and can execute deposits/withdrawals via the wallet skill.

Data source: DeFi Llama Yields API (free, no key, real-time).

Supported Protocols & Chains

Stablecoins tracked: USDC, USDT, DAI, USDS, sDAI, sUSDe, GHO

Core Workflows

1. Scan & Report

When user asks "what are the best yields?" or "where should I put my USDC?":

1. Run python3 skills/yield-optimizer/scripts/scan_pools.py
2. Present top opportunities grouped by risk tier
3. Highlight the best rate per chain and the overall best

Risk tiers:

• Tier 1 (Safe): Aave V3, Compound V3, Spark — established, audited, >$100M TVL
• Tier 2 (Moderate): Morpho, Euler, Fluid — newer but audited, >$10M TVL
• Tier 3 (Aggressive): Pendle, smaller pools — higher yields, more complexity

2. Recommend Allocation

When user has a specific amount to deploy:

1. Scan current rates
2. Filter by user's risk preference (default: Tier 1-2)
3. Suggest allocation:
   • Conservative: 100% best Tier 1 pool
   • Balanced: 70% Tier 1 / 30% Tier 2
   • Aggressive: 50% Tier 1 / 30% Tier 2 / 20% Tier 3
4. Show expected annual yield in dollar terms

3. Execute Deposit

When user confirms a recommendation:

1. Check wallet balances via wallet skill (wallet_balance on target chain)
2. If USDC is on wrong chain, suggest bridging (across-bridge skill)
3. For Aave V3 deposits, use the contract addresses from references/contracts.json
4. Execute: approve USDC → supply to Aave pool
5. Verify position via aToken balance read

Prerequisites — Wallet Policy: Before any on-chain execution, load the wallet-policy skill and propose the standard wildcard policy.

4. Monitor & Rebalance

For scheduled autonomous operation:

1. Scan yields every hour
2. Compare current position APY vs best available
3. If delta > 0.5% AND gas cost < 0.1% of position — recommend rebalance
4. Log all decisions (hold/rebalance) with reasoning

Schedule via: schedule_task(command="python3 skills/yield-optimizer/scripts/scan_pools.py --json", schedule="every 1 hour")

Decision Logic

scan pools → filter (TVL > $100K, supported protocols) → rank by APY

IF no current position:
  → recommend deposit to best risk-adjusted pool

IF current position exists:
  apy_delta = best_available - current_apy
  IF apy_delta < 0.5%: HOLD (not worth gas)
  IF gas_cost > 0.1% of position: HOLD (too expensive)  
  ELSE: REBALANCE (withdraw → bridge if needed → deposit)

Key Gotchas

• APY ≠ APR. DeFi Llama returns APY (compounded). Don't double-compound when projecting.
• Reward APY is temporary. apyReward (token incentives) can vanish overnight. Weight apyBase higher for stability.
• TVL matters. High APY + low TVL = likely unsustainable or risky. Filter >$100K minimum, prefer >$10M.
• Gas costs eat small positions. Ethereum mainnet Aave deposit ≈ $5-15. A $500 position rebalancing for 0.3% more APY loses money.
• Bridge time. Cross-chain moves take 2-15 minutes. Don't show as instant.
• Pendle yields are structured. They represent fixed-rate or leveraged yield — not simple lending. Flag this to users.

Script Reference

Integration with Other Skills

• wallet — balance checks, transaction execution
• across-bridge — cross-chain USDC transfers when rebalancing
• wallet-policy — ensure policy is set before on-chain operations
• coinglass / market-data — broader market context for yield decisions