CL vault system: production fixes, pilot tooling, and cleanup - #2

Summary

Complete CL (Aerodrome Slipstream concentrated-liquidity) vault system for Base: production-hardened contracts, pilot deployment tooling, and a full dead-code cleanup pass.

Contract changes

• CLVault — NAV counts vault + strategy idle; deposit reads NAV before token transfer (fixes ~5% share-dilution); range-aware idle rebalance via planSwapForMint; position NFT re-staked into the gauge after every deposit/withdraw/rebalance; rebalance tail runs a full doHardWork so accumulated strategy idle is absorbed into the new position; sweepDust replaced by sweepStrayToken (token0/token1 blocked — they back PPS)
• CLRebalanceHelper — planSwapForMint plans at spot (not TWAP) with a one-step price-impact correction using pool active liquidity; burn/mint mins take per-side min(TWAP, spot) so out-of-range burns don't revert; precision-edge clamps at range boundaries
• AerodromeCLStrategy — idle absorption decoupled from the reward threshold (_absorbIdleIntoPosition runs on every doHardWork); stakePosition for vault-driven restake; preInteract dust sweep keeps strategy idle inside NAV/payout math
• CLWrapper — fee-aware previews, per-leg swap slippage guard (WrapperSwapBelowPrecision fails loud instead of bleeding value), truncation guard, immutable token caching, wrapper always ends interactions at zero balance
• CLChainlinkChecker — Chainlink Automation upkeep for rebalances; replaces the Gelato-style CLRebalanceChecker (deleted)

Cleanup

82-agent audit with adversarial verification; 73 confirmed findings applied: dead functions, legacy paused/sellFloor/slot-array machinery, unused events/imports/params, 4 dead files, interface trims, stale comments. CLVault bytecode 22,422 bytes (2.1 KB under EIP-170). Zero CL compile warnings.

Deploy tooling

Bridge-Storage flow (EOA does setup, final step flips every contract to the multisig-governed Storage), recovery script for partial deploys, checker deploy script, pilot configs for cbETH/ETH1 and tBTC/cbBTC.

Test plan

• 169 passing across 12 CL suites at FORK_BLOCK=32897925: user-interaction audit (34), wrapper audit (26), user fairness both vaults (32), user deep-invariants (17), rebalance deep-dive incl. 100-case fuzz (8), live-pool planner comparison both pools (8), adversarial guards (4), mins helper (8), invariants (2), cbeth-eth1 (2), wrapper (8), live-controls (20)
• Rebalance planner on live pools: avg/max leftover 0 bps (cbETH & BTC) vs legacy ~4,100 bps avg

🤖 Generated with Claude Code