Tools I Built
Because They Didn't Exist
A lot of my work looks boring if you only value "apps". But tooling is what makes systems stable and scalable.
Over time, I kept building pieces that didn't exist:
Deterministic Create2 Patterns
The problem: how do you give a user a deposit address on Chain B before you've even deployed a contract there? Create2 lets you compute the address deterministically. I built patterns that let Gasyard show users where to send funds, then deploy the receiving contract just-in-time when the deposit arrives.
Solver Rebalancing & Routing Engines
Cross-chain protocols talk about "solvers" like they're magic. They're not. Someone needs to move liquidity between chains, optimize for gas costs, and handle failures gracefully. The rebalancing engine I built tracks inventory across 10+ chains and executes rebalancing when thresholds hit.
Atomic Swap+Bridge Executors
Most bridges move Token A from Chain X to Chain Y. But users want to end up with Token B on Chain Y. The atomic executor bundles swap and bridge into a single flow, so if one fails, everything reverts. One transaction, one failure mode.
Vault Automation
Inspired by Safe/Zodiac patterns. How do you let automated systems move treasury funds without giving them full control? Policy-based permissions: this address can move up to X tokens per day, only to whitelisted destinations, with time delays for large amounts.
x402plus Payment Standards
The newest experiment. HTTP 402 was supposed to be "payment required" but never got used. x402plus is my take on making payments programmable: agents and services that can request, verify, and settle payments without human intervention.
Why Tooling Matters
Most people think "tooling" means libraries you import. For me, tooling is the infrastructure that makes systems reliable when they're under pressure: when liquidity is moving, when users are impatient, when chains have different finality assumptions.
The Create2 patterns weren't academic exercises. They solved a real problem: how do you give a user a deposit address that's predictable across chains, before the contract even exists? The solver engines came from watching rebalancing fail when routes were hardcoded. The atomic executors emerged because "bridge then swap" is two failure points, not one.
Every tool here came from hitting a wall on a real project and deciding "this should exist, and it doesn't."
The x402plus direction is the newest layer: making payments programmable for AI agents and automated systems. It's early, but it builds on the same principle. Tooling should solve problems that keep showing up.