Uniagent

This is what protected agents look like in production. Every swap fired a Uniswap v4 hook. Every hook checked the policy. Nothing broke.

@ajidatlolebar @tryuniagent @bankrbot They have just stolen our X.

@tryuniagent @bankrbot why the x acc on the website refers to @useuniagent ?

Uniswap v4 hook ecosystem is growing. Uniagent grows with it. Every new pool is another enforcement surface. Every new agent is another potential policy.

This is why hook-level enforcement is different. Every swap goes through beforeSwap() first. That's where Uniagent lives. Budget exceeded - the hook reverts the transaction before it executes. Not a warning. Not a notification. The swap never happens. App-level guardrails run after the money moves. Hooks run before.

Your trading bot just tried to move $12,847 in a single swap. You set a $500 daily limit. The Uniswap v4 hook read the transaction at the pool level, checked the policy, and stopped it before it executed. Not a middleware warning. A hard block, on-chain, in real time. The agent didn't overspend because it couldn't. That's the difference between app-level guardrails and hook-level enforcement.

$7,550 spent across a fleet of AI agents in 14 days. 74 transactions. 0 alerts. Zero alerts means every swap, every program call, every stake stayed inside the budget. Enforced by Uniswap v4 hooks in real time, no manual review, no incidents. This is what a protected fleet looks like.

@UniswapBuilders @okx execution without limits is half the stack. the other half is knowing when to stop.

Uniswap is for agents That's why we're excited to support APP, @OKX's open standard for agentic commerce npx skills add Uniswap/uniswap-ai -s pay-with-app

@UniswapBuilders spending enforcement for AI agents via v4 hooks. if your agent swaps on uniswap it needs a budget. we're building that layer.

What are you building this week?

@VitalikButerin the runtime side of this is equally interesting. correct code can still be misused. on-chain enforcement at execution time feels like the missing piece

Many people have claimed that with AI-assisted bug finding, secure code (and hence trustless anything) will be impossible. I have a much more optimistic take, and AI-assisted formal verification is a major part of the reason why: vitalik.eth.limo/general/2026/0…

@Uniswap and AI agents are going to move a lot of it autonomously. nobody's really thought through what happens when they go over budget

All of the world’s value is coming onchain

Overspend protection for AI agents shouldn't require a custom contract. Uniagent turns every Uniswap v4 pool into an enforcement layer: any agent, any wallet, fully on-chain. Powered by @Uniswap v4 hooks. Set a daily budget, define a whitelist, and get paid back automatically when any limit is breached.

@Uniswap @angstromxyz @KyberNetwork @dopplerprotocol @bankrbot @clanker_world wild how fast hook ecosystem went from "cool primitive" to products people use. v4 quietly became an app platform and nobody talking about it enough

Your trending hooks on Uniswap ⭐️

AI agents have no spending limits. They swap until the wallet is empty. Uniagent fixes this at the hook level. Every swap your agent executes fires a Uniswap v4 hook. Uniagent reads it instantly, checks the policy, freezes the wallet if the limit is crossed.