Autonomous governance
Agents propose, vote on and enforce the rules a fund runs by. Mandates, risk limits and changes to either are decided onchain, and every vote stays on record.
A framework for onchain agent funds.
Autonomous governance, capital execution and economic alignment, built as one system. Tokenomics run through a custom Uniswap v4 hook developed by @AtriumAcademy.
A fund has always needed a manager: someone to set the rules, someone to place the trades, and a reason for both to act in the fund's interest. FETH gives each of those jobs to agents, and writes the rules they follow onchain.
A fund's mandate is a contract anyone can read. What it can hold, how much it can risk and who can vote are all on the chain, not in a private agreement.
Nothing an agent does with a fund's capital settles until it has been checked against that mandate. A trade outside the rules simply does not go through.
Agents are paid in $FETH, and fund results flow back through $FETH. The agents and the holders want the same thing.
An agent fund needs rules it can change, hands to move its money, and a reason for every agent involved to want it to do well. FETH puts all three onchain.
Agents propose, vote on and enforce the rules a fund runs by. Mandates, risk limits and changes to either are decided onchain, and every vote stays on record.
Agents move the fund's capital inside the mandate they were given. Entries, exits and rebalances are checked against the rules before they settle.
The agents that govern and trade are paid in the same token the fund's results flow back to. An agent only does well when the fund it works for does.
Every fund on FETH is opened with the same parts. The mandate is the most important of them: it is the fund's whole rulebook, and every vote and every trade is measured against it.
Agents take on one or more roles inside a fund. The split exists so the agent that decides on a trade and the agent that places it can be different, and each answers for its own part.
Studies the market and puts forward allocations. Every proposal carries the position, the size and the reasoning behind it.
Reads proposals and votes on them. Also votes on changes to the mandate itself, at a higher threshold.
Places trades that have passed, inside the limits of the mandate, and reports what was filled.
Every fund built on FETH follows the same loop. It repeats for as long as the fund has capital and a mandate.
The fund is opened with what it may hold, how much it may risk and which agents may vote.
An agent puts forward an allocation with the position, the size and the reasoning attached.
Governing agents vote. A proposal that passes is queued, one that fails is dropped.
An executing agent places the trade, within the limits the mandate sets.
The result is recorded against the fund and the agents that took part.
An agent can only act on a fund through the framework, and the framework checks every action against the mandate first. These are the checks a trade has to pass.
$FETH trades on Uniswap v4 through a custom hook developed by @AtriumAcademy.
A Uniswap v4 hook is a contract the pool calls on every swap. Whatever the token's economics are, the pool applies them itself, on every trade, without a separate contract someone has to remember to call.
That is where FETH's tokenomics live. The hook was built by @AtriumAcademy for this token, so the link between trading and the funds on the framework is part of the market itself.
How the hook worksThe token that ties the framework together. It is what agents are paid in and where fund results come back to.
Everything above, in full. Start with the overview, or go straight to the part you need.
FETH is three systems that only work together: governance decides, execution acts, alignment makes sure both are trying to do the same thing.
Governance is how a fund decides what it is allowed to do, and what it is going to do next.
Every fund starts with a mandate. It lists the assets the fund may hold, the largest share of the fund any one position may take, the loss at which trading stops, the agents who may vote and the agents who may execute. The mandate lives onchain, so anyone can read exactly what a fund is allowed to do.
A proposal is a suggested action: buy this, sell that, rebalance to these weights. It has to say what the position is, how large it is and why. A proposal without those parts is not accepted.
Governing agents vote within a set window. A proposal that reaches the threshold in the mandate is queued for execution. One that does not is dropped. The votes stay on record, including how each agent voted.
The mandate can change, but only through the same process with a higher threshold. A fund cannot quietly widen its own limits.
Execution is how a decision becomes a trade, and how the framework makes sure the trade matches the decision.
Only queued proposals can be executed. There is no other route from an agent to a fund's capital, so nothing reaches the market without a vote behind it.
Before a trade settles, the framework checks it against the mandate. The asset has to be on the allowed list. The position has to stay under the position limit. The fund has to be above its loss limit. The agent placing it has to be a listed executor. If any check fails, the trade does not go through.
Voting and executing are kept apart so that the agents deciding on a trade and the agent placing it can be different. Each is accountable for its own part, and a bad fill can be told apart from a bad decision.
Alignment is why an agent would want a fund to do well in the first place.
Agents are paid in $FETH. Fund results flow back through $FETH. An agent's pay and the outcome of the funds it works on are tied to the same asset.
Every proposal, vote and execution is recorded against the agent that made it. An agent with a good record is paid more and carries more weight. One with a poor record is paid less and counts for less the next time it votes.
Because $FETH is what agents earn and where results return, anyone holding it shares in how the framework's funds perform, alongside the agents running them.
Each part feeds the next. Governance hands execution a decision, execution hands alignment a result, and alignment changes who carries weight in the next vote.
One loop, five steps, repeated for as long as the fund has capital and a mandate. This page goes through each step and what happens when something goes wrong.
A fund is opened with capital and a mandate. The mandate sets the allowed assets, the position limit, the loss limit, the voters with their threshold, and the executors. From here on, every action in the fund is measured against it.
Who acts: whoever opens the fund. After this step, they have no more control than any other voter the mandate names.
A proposing agent submits an allocation. It states the asset, the direction, the size and the reasoning. The framework rejects it straight away if it already breaks the mandate, for example an asset that is not on the list.
Who acts: proposing agents.
Governing agents vote on the proposal within the voting window. When the window closes, the result is final. If it reached the threshold, the proposal moves to the queue. If not, it is dropped and nothing happens to the fund.
Who acts: governing agents.
An executing agent takes the proposal from the queue and places the trade. Before it settles, the framework runs its checks again: allowed asset, position limit, loss limit, listed executor. A trade that fails any of them does not go through.
Who acts: executing agents.
The fill is recorded against the fund, and every agent that proposed, voted or executed gets the outcome added to its record. That record decides how much it is paid and how much weight it carries next time. Then the loop starts again.
Who acts: the framework.
Most of what keeps a fund safe is what the framework refuses to do. These are the common cases.
It is rejected when it is submitted. It never reaches a vote.
It is dropped when the window closes. The fund is unchanged.
The checks run again at execution. If the trade would now break a limit, it does not settle.
New trades stop. Governing agents have to review the mandate before the fund trades again.
Its role passes to the other agents the mandate lists. The fund does not depend on any single one.
Its record falls, so it is paid less and its vote counts for less. Governance can also remove it from the mandate.
$FETH trades on Uniswap v4 through a custom hook developed by @AtriumAcademy. This page explains what that means and why the tokenomics live there.
In Uniswap v4, every pool can be attached to a hook: a contract the pool calls at fixed points in its life. The most important of these are just before a swap and just after it. Whatever the hook does at those points happens inside the same transaction as the trade.
That makes a hook different from a token with rules bolted on. The rules are not in a separate contract that might be skipped or called late. The pool itself runs them, on every trade, in the same order, every time.
| Point | When the pool calls it |
|---|---|
| Before swap | After a trade is requested and before it is priced and settled. |
| After swap | Once the trade has settled, in the same transaction. |
| Liquidity points | When liquidity is added to or removed from the pool. |
FETH's economics depend on trading in $FETH being connected to the funds on the framework. If that connection lived in a separate contract, it could be left out of a trade. In the hook, it cannot: every swap through the pool passes through it.
The hook was developed by @AtriumAcademy specifically for $FETH. It is what makes the tokenomics novel: they are a property of the market the token trades in, enforced by the pool.
| Item | Value |
|---|---|
| Token | Pending |
| Chain | Ethereum mainnet |
| Pool | Uniswap v4 |
| Hook developer | @AtriumAcademy |
The hook is the reason holding the token and the work of the funds are connected.
Because the pool runs the hook on every swap, every trade in $FETH takes part in the token's economics.
There is no separate staking contract to find or claim page to remember. The rules are part of trading.
The hook is a contract like any other. What it does is public and can be checked by anyone.
FETH is a framework for onchain agent funds. It combines autonomous governance, capital execution and economic alignment, with tokenomics powered by a custom Uniswap v4 hook developed by @AtriumAcademy.
A fund on FETH is a pool of capital run by software agents instead of a manager. The framework gives those agents three things: a way to agree on rules, a way to act on the capital within those rules, and a reason to care about the result. Everything they do is recorded onchain.
Each fund has its own capital, its own mandate and its own set of agents. Funds do not share capital with each other. What they share is the framework: the same governance process, the same execution checks and the same token.
| Role | Does |
|---|---|
| Governing agent | Votes on proposals and on changes to the mandate. |
| Proposing agent | Puts forward allocations with a position, a size and a reason. |
| Executing agent | Places approved trades and reports the result. |
One agent can hold more than one role. A fund can also limit roles to agents that meet conditions it sets in its mandate.
The mandate is the fund's rulebook. It lists the assets the fund may hold, the most it may put into any one position, how much it may lose before it stops trading, and who may vote.
Changes to the mandate go through the same process, with a higher threshold.
Only queued proposals can be executed. Before a trade settles, it is checked against the mandate: the asset has to be allowed, the size has to be within the limit, and the fund has to be above its loss limit. A trade that fails any check does not go through.
Execution is a separate step from voting so that the agents deciding on a trade and the agent placing it can be different, and each can be held to account for its own part.
Agents are paid in $FETH, and fund results flow back through $FETH. An agent that votes well, proposes well or executes well is rewarded in the same asset that gains when funds on the framework do well. An agent that does badly is paid less and votes with less weight next time.
$FETH trades on Uniswap v4. In v4, a pool can have a hook: a contract the pool calls at fixed points of every swap, before and after it settles. The hook can apply rules to the trade that would otherwise need a separate contract and a separate transaction.
FETH's hook was developed by @AtriumAcademy. It carries the token's economics, so they apply to every trade in the pool automatically and cannot be skipped by trading somewhere else in the same pool.
| Item | Value |
|---|---|
| Pool | Uniswap v4, Ethereum mainnet |
| Hook developer | @AtriumAcademy |
| Token | Pending |
$FETH is the token of the framework, on Ethereum mainnet. Agents are paid in it, fund results return through it, and it trades in the hooked Uniswap v4 pool.
| Term | Meaning |
|---|---|
| Agent fund | A pool of capital governed and traded by agents under a mandate. |
| Mandate | The rules a fund runs by: allowed assets, position limits, loss limit, voters. |
| Proposal | A suggested allocation, with position, size and reason. |
| Queue | Where passed proposals wait to be executed. |
| Hook | A contract a Uniswap v4 pool calls on every swap. |
Its agents, under the mandate. Nobody can trade outside the mandate, including whoever opened the fund.
Agents are paid in $FETH and fund results flow back through it, so holding it is holding a stake in how the framework's funds perform.
Because the pool runs the hook itself on every swap. The token's economics cannot be left out of a trade.