Bane.
Launching on Pons V2 · $BANE
A 3% tax on every trade keeps buying tokenised stocks into a shared basket. An autonomous agent sets the weights. MandateGuard rejects anything outside the mandate. The agent's key can propose, and it can never withdraw.
What happens every epoch
Every trade pays 3%. Eighty percent of it buys tokenised stocks into the holders' basket, thirteen funds the agent's inference and gas, seven goes to the protocol. Creator fees add another one percent in USDG. Nobody presses anything and the flow never stops.
Your share of the basket accrues as balance multiplied by holding time, never by snapshot. A bot that buys one block before the close earns nothing at all. There is no distribution schedule to snipe, so claim whenever it suits you.
Before anything moves, the agent publishes a keccak256 of the epoch, the weights, its reasoning and a nonce. Next epoch the text is opened and checked against that hash. Reasoning lands first and the outcome second, never the other way around.
The whitelist
Drag them. Throw them. The guard puts them back inside the mandate.
One epoch, start to finish
The agent picks its weights and writes the hash on chain before a single swap happens. The reasoning is locked in the open where nobody can edit it afterwards, including us.
MandateGuard runs the vector through eight checks. Whitelist only, weight under the cap, no position below five percent, turnover inside the budget, one rebalance per epoch. Any failure cancels the epoch and the basket stays exactly where it is.
What survives goes through the Uniswap v4 pools as a single batch. Slippage over the limit reverts the whole thing, so there are no partial fills. Next epoch the thesis is revealed and re-hashed against the seal.
This node is running it now
Every hash lands before the swap it describes. The thesis behind it opens one epoch later and is checked against the seal, so a call the agent got wrong stays on the record next to the one it got right.
The downside
The worst it can do is hold whitelisted stocks in bad proportions, and move at most a fifth of the basket in a day. Withdrawing assets is not a function its key has. There is no path from a bad decision to an empty vault.
Down twenty-five percent across seven epochs and the breaker freezes the agent on its own, parking the basket in USDG. Switching it back on takes a holder vote and a three-day timelock.
The limits, the whitelist and the agent itself are all replaceable by vote. Three days of timelock sit between the decision and the change, which is time for anyone who disagrees to leave first.
These are the numbers the contract checks against, published rather than described. The console shows every vector measured against them, epoch by epoch.
mandate: whitelist: 8 tokenised equities + USDG max_weight: 35% min_position: 5% max_turnover: 20% NAV per epoch rebalances: 1 per epoch breaker: −25% over 7 epochs timelock: 3 days agent_key: propose only
Rollout
Static weights. Accrual and claims are running, the agent is off.
Thirty epochs of sealed public calls and zero execution. The hypothetical book is charted next to the real one.
Execution goes live at seven percent NAV turnover and a twenty percent weight cap.
A holder vote unlocks thirty-five percent weights, twenty percent NAV turnover and new whitelist entries.
The whole thing is running right now
The console shows the basket rebalancing, the guard rejecting bad vectors, the sniper bot earning nothing, the breaker and the votes.