risks read before buying

What can go wrong

Nothing is deployed yet, and when it is, it will be beta software holding real money. Every number below is the honest one, and the ones that argue against this design are on the page rather than in a footnote.

Two things to read before anything else

The contracts are unaudited, and Flap's Guardian can replace the implementation through the beacon this vault runs behind. Everything else on this page is downstream of those two.

liquidation distance by target leverage
How far BNB has to fall before each leverage target is liquidated on Venus, at its 80% collateral factor. The four rows read: 2x at health 1.600 surviving a 37.5% drop; 3x, the level this vault builds, at health 1.200 surviving 16.7%; 4x at health 1.067 surviving only 6.3%; and 5x at health 1.000, where the position is already at the liquidation point before the price has moved at all, which is why this vault hard-codes 3x rather than offering 5x with a warning.
the list worst first

Everything that can hurt you

A 16.7% move liquidates the position

Not "may be affected by volatility". Liquidated: Venus sells the collateral, and the treasury is permanently smaller. The vault reduces leverage on its own when health falls under 1.10, and waives its cooldown to do it, but a fast enough move does not wait for a five-minute wake.

Flap's Guardian can replace the implementation

The vault runs behind a beacon proxy whose owner is Flap's Guardian, and that upgrade path is the emergency mechanism — it is the reason the vault ships no emergency-withdraw function of its own. An upgrade can change every constant this site quotes, add a pause, or add a withdrawal. This is the single largest trust assumption in the design.

We are not trustless. We changed who you trust.

Besides the Guardian above: the position borrows from Venus and trades on PancakeSwap, and it is woken by Flap's trigger service. If any of those fails, the treasury is affected. This is a better trade than one operator's private key. It is not the absence of one.

Settlement costs the treasury while nothing happens

Each wake costs 0.0002 BNB whether or not there is work. A wake that finds nothing books the next one an hour out instead of five minutes, which slows the bleed but does not stop it. A token with very little trading slowly spends its own treasury staying awake.

Leverage decays

A 3× position loses value in a choppy market even if BNB ends where it started, and the carry is −4.87% a year before any price move. This is arithmetic, not a bug.

Every step needs someone to show up

Flap's service wakes the vault, and anyone may restart the chain if it stops. Both paths pay. But nobody is obliged to appear, and a trigger call that fails is recorded as failed.

Dividends have a holding threshold

A balance under the threshold earns no dividend share at all. That rule belongs to Flap's dividend contract, not to this vault, but it decides whether you can be paid.

The contracts are unaudited

They have been reviewed adversarially against themselves many times and checked rule by rule against Flap's specification. That is not an audit and we will not call it one.

what is fixed and therefore unfixable

Four things nobody can change after launch

Fixed at launchConsequence
The project addressReceives 40% of every gain, forever. The vault has no setter.
The tax split80 / 20 is written into the token. It cannot be retuned later.
Tax rate and duration2% for 100 years.
The dividend thresholdDecides who is eligible to be paid, for the life of the token.

This cuts both ways. Nobody can quietly raise the project's share — and nobody can fix a number that turns out to be wrong, either. Changing any of them means a new token.

Everything above is checkable

Including the numbers that make this look worse.

How to verify