A BTC stake of forty billion, stuck in the hands of nine signers
After reading through Babylon’s script implementation, the first thought that came to mind was: Bitcoin script’s support for Covenants is like a phone from twenty years ago that can’t send video—not that it doesn’t want to, it just really can’t.
Bitcoin is still doing pretty well; BTC is up again!
In Babylon’s staking process, there’s something called a Covenant Committee. It’s not a custodian—official docs emphasize this repeatedly. The funds really do stay in your wallet, and the script hard-codes the legal destination of the funds, so the committee can’t move them. But to activate staking, you have to collect the signatures of this group; only after enough signatures can you start earning BABY. If you want to unstake early, sorry, you still need their signatures.
These 9 keys are currently public in the testnet docs: Babylon Labs itself holds three seats, and the rest are Cubist, Informal Systems, Zellic, RockX, AltLayer, and CoinSummer Labs. The mainnet composition is probably similar, but it’s not fully transparent yet.
The most serious problem is hidden in the details. Zellic’s audit report mentioned a scenario: if one committee member deliberately returns an invalid signature, because it responds the fastest, it may actually be accepted first, causing the entire unstaking transaction to fail outright. Not to mention the case where everyone is passively uncooperative—then the stake is completely locked, and you can only wait for the original timelock to expire naturally, which could be weeks at best or months at worst.
Once mainnet starts running, hundreds of billions of dollars in liquidity will be held up by these 9 nodes. No matter how solid EOTS’s cryptographic logic is, it can do nothing against "the signer just doesn’t do the work."
How do you solve this? A dashboard that publicly shows members’ signature latency and failure rates in real time should be the bare minimum. After all, funds being impossible to steal is the baseline; whether transactions can be completed on time is the lifeline. These two things can’t be bundled together and brushed off with a vague "self-custody." @BabylonLabs_io $BABY #baby
After reading through Babylon’s script implementation, the first thought that came to mind was: Bitcoin script’s support for Covenants is like a phone from twenty years ago that can’t send video—not that it doesn’t want to, it just really can’t.
Bitcoin is still doing pretty well; BTC is up again!
In Babylon’s staking process, there’s something called a Covenant Committee. It’s not a custodian—official docs emphasize this repeatedly. The funds really do stay in your wallet, and the script hard-codes the legal destination of the funds, so the committee can’t move them. But to activate staking, you have to collect the signatures of this group; only after enough signatures can you start earning BABY. If you want to unstake early, sorry, you still need their signatures.
These 9 keys are currently public in the testnet docs: Babylon Labs itself holds three seats, and the rest are Cubist, Informal Systems, Zellic, RockX, AltLayer, and CoinSummer Labs. The mainnet composition is probably similar, but it’s not fully transparent yet.
The most serious problem is hidden in the details. Zellic’s audit report mentioned a scenario: if one committee member deliberately returns an invalid signature, because it responds the fastest, it may actually be accepted first, causing the entire unstaking transaction to fail outright. Not to mention the case where everyone is passively uncooperative—then the stake is completely locked, and you can only wait for the original timelock to expire naturally, which could be weeks at best or months at worst.
Once mainnet starts running, hundreds of billions of dollars in liquidity will be held up by these 9 nodes. No matter how solid EOTS’s cryptographic logic is, it can do nothing against "the signer just doesn’t do the work."
How do you solve this? A dashboard that publicly shows members’ signature latency and failure rates in real time should be the bare minimum. After all, funds being impossible to steal is the baseline; whether transactions can be completed on time is the lifeline. These two things can’t be bundled together and brushed off with a vague "self-custody." @BabylonLabs_io $BABY #baby
