Markets
BTC $55,761 -9.36% ETH $1,549 -10.24% BNB $511 -9.99% XRP $0.8707 -9.13% SOL $65.33 -9.41% LINK $8.07 -8.06% BTC $55,761 -9.36% ETH $1,549 -10.24% BNB $511 -9.99% XRP $0.8707 -9.13% SOL $65.33 -9.41% LINK $8.07 -8.06% TOTAL $1.95T BTC.D 57.2%

Ethereum

Ethereum’s Builder-Level Transaction Filtering Reopens the Censorship Argument

A tool that can stop a suspicious transaction before a block is assembled raises the question of who writes the rules and how they can be changed.

A security system that lets Ethereum applications block harmful transactions before they settle has moved from concept to production, and with it a familiar argument about how much filtering a neutral settlement layer can absorb.

The tool comes from Phylax and is integrated with block builder Titan Builder. Applications write rules; Titan checks transactions against them while assembling a block and decides whether to include them. The architecture matters because it operates at the builder level rather than inside the protocol, so no fork or validator vote is required — a single builder’s policy reaches every block it produces, and Phylax says Titan produces roughly one in every two Ethereum blocks.

The pitch is prevention rather than response. Phylax illustrates it with a DeFi vault that would release $1 million to a user whose shares were worth $1,000: a circuit breaker capping outflow per transaction stops the drain before it is finalised. Applied more broadly, the same mechanism could have limited losses in incidents such as the $6.3 billion Binance hack and a $380 million Bitget exploit, where attackers moved funds through DeFi protocols after the theft.

The objection is that a system able to block a transaction is also able to freeze a contract’s functions, and that the decision rests with block builders rather than with the applications’ users or the protocol’s governance.

Phylax founder and CEO Odysseas Lamtzidis says protocols can limit that risk by configuring the system to fail open — allowing transactions through when the security checks cannot complete, instead of blocking them by default. That choice is left to the deploying application, which is precisely where critics say it should not sit: they argue fail-open behaviour should be mandatory, and that time delays should be required before security settings can be changed, so a safeguard cannot be switched off in a single transaction.

There is a structural question underneath. Ethereum’s block-building market is already concentrated among a small number of builders, and a filtering layer that only becomes effective once a large builder adopts it deepens that dependence. Researchers have also asked whether the arrangement hands one company effective discretion over which transactions land, and what happens to applications on the wrong side of a rule they did not write.

For DeFi developers the tension is practical rather than philosophical. An exploit that drains a vault in one block leaves no room for after-the-fact monitoring, so the appeal of pre-settlement blocking is obvious. But the same mechanism can stop a legitimate transaction when a rule is misconfigured, and a protocol that cannot operate because its security provider is offline has traded one failure mode for another.

The coming decisions will settle much of it: whether fail-open is the default or an option, whether delays are compulsory, how rule sets are published, and whether users of a protocol get any say over the filters applied to their transactions. Until those are answered, the tool reads as both a genuine defence and a permissioning layer, depending on who is configuring it.

Share this article

Viber uses viber://forward?text= when the app is installed.