KEY POINTS
- Phylax and Titan Builder announced an integration on Nov. 10 that lets DeFi apps set rules to keep harmful transactions out of blocks.
- Critics warn that apps relying on builder attestations could become freezable and that the model could deepen builder market concentration.
- Phylax says the system is opt-in and recommends a fail-open default, though protocols can choose otherwise.
Phylax, a security company, has introduced a system that lets decentralized finance protocols block potentially harmful transactions before they are finalized on Ethereum. Cointelegraph reported that the tool has also set off a debate over censorship and the concentrated market for Ethereum block builders.
The company announced on Nov. 10 an integration with Titan Builder. Under it, applications can define rules that exclude transactions likely to cause damaging outcomes while a block is being assembled. Phylax says Titan produces one out of every two Ethereum blocks, and the builder checks transactions against the rules before including them.
The aim is to stop exploits from settling at all, rather than depending on monitoring or emergency responses once an attack is underway. Phylax gives the example of a flawed vault that could wrongly send $1 million to a user whose shares are worth $1,000. A circuit breaker could cap how much leaves the vault in one transaction. The report says similar safeguards might have limited losses in large incidents such as the $6.3 billion Binance hack and the $380 million Bitget exploit.
Digital assets lawyer Gabriel Shapiro cautioned that protected applications could become effectively freezable. Smart contracts can be set to reject any transaction not included in a block from a builder that ran the required checks. If those builders stopped issuing attestations, the protected functions could keep rejecting transactions until the attestations returned.
Phylax founder and chief executive Odysseas Lamtzidis said protocols can avoid this by letting transactions through when the security checks are unavailable. He described that fail-open behavior as the suggested default, but it is optional, and some protocols may favor safety over availability. Shapiro argued that fail-open protections should be mandatory and that changes to the safeguards should be subject to time delays.
Researchers also raised concerns about market structure. Fan Zhang said the approach has appealing security benefits, but that it has become workable only because Ethereum's builder market is so centralized. He warned that adding code at such a critical point creates new failure modes, including implementation errors and the theft, loss or impersonation of builder signing keys. Sen Yang said reliance on a few builders for enforcement could give them a competitive edge and worsen concentration. He added that the proposed enshrined proposer-builder separation upgrade, slated for the upcoming Glamsterdam upgrade, would not directly address concentration. Transactions excluded by dominant builders could still be included by others, which would likely mean delays rather than permanent exclusion.
Phylax maintains that the system creates no new authority over transactions because applications set their own rules and participation is explicitly opt-in. Lamtzidis compared the checks to a require() statement in a smart contract and said censorship implies arbitrary decisions or coercion. Zhang drew a similar line, saying voluntary filtering by applications is acceptable because users can leave, while externally forced censorship is the real problem.
Stefanie Boss, a co-author of research on censorship resistance in Ethereum's consensus mechanism, considers the risk acceptable because the rules target outcomes rather than the identity of the sender. She said the harm from hacks outweighs the harm of such exclusions. She also warned that the rules could be turned toward enforcing sanctions or politically motivated exclusions, and said safeguards are needed to keep them fair and focused on security.