TechFlow reports that on August 21, according to CryptoSlate, researchers from USENIX Security publicly disclosed a clock attack vulnerability targeting Solana's Proof of History (PoH) mechanism on August 12. The vulnerability had been privately reported to the Solana development team as early as December 2025. Research shows that a malicious scheduling leader can manipulate the PoH logical clock through "re-anchoring," slowing the progression of logical time. This allows the attacker to secure a longer transaction selection window within physical time and leverage the TowerBFT fork-choice mechanism to isolate an honest leader's blocks, requiring less than 33% stake. The Alpenglow security competition under Anza, featuring a 50,000 SOL prize pool, concluded on August 19. However, since the contest rules excluded "behaviors triggerable only when Alpenglow is not active," this vulnerability was not included in the evaluation scope.
The Solana development team stated they are aware of the relevant behavior, considering the probability of the worst-case scenario occurring under current conditions to be low, and expect the Alpenglow upgrade to fundamentally eliminate the attack's prerequisites. Currently, the Alpenglow code has been integrated into the Agave 4.2 client but has not yet been activated on mainnet, and is expected to go live officially with Agave 4.3. Prior to this, the transition-period risks of the vulnerability have not yet received public implementation-level analysis or official response.





