Arbitrum Halts New Stylus Contract Activations After AI-Assisted Vulnerability
Arbitrum has temporarily paused new activations of Stylus contracts on its Arbitrum One and Nova networks. The decision came in response to AI-assisted development tools that have made it easier to create attacks targeting hand-crafted WebAssembly (WASM) programs. The emergency measure, enacted by the Arbitrum Security Council at 11:30 EST on October 2, 2026, aims to prevent new vulnerable contracts from entering the network while allowing existing ones to operate normally.
The restriction focuses solely on activation, meaning existing Stylus contracts remain active and users can continue interacting with them. Developers can also renew active contracts through the existing permissionless keepalive mechanism. The pause does not affect Solidity-based development, which continues unaffected. Arbitrum implemented the restriction through a gas configuration change, making new activations economically impractical without requiring a network upgrade.
In addition to the Stylus pause, Arbitrum introduced a safeguard for its one-step proof mechanism under BoLD. The new OspSoundnessGuard can pause settlement between Arbitrum One and Ethereum if conflicting proofs are detected, ensuring network integrity. This measure is precautionary, as no exploits have been reported against the production network.
Developers face a temporary constraint on deploying new Stylus contracts, while users of existing applications can continue using them without disruption. Arbitrum has not announced when Stylus activations will resume but plans to work with ArbitrumDAO on the timing and process for restoring them. The incident underscores the growing security challenges in blockchain development as AI tools lower the barrier to sophisticated attacks.