Arbitrum Hits the Brakes on Fresh Stylus Activations
Arbitrum's October 2, 2026 action restricted fresh Stylus activations on One and Nova. Deployment, active code and a separate conditional settlement guard need distinct treatment.
Prepared automatically from the linked sources. The category byline does not confirm personal authorship or review. Editorial policy.
Arbitrum’s October 2, 2026 emergency action paused fresh Stylus contract activations on One and Nova. This historical account, as of October 5, separates that restriction from a network shutdown. The official pause notice identifies activation as the step that makes a Stylus contract callable.
Deployment is a different step
The activation documentation describes two stages: deployment stores processed WebAssembly code onchain, while activation registers it with the ArbWasm precompile. Crucially, a newly deployed contract whose codehash matches an already-activated contract automatically uses the existing activation. Editorial analysis: The pause therefore cannot be described as preventing every new Stylus contract instance from becoming usable. The code’s activation state matters.
The Council made fresh activation prohibitively expensive by raising its activation gas requirement on both chains, according to its action report. That particular change adjusted an ArbOS configuration parameter and required no ArbOS upgrade. It restricted new activations, including application versions requiring that step, rather than deactivating all existing programs.
What the reported risk established
The Council linked its precaution to increasingly sophisticated AI-assisted attacks involving hand-crafted WASM programs outside the standard compiler toolchain. It described programs designed to produce unexpected behavior, including degrading chain performance. Previous ArbOS releases had patched reported bugs and attacks, while the pause addressed programs not yet patched by the current version. Editorial analysis: The report did not establish that those risks were unpatchable or give a deadline for a fix. Its description of a changing attack environment also did not quantify how many attacks had succeeded.
In the October 2 report, known Stylus bugs primarily concerned chain liveness, including denial of service. The Council said it had not discovered an attack permitting theft of user funds. Editorial analysis: That time-qualified finding does not establish that all possible fund-loss vulnerabilities had been excluded.
Active code, expiry and the separate guard
The guide says active Stylus contracts keep running until expiry. Several instances can share one codehash and activation. Expiry clears activation state while the code remains onchain; a stored deployment therefore need not remain executable.
CryptoSlate’s account describes permissionless keepalive renewal before an active program expires. Reactivating an expired program, or one requiring reactivation after a Stylus version change, remained blocked during the pause. The same report identifies ordinary Solidity deployment and execution as unaffected. Renewal of a still-active program is distinct from reactivating one that has already expired.
The October 2 action also installed a separate BoLD one-step-proof safeguard on Arbitrum One. Anyone could present two conflicting answers to the same step of an open challenge. If the proof accepted both, the guard could put One’s settlement to Ethereum on hold. One would continue processing, while messages not yet confirmed on Ethereum, including withdrawals, would wait. Installing this safeguard did not itself suspend withdrawals: the delay depended on the conflict condition being met. This conditional measure is distinct from the Stylus activation restriction.
Reopening and editorial analysis
The developer notice said the Arbitrum Foundation would work with ArbitrumDAO on next steps for reopening activations while restricting hand-crafted WASM programs. The saved notice contains no reopening date. It does not establish that a reopening decision had already been taken or prescribe additional mitigation measures.
Editorial analysis: The narrower description is also the more useful one. An activation restriction identifies a particular execution prerequisite; it does not, by itself, describe the operational state of every application or the entire network. Reporting the restricted step makes the scope legible without turning a precaution into a safety guarantee.
Editorial analysis: For a deployment using identical code, the relevant distinction is whether the shared activation remains valid. Describing every new contract instance as blocked would erase the reuse mechanism documented in the activation guide.
Which stage does a particular case involve: storing code, using a valid shared activation, renewing before expiry, or requesting fresh activation?