Crypto projects line up for Anthropic's new AI bug-sniffing service 🐛🔍

Several cryptocurrency projects, including Ethereum client developer Nethermind and Bitcoin and Lightning wallet ZEUS, submitted enrollment requests on Friday for Anthropic's newly launched OSS Scanner, an opt-in program that provides open-source developers with vulnerability reports generated by the company's most powerful AI models. Anthropic announced the program on Thursday, stating that the reports will be produced by "our strongest models (including Claude Mythos) to give open-source projects the largest defensive advantage." Pull requests to the OSS Scanner GitHub repository show Nethermind applied for audits of its entire repository, while ZEUS requested examination of its self-custodial wallet for weaknesses affecting payments, private keys and connection to Lightning Services. Another applicant, VirtEngine, is described in its pull request as a decentralized cloud computing marketplace structured as a Cosmos SDK chain. None of the pull requests had been merged at the time of publication.

OSS Scanner extends Anthropic's earlier Project Glasswing work and is intended to shorten the timeline for sharing vulnerabilities with open-source maintainers. The company said it already scans open-source software and sends reports following human review, but that the manual process is slow "so we're not always able to share vulnerabilities as quickly as we would like." Under the new program, reports will be delivered to participating projects as soon as their code has been scanned. Anthropic said projects will be assessed case by case, based on factors including their importance to infrastructure and user security, exposure to remote attacks and the number of users or dependent projects. Other applicants listed in the pull requests include developers of AI assistants, agent-security tools and machine-learning infrastructure, alongside projects providing software development tools, cloud storage and energy-system controls.

The applications come as crypto firms seek access to frontier AI models to defend against increasingly capable attackers. Bitcoin swap provider Boltz suspended operations in August, citing attackers developing exploits faster than its team could patch them, while crypto-payment service PayPerQ reported repeated attacks it suspected were AI-powered. Anthropic warned on Thursday that AI may favor attackers in the near term, noting that exploitation is becoming cheaper while verification and patching remain dependent on people and continue to move slowly.