Anthropic has launched a new program called OSS Scanner, designed to help open-source developers identify and mitigate software vulnerabilities. The service utilizes the company's most powerful AI models, including Claude Mythos, to generate automated security reports for participating projects.
Several crypto-related entities have already submitted applications to join the program. Among the applicants are the Ethereum client developer Nethermind, the Bitcoin and Lightning wallet provider ZEUS, and the decentralized cloud computing marketplace VirtEngine.
The initiative aims to accelerate the vulnerability reporting process, which was previously slowed by manual human reviews. Anthropic stated that the service will deliver reports to projects immediately after code analysis is completed, providing a defensive advantage against potential exploits.
Anthropic will evaluate applicants on a case-by-case basis. The selection criteria focus on the project's importance to infrastructure, user security, potential exposure to remote attacks, and the number of dependencies from other systems.
The move comes as crypto firms face an increasing number of AI-assisted attacks. Companies like Boltz and PayPerQ have recently reported incidents where attackers developed exploits at a speed that outpaced manual patching efforts.
