Ethereum and Bitcoin Developers Seek Access to Anthropic’s AI Open-Source Security Scanner

Frank Joaquin

Date: October 11, 2026

Ethereum engineering company Nethermind and Bitcoin wallet project ZEUS have applied to use Anthropic’s new AI-assisted vulnerability-finding service for open-source software.

Anthropic launched the opt-in service to scan eligible public repositories with its Frontier Red Team’s tooling. The company says accepted projects can receive vulnerability reports generated by its models without charge. The OSS Scanner portal and public repository document the system and application process.

Nethermind requested review of its repository, while ZEUS identified concerns relevant to payment and private-key handling in its Bitcoin and Lightning wallet. Cointelegraph’s report said the applications had not been merged at publication time.

The scanner’s output is model-generated and is not automatically a validated security finding. False positives, incomplete coverage and missed vulnerabilities remain possible. Projects must reproduce and assess any report before treating it as a confirmed defect or deploying a fix.

The crypto relevance is operational rather than token-based. Open-source wallets and clients expose large codebases where a subtle implementation error can affect assets directly. An AI system that reliably expands review coverage could shorten the path from vulnerable code to responsible disclosure, but only if human maintainers can verify and remediate the results safely.

TOKEN RECON ASSESSMENT

This is a credible AI-plus-crypto development because the models are being directed at live blockchain infrastructure rather than attached to a speculative token narrative. The evidence threshold for success is higher than application acceptance. A useful result would be a responsibly disclosed, independently reproduced vulnerability and a verified patch, with no sensitive exploit detail released before users are protected.