[DigitalToday reporter Chi-gyu Hwang (황치규)] Anthropic has updated its security service, 'Claude Security'. The update allows it to continuously check connected code repositories without people having to start inspections each time.

According to a recent report by Cybersecurity News, the update means Claude Security continues to check whether vulnerabilities already found have actually been fixed even after developers change code.

If a fix is confirmed, it marks the issue as "resolved". If the same issue appears again, it reverts the status to "unresolved".

The Claude Security update was introduced as an open beta for Claude Enterprise customers. Anthropic said that since the open beta in April, thousands of organisations have found more than 1 million vulnerabilities with Claude Security. The figure has not undergone independent verification.

Claude Security can be used after a security or development team creates a project, connects a code repository and sets the scope of code to scan. GitHub connects through the Claude GitHub app, and GitHub Enterprise Server for companies can also be used. Other code management services such as GitLab connect by issuing an access token.

Anthropic explained that it conducts security analysis using its own models, including Claude Mythos.