Rubrik (NYSE: RBRK) on Tuesday announced the launch of Rubrik Code Guardian, an AI-powered code security testing service designed to detect vulnerabilities in customer software.
The Palo Alto–based enterprise data security firm said the service is built on Anthropic’s Claude Mythos 5 model and operates through a Rubrik software harness constructed specifically for testing code. Rather than scanning live production environments, the tool runs its analysis on air-gapped, secure clones of code repositories.
According to Rubrik, the system tests for multi-step vulnerability chains that attackers could exploit, identifying chains that span across files, services, authentication patterns and cloud boundaries. The platform then validates findings for exploitability and ranks them based on both technical exploit potential and business impact. Confirmed critical issues are pushed into developer workflows through integrations with Jira and GitHub, accompanied by file-level remediation guidance. Rubrik also said the service maintains recovery workflows capable of restoring code in the event of a security incident.
Arvind Nithrakashyap, Rubrik’s co-founder and chief technology officer, said the company is using frontier AI to scale vulnerability detection beyond the capacity of traditional code scanning tools. "We took Anthropic's powerful model and built a Rubrik software harness to test on our code," Nithrakashyap said.
Michael Moore, cybersecurity lead at Anthropic, stated that Rubrik is among the partners working to deploy Claude Mythos 5's capabilities for defensive purposes.
Rubrik Code Guardian is currently accepting select design partners for a private preview. The company added a disclaimer that unreleased services may not become generally available on schedule — or at all.
