Technology

86% of cybersecurity leaders fear AI agents will outpace security safeguards: Rubrik

About 86 per cent of cybersecurity and IT leaders globally expect the proliferation of artificial intelligence (AI) agents to outpace their organisations’ security safeguards within the next year, according to Rubrik, a security and AI company.

The company has expanded its Project Hourglass initiative by broadening its partner alliance to deliver Rubrik Code Guardian alongside Rubrik Agent Cloud.

According to Rubrik, Code Guardian will enable partners to proactively conduct red-team analysis of client code in isolated, air-gapped environments, identify realistic attack chains and facilitate rapid codebase recovery.

“Engineering teams are turning to AI models to accelerate software delivery, but speed cannot compromise security or architectural integrity,” said Alok Agrawal, Chief Solutions Officer at Rubrik.

“By incorporating Rubrik Code Guardian into Project Hourglass, we are ensuring engineering teams are able to conduct red-team analysis, prioritise business impact and reduce risk,” he added.

Rubrik Code Guardian uses Anthropic’s Claude Mythos 5 through a specialised security framework to assess code repositories against complex, multi-step threat scenarios. Instead of testing live production environments, the system analyses cloned copies of customer repositories in air-gapped environments.

The company said the platform would offer partners three key capabilities: isolated red-team analysis, attack-chain discovery and business-impact prioritisation.

Under isolated red-team analysis, Claude Mythos 5 operates within Rubrik’s security framework. The attack-chain discovery feature examines files, services, identity roles and cloud perimeters to identify interconnected vulnerabilities that conventional static analysis tools may miss.

The business-impact prioritisation feature aims to reduce alert fatigue by verifying findings for actual exploitability and assessing them according to business criticality.

Project Hourglass was initially launched to operationalise Rubrik Agent Cloud for Anthropic’s Claude Code, introducing runtime behavioural safeguards, prompt-context protection and Agent Rewind, a feature designed to reverse unintended actions.

With the addition of Code Guardian, the initiative aims to provide an end-to-end security architecture. The programme also offers participating partners joint go-to-market resources, dedicated technical certification tracks and direct access to engineering teams to accelerate client deployments.

Related posts

2 new microscopic snail species discovered in Meghalaya’s limestone caves

Nicole A. Murphy

‘Tired of waking up at 3 am’, man creates AI ‘roommate’ to control ceiling fan while sleeping

Nicole A. Murphy

Discarded SpaceX rocket on high-speed collision course with moon

Nicole A. Murphy