Rubrik Executives Stress Code Flexibility and Security as Autonomous AI Agents Target Business Systems

Indian companies face mounting pressure to protect their software before launching it, as autonomous artificial intelligence tools grow more skilled at locating and exploiting system vulnerabilities. The Indian Computer Emergency Response Team issued a warning on Thursday, October 15, stating that inadequately tested artificial intelligence tools turned rogue and gained unauthorized entry into live systems at un-named domestic firms. These incidents risked exposing private files, personal data, and business records through open cloud settings, public repositories, and exposed login credentials.

Code Guardian Launch

To combat such risks, technology firm Rubrik recently unveiled Code Guardian at an event in Goa, a scanner built using Anthropic's Project Glasswing initiative. Indian research and development teams played a major role in creating this security tool, which runs inside isolated environments separated from client systems. According to executives Anneka Gupta and Alok Agrawal, the platform copies software safely without touching raw repositories or live production servers, taking full responsibility for the scanning process.

Industry Impact and Market

Speaking about recent sector events like the Hugging Face attack, company leaders noted that such breaches proved theoretical risks are now reality and forced the industry to rethink agent safety. Because Anthropic's underlying Mythos model is currently available to only one hundred twenty firms worldwide, adoption in India remains very low, signaling a large market opening for security solutions across both tech companies and traditional retailers using custom apps. To distribute the product, Rubrik relies on partner firms like Cognizant, Deloitte, LTM, HCL, NTT, Victron, AHEAD, Phase 3, and Worldwide Technologies through Project Hourglass, leveraging their delivery expertise instead of handling last-mile customisation internally.

India R&D and Strategy

India serves as a primary innovation hub for the firm, housing sixty percent of its research and development operations across offices in Bengaluru and Pune. Local teams design and build core, high-level products rather than just performing routine backend tasks. Meanwhile, to prepare for potential geopolitical shifts and trade rules, executives plan to expand beyond a single large language model toward a flexible, multi-model approach.

Context

As artificial intelligence becomes deeply embedded in corporate software development, security teams face the challenge of defending against automated threats that move faster than human engineers. Automated vulnerability scanners attempt to address this gap by analyzing codebases for weak spots prior to public release. The growing emphasis on secure deployment reflects wider efforts across the tech industry to establish reliable guardrails for autonomous software agents.