Rubrik Launches Agent Cloud to Govern Claude Code Deployments
Rubrik Agent Cloud adds governance, immutable repo snapshots, and an "Agent Rewind" feature for Claude Code and Claude Cowork. Targets organizations where AI agents can commit, push, and deploy without human approval.
Rubrik shipped a purpose-built governance layer for Anthropic's Claude Code and Claude Cowork. Rubrik Agent Cloud addresses the security gap between traditional DevSecOps controls and AI agents that can modify code, push commits, and deploy at machine speed without a human in the loop.
What it does
The service has four components:
- Governance engine — real-time policy enforcement on agent actions. Define what agents can access and what they can do.
- Agent inventory — visibility into all deployed agents, their permissions, risk profiles, and policy violations across the organization.
- Agent Rewind — roll back unintended agent actions. Undo a bad commit, a destructive refactor, or a rogue deployment.
- Immutable repo snapshots — backups of GitHub and Azure DevOps repos stored outside the repository environment. Recover from deleted branches or overwritten history.
It also backs up Claude's agent configuration itself — system prompts, tool permissions, skills, and key files at org, repo, and user levels. Configuration drift monitoring flags unauthorized changes before they spread.
Why it exists
The product exists because enterprises are adopting Claude Code faster than their security teams can build controls. Rubrik's CPO Anneka Gupta said: "Organizations are adopting Claude faster than any agentic technology we have seen, and every security leader asks the same question: how do we stay in control when an agent can act?"
The threat model includes rogue commits, repo ransomware, prompt injection, and
IP exfiltration. When an agent can push to main without review, those threats
compound.
The bottom line
Agent governance is becoming its own product category. Rubrik Agent Cloud validates that enterprises aren't just experimenting with Claude Code. They're deploying it widely enough to need enterprise-grade recovery, audit, and policy controls. If your team is running Claude Code agents in production, the question isn't whether you need governance. It's whether you're building it yourself or buying it.