Rubrik Agent Cloud
by Rubrik
AI agent discovery, runtime policy and rollback from the Rubrik data security platform
Rubrik Agent Cloud is an AI agent governance and recovery product from Rubrik (NYSE: RBRK) for security, platform and data-protection teams running agents in production. It finds the agents in an environment, enforces policies on their prompts, responses and tool calls, and uses Agent Rewind to roll back destructive agent actions to a prior healthy snapshot.
Rubrik Agent Cloud is a control layer that sits between applications, AI agents and the underlying large language models, built by Rubrik, the NYSE-listed data security company led by co-founder and CEO Bipul Sinha. It launched in October 2025 as a limited early-access waitlist organised around three functions: Agent Monitor, Agent Govern and Agent Remediate. Rubrik announced general availability on February 18, 2026. The GA release covers three areas. Monitoring scans the environment to build an inventory of agents, profiles their risk and shows which tools and data each agent can reach. Governance applies policies to agent inputs and to outputs and tool calls; buyers get predefined guardrails such as PII detection and can write custom policies in natural language, which Rubrik enforces with its own small language models. Remediation is Agent Rewind, which matches an agent's destructive action to an earlier healthy snapshot so teams can undo it, with examples such as a dropped database or deleted Salesforce opportunities. Coverage at GA names backend integrations for Microsoft Copilot Studio, support for agents built with LangChain, OpenAI and Claude, a standalone AI gateway for custom agents, and discovery for agents running on endpoints. At its Rubrik Forward conference in June 2026 the company made Agent Cloud for Claude Code and Claude Cowork generally available, adding the SAGE semantic governance engine, immutable snapshots of GitHub and Azure DevOps repositories, and backup and drift monitoring for system prompts, tool permissions and files such as CLAUDE.md. Six integrators (Cognizant, Deloitte, LTM, HCL Technologies, NTT Data and Wipro) signed on to resell and embed the Claude Code offering. Rubrik's pitch rests on recovery: most agent security tools stop at detection and blocking, and Rubrik's backup heritage lets it restore the data an agent damaged.
A CISO or head of platform security who already runs Rubrik for backup and now has agents with write access to production systems or code repositories.
One inventory of agents plus runtime policy on what they say and do, with the ability to roll back the data and repo changes an agent got wrong.
At a Glance
- Category
- Governance & Security
- Pricing
- Contact for pricing
- Target Market
- CISOs, CIOs, Security Engineers, Platform Engineering Teams
- Deployment
- Cloud-first
- Team Size
- 500+
Key Features
- ✓Agent discovery and inventory
Automatically scans the environment to build an inventory of agents, profile their risk and show which tools and data they access.
- ✓Prompt, response and tool-call policies
Enforces predefined guardrails such as PII detection and custom natural-language policies on agent inputs and outputs, using Rubrik small language models.
- ✓Agent Rewind
Matches a destructive agent action to a prior healthy snapshot so teams can undo changes such as a dropped database.
- ✓Codebase resilience for Claude Code
Keeps immutable snapshots of GitHub and Azure DevOps repositories outside the repo, so force-pushes or deleted branches can be rolled back.
- ✓Agent configuration backup and drift monitoring
Backs up and watches system prompts, tool permissions, skills and files such as CLAUDE.md for unexpected changes over time.
- ✓AI gateway and endpoint discovery
A standalone AI gateway covers custom-built agents, and discovery support finds agents running directly on employee endpoints.
Use Cases
- •Undoing a destructive agent action
An agent deletes records or drops a table, and the team uses Agent Rewind to restore the last healthy snapshot.
- •Governing coding agents
Engineering security puts Claude Code and Cowork under policy and keeps immutable repo snapshots to recover from bad force-pushes.
- •Building an agent inventory
Security teams discover agents across Copilot Studio, custom builds and endpoints to see which data and tools each can reach.
- •Blocking sensitive data in prompts
PII detection and custom natural-language policies stop agents from sending or returning regulated data through prompts and tool calls.
Ideal For
Best For
- ✓Existing Rubrik Security Cloud customers adding governance for AI agents
- ✓Teams running Claude Code or Claude Cowork against production repositories who need repo snapshots and rollback
- ✓Enterprises with agents built in Microsoft Copilot Studio that need an agent inventory and access visibility
- ✓Security teams that want natural-language guardrails such as PII detection on agent prompts, responses and tool calls
- ✓Organisations that need an audit trail of agent actions for compliance reviews
Not Ideal For
- ✗Buyers who want published self-serve pricing; no list price for Agent Cloud appeared in any source reviewed
- ✗Teams with no Rubrik footprint, since Agent Rewind depends on Rubrik snapshots of the affected data
- ✗Buyers who need independent production reviews before purchase; no G2 or Hacker News discussion of Agent Cloud was found as of November 2026
Deployment
Market Analysis
Pros
- ✓Combines discovery, runtime policy and recovery in one product
- ✓Builds on Rubrik's existing backup and data security footprint, so current customers add a module instead of a new vendor
- ✓Rollback for coding-agent mistakes on GitHub and Azure DevOps repositories is uncommon among agent security tools
Cons
- ✗No published pricing or packaging in any GA coverage reviewed
- ✗Remediation value depends on Rubrik already snapshotting the affected systems
- ✗Coverage is largely vendor-sourced; one iTWire piece ran as an advertisement and no independent user reviews were found
- ✗Parts of the wider Rubrik AI line remain in preview, and Rubrik's own safe-harbour language says announced features may not ship on schedule
Pricing
Rubrik Agent Cloud
Contact for pricing
- ✓Agent discovery and inventory
- ✓Runtime policy on prompts, responses and tool calls
- ✓Agent Rewind remediation
None of the GA coverage reviewed (February and June 2026) published a price, metering unit or packaging for Agent Cloud. Expect a sales-led quote, and expect Agent Rewind to require Rubrik snapshot coverage of the affected systems, which is licensed through Rubrik's wider platform.
Security & Compliance
Sources
This page was written from 4 sources, 4 on domains other than rubrik.com.
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Pretxt
Runtime AI agent supervision that enforces plain-language business policies in milliseconds, without an LLM judge
Portal26
AI adoption management platform for shadow AI discovery, AI governance and Claude usage control
Akeyless Agentic Runtime Authority
Intent-based runtime access control for AI agents: credentialless access, per-action policy checks and a session kill switch
Darktrace / SECURE AI
Behavioral monitoring of enterprise AI use: shadow AI blocking, prompt risk analysis and AI agent identity mapping