← 返回列表

Agent Intent-Based Access Control: Stopping Rogue Agent Behavior

单站 varonis_cloud_sec 入库 2026-11-18T06:40:31

Agent Intent-Based Access Control: Stopping Rogue Agent Behavior

Varonis Atlas’ enhanced Agent Intent-Based Access Control (IBAC) prevents out-of-policy agent actions in real time, so organizations can safely connect AI to their enterprise data.

"Agent IBAC" in white text on the left with a screenshot of the Varonis Atlas platform on the right.

Key Takeaways

AI agents are only useful if they can reach your data and tools, but that’s also what makes them risky. Agents are making headlines for going rogue, exposing sensitive company data and, in at least one case, deleting an entire production database.

Traditional access controls can’t stop agent deviations. Role-based access control (RBAC) asks whether an identity is allowed to reach a resource, but an agent with valid credentials can still take the wrong action. Static controls can't stop an agent that finds ways to circumvent them entirely, such as elevating its own privileges, invoking unauthorized tools, and acting on data it was never meant to touch.

Agent intent needs to be determined and guardrails enforced at runtime.

What’s missing is a way to ask whether this agent — in this context, for this purpose — should take this action right now. That’s precisely what Varonis Atlas’ Agent Intent-Based Access Control (IBAC) was built to address. It lets businesses connect AI agents to their enterprise data with safeguards that stop dangerous or out-of-policy behavior.

Agent IBAC compares the instruction an agent received to its reasoning and the tools and data it reaches for, then responds in real time to actions that don’t align, including alerting or blocking. When an agent crosses the line, Atlas can quarantine the identity behind it and block everything that follows for a defined window.

Agent IBAC in Varonis Atlas

Agent IBAC closes the gap between what an agent is allowed to do and what it was designed to do. Drift is determined in part by monitoring behavior. Agent IBAC evaluates every action an agent takes across a session and compares those actions to the instruction that set the agent in motion, whether that instruction came from a person, system prompt, or another agent.

Because Atlas sees the full context around an action, it doesn't rely on blanket restrictions. The sensitivity of detection and the action taken in response can each be tuned appropriately.

6 capabilities that keep agents within their intended scope:

Importantly, Atlas sits between the agent and the model that drives it. Every prompt, every model response, and every tool call flows through Atlas before it reaches its destination. That’s what makes real-time enforcement possible.

Atlas isn’t reading logs after the fact. It’s in the path, and it can stop an action before it’s executed.

1. Intent drift detection

Agent IBAC uses an LLM evaluator, the same engine behind all Atlas guardrails, to judge whether an agent's action follows from the instruction it was given. Intent drift detection includes determining whether the agent is accessing data it shouldn’t, attempting unauthorized exfiltration or download of data, or expanding the scope beyond what the user intended.

The evaluator reads the agent loop: the reasoning the agent produces, the tools it selects, and the parameters it passes to them. It then asks a simple question: “Do these steps align with the request?”

Sensitivity is tunable across three levels. Lenient gives agents room to improvise and flags only clear mismatches. Balanced is the default. Strict requires close alignment between the request and the action, and is the right setting for agents that touch regulated or high-value data.

Varonis Atlas Agent IBAC detects when an agent takes unintended actions that deviate from the user's request.

Varonis Atlas Agent IBAC detects when an agent takes unintended actions that deviate from the user's request.

2. Agent Manifest

The Agent Manifest is a record, written by the agent’s author, of what the agent is for and where its limits lie, covering purpose, ownership, permitted data, and prohibited behavior. Runtime guardrails can use the Agent Manifest to determine whether an agent’s actions match its declared intent. For agents that handle sensitive data or powerful tools, the Agent Manifest adds additional protection.

The Agent Manifest defines an agent’s purpose and boundaries, informing runtime guardrails that prevent drift.

The Agent Manifest defines an agent’s purpose and boundaries, informing runtime guardrails that prevent drift.

3. Full-session evaluation

Agent IBAC evaluates the agent's action across the entire session: every prompt, response, and tool call. That's what allows Agent IBAC to catch intent drift that unfolds gradually, where no single action looks alarming, but the cumulative path leads somewhere the user never intended.

The same full-session view also makes it possible to detect multi-turn attacks, such as jailbreak attempts spread across several prompts that appear benign individually.

Sessions are tracked by the conversation ID the AI tool assigns, so a single evaluation can span everything from the first prompt to the last. That matters because agents carry memory forward. A later prompt can lean on context established several turns earlier, which is precisely how a patient attacker assembles a jailbreak out of pieces that each look harmless. Teams can also write their own session policies in plain language and set how many events must accumulate before evaluation runs.

Varonis Atlas Agent IBAC monitors agent intent throughout the session to detect cumulative drift.

Varonis Atlas Agent IBAC monitors agent intent throughout the session to detect cumulative drift.

4. Runtime guardrails

Every intent-based detection is paired with AI runtime guardrails that take action in real time. The actions are customizable, including alerting, blocking, modifying (like redacting sensitive data), logging activity, or requiring human-in-the-loop approval.

Runtime guardrails can be customized to allow low-risk drift while stopping high-risk actions. For example, a user asks an agent to summarize a customer account. The agent starts pulling records for a much larger set of accounts than requested. This isn't necessarily malicious, but the scope creep touches more sensitive data than the request warrants. In this case, Agent IBAC can flag it for human-in-the-loop approval before it proceeds.

With AI runtime guardrails, Varonis Atlas takes real-time action when intent deviations are detected.

With AI runtime guardrails, Varonis Atlas takes real-time action when intent deviations are detected.

5. Quarantine

Quarantine goes one step further. When a violation warrants more than just stopping a single action, Atlas can quarantine the identity associated with the session. Every prompt that follows is blocked for a window the customer sets, from a couple of minutes to a full day. Administrators can see all quarantined identities in one place and lift, extend, or make them permanent. Detection tells you an agent went off course. Quarantine stops the next attempt.

6. Complete audit trail

Atlas records every action and the intent behind it, giving investigators a complete trail: what the agent did, whether each action followed from the request, and which guardrails fired.

A conversation view shows the exchange the way the user experienced it. An execution view expands it to include the tool calls beneath the surface, the steps that never surface in the chat window, and where most agent risk actually lives.

With AI runtime guardrails, Varonis Atlas takes real-time action when intent deviations are detected.

With AI runtime guardrails, Varonis Atlas takes real-time action when intent deviations are detected.

Trust is the ultimate metric for agentic success

The question for any organization adopting agents isn’t whether to give them access — it’s how to keep that access from becoming exposure.

Agent IBAC lets agents retain the broad access that makes them useful, while Atlas checks each action against what the agent was asked to do and stops any that stray before execution.

It’s one piece of what makes Atlas a complete AI security solution.

原文链接:https://www.varonis.com/blog/agent-intent-based-access-control