Manage Human Approval Requests
Overview
Human Approval is a Rule Behaviour option for guardrail policies. When the policy is triggered, Akto blocks the attempt and sends it to a reviewer instead of only alerting after the fact or blocking permanently. The request is routed to the approvers you designate, who approve or deny it. Approving the specific actor, always or for a fixed number of days, allows future attempts through. Every decision is recorded in the approval history.
How It Works
You create or edit a guardrail policy and set Rule Behaviour to Human Approval.
The next time that policy's condition is triggered, Akto blocks the attempt, the same as a Block rule would.
The event is logged to the Needs Approval tab in Guardrail Activity, instead of Active.
The request is routed to the designated approvers configured for the policy.
An approver opens the event and chooses Approve or Deny.
On Approve, the approver chooses how long the approval lasts: Always, or a specific number of days. For that duration, Akto stops enforcing that policy against the approved actor. Once the window expires, enforcement resumes and the next attempt is blocked and logged to Needs Approval again.
On Deny, enforcement stays in place and the actor's attempts keep being blocked.
Each decision is added to the approval history, so you can see who approved or denied a request and when.
Set a Policy's Rule Behaviour to Human Approval
Navigate to Guardrails → Guardrail Policies.
Open an existing policy, or select Create Guardrail to configure a new one.
Go to the Scope section and locate Rule Behaviour.

Select Human Approval.
Save the policy using Create Policy or Update Guardrail.
For the rest of the policy configuration, see Create Guardrail Policies.
Review a Human Approval Request
Navigate to Guardrails → Guardrail Activity.
Select the Needs Approval tab. This lists every blocked attempt that is waiting on a reviewer decision.
Review the event's columns for context:
Severity
Risk level assigned to the detection
Actor
The user, host, or endpoint that triggered the policy
Username
The signed-in user associated with the attempt, if available
Policy Triggered
The guardrail policy that blocked the attempt
Detection Type
The layer the violation was detected at (e.g. Session)
Rule Violated
The specific rule within the policy that matched (e.g. BlockedHost)
Behaviour
Shows Human Approval for events in this tab
Compliance
Compliance frameworks associated with the violated rule
Collection
Associated data collection
Detected
Timestamp when the attempt occurred
Select Approve or Deny in the Action column for the event. Only the designated approvers for the policy can act on the request.
In the confirmation dialog, review the actor named (e.g. a specific server or domain) and the policy it will bypass.
Under Approve for, choose:
Always – Permanently stop enforcing this policy against this actor.
Number of days – Enter a duration. Enforcement resumes automatically once it elapses.

Select Approve to confirm.
Track Approval History
Every approve and deny decision is recorded, so you can audit who handled each request, what they decided, and when. Use the approval history to review past decisions and confirm that approvals are being granted for an appropriate duration.
What's Next
Create Guardrail Policies: configure the full set of detection and scope settings for a policy.
Guardrail Activity Detailed View: review the full event context available for any guardrail activity, including Needs Approval requests.
Last updated