Team: Huntress Managed Endpoint Security Posture Management (ESPM)
Product: ESPM
Environment: Eligible Windows endpoints
Summary: Learn how RMM Guard safely progresses from Learning Mode through Blocking Preview to full Blocking Mode without disrupting legitimate administrative tools.
In This Article
Overview
Phase 1: Learning Mode
- Instance-Level Tool Identification
Phase 2: Blocking Preview Mode
- Transition Requirements
- Reviewing Tools in Preview
Phase 3: Blocking Mode
- Enabling Blocking Mode
- Execution & Enforcement Logic
- Unrejecting a Tool
Adding New Devices and Organizations
Enabling, Disabling, or Reverting RMM Guard
Overview
RMM Guard prevents unauthorized remote access and administration tools from executing while ensuring verified management software continues to run unimpeded. To avoid operational disruption, blocking does not activate immediately. Instead, RMM Guard uses a progressive multi-stage rollout:
Learning Mode: Passive observation and discovery.
Blocking Preview Mode: Telemetry simulation ("Would Block") to validate policy definitions.
Blocking Mode: Active enforcement on verified endpoints.
Important: RMM Guard Blocking Mode requires manual administrative activation once readiness criteria are satisfied. For prerequisite system configurations, see Huntress Managed ESPM Readiness and Requirements.
Phase 1: Learning Mode
When RMM Guard is initially enabled, it operates strictly in Learning Mode. In this phase, the Huntress agent scans endpoints to build a comprehensive inventory of all active Remote Monitoring and Management (RMM) and remote access utilities.
Enforcement: Disabled.
Agent Behavior: Passive observation and telemetry ingestion only. No processes are interrupted, blocked, or altered.
Instance-Level Tool Identification
RMM Guard identifies specific tool instances, rather than relying solely on file names or installation paths. For example, with utilities like MeshCentral, RMM Guard parses the specific server domain or endpoint the agent connects to. This allows organizations to approve their own managed deployment while rejecting unauthorized or rogue instances of the identical binary.
Viewing Instance Telemetry: Navigate to ESPM > RMM Guard, select a tool, and open the details panel. The Instances section provides a read-only breakdown of observed endpoints and connection profiles.
Granular Rejection: Rejecting a specific instance blocks only that target configuration, even if it shares a directory or executable path with an approved instance.
Phase 2: Blocking Preview Mode
Blocking Preview serves as a non-disruptive rehearsal. The Huntress agent evaluates processes against approval rules and records what would have been terminated without intercepting live execution.
Transition Requirements
Your account enters Blocking Preview when all of the following are true:
- at least one RMM tool is marked Approved (at the account or any organization level);
- no RMM tool that is new to the account has been seen in the last three days;
- RMM Guard has been active on the account for at least three days. An account that was previously reverted out of Preview or Blocking does not re-enter Preview automatically.
Reviewing Tools in Preview
Use this phase to audit your software inventory in ESPM > RMM Guard:
| Tool Status | Definition | Operational Impact |
| Approved | Software actively maintained and authorized by your IT/security staff. | Permitted to run without interruption. |
| Rejected | Software unauthorized for business use. Decisions are queued. | Logged as Would Block; execution continues until Blocking Mode is enabled. |
| Unreviewed | Newly observed or pending classification. | Monitored without disruption until classified. |
Note: Any RMM tool detected for the first time after entering Blocking Preview is classified as blocked by default and requires administrative review before execution is allowed.
To inspect simulated blocks:
Navigate to ESPM > RMM Guard.
Select Executions Blocked > View all blocked events.
Filter by Would Block to review actions before enforcing live restrictions.
On the Blocked and Would Block lists you can also filter by All, Rejected, or Auto Rejected, and choose a 7, 14, or 30 day period.
Phase 3: Blocking Mode
Once your inventory is audited and at least one endpoint meets the readiness requirements, the Setup Blocking button becomes available under ESPM > RMM Guard.
Enabling Blocking Mode
Navigate to ESPM > RMM Guard.
Select Setup Blocking.
Confirm deployment across eligible devices.
Staged Activation: Blocking is applied on a per-endpoint basis. Only endpoints that meet all policy and health prerequisites start blocking immediately. Devices still undergoing stability soak cycles will begin blocking automatically as they qualify.
Execution & Enforcement Logic
Once Blocking Mode is active, execution is evaluated according to the following logic:
Approved Tools: Continue running normally without operational impact.
Rejected Tools: Blocked immediately upon execution.
Auto-Rejected Tools: Tools first seen on an endpoint after Blocking Preview or Blocking was enabled are rejected automatically, shown as Rejected (Auto), and blocked until you approve them or confirm the rejection.
-
Pre-Existing Unreviewed Tools: Discovered prior to activating Blocking Mode are generally permitted to minimize disruption. However, execution is blocked if:
The file runs from high-risk paths (e.g.,
\Downloads,\Temp,\AppData\Local\Temp).The binary shares attributes (such as digital signing certificates) with a previously rejected tool.
New Tools Discovered After Blocking is enabled: Blocked by default until reviewed and explicitly approved by an administrator.
Unrejecting a Tool
If an administrative utility was marked as rejected or auto-rejected in error:
Navigate to ESPM > RMM Guard.
Select the tool to open its slide-out details panel.
Select Unreject to return the tool to Unreviewed. If Huntress auto-rejected the tool, you can instead select Confirm Rejection to keep it rejected as your own decision.
Adding New Devices and Organizations
Endpoints and organizations added after Blocking Mode is enabled adopt security policies gradually to prevent unexpected outages:
Policy Inheritance: New endpoints immediately inherit all existing account-wide approval rules.
-
Readiness Soak Period: The device enters an onboarding evaluation phase, requiring:
A system reboot following initial policy deployment.
A 14-day baseline monitoring period.
A 7-day stability window with zero new RMM detections and zero "Would Block" events.
Automatic Cutover: When all criteria are met, the endpoint automatically transitions into live Blocking Mode without manual intervention.
Manual Override: If you don't want to wait for an endpoint to meet these requirements, an account admin can override readiness for that endpoint from its Blocking Readiness panel. The failed checks remain visible.
Enabling, Disabling, or Reverting RMM Guard
Scope and Exclusions
Enrollment Scope: Configure managed coverage under ESPM > Settings. ESPM can be enabled account-wide, with granular inclusion or exclusion overrides applied at the Organization or Endpoint level.
Enforcement Scope: Blocking mode is an account-wide setting. Which organizations and endpoints are covered is controlled by the enrollment overrides under ESPM > Settings (See above). Once Blocking is on, an account admin can revert or resume blocking on a single endpoint from the Endpoint Coverage grid. Approve and Reject decisions can also be made for a specific organization, and an organization's own decision takes precedence over the account's..
Reverting to Learning Mode
If unexpected behavior occurs or widespread audits are required:
Navigate to ESPM > RMM Guard.
Locate the Protection Status column.
Select Revert to Learning.
An endpoint stops blocking once it receives the updated policy, and Huntress retries any revert that doesn't report back as confirmed. Existing software classifications, instance records, and approval selections remain preserved.
To stop blocking on only one endpoint, an account admin can revert it from the Endpoint Coverage grid. If the account is still in Preview or Blocking, that endpoint goes back to Blocking Preview, not Learning.