Skip to content
MoftgateMoftgate Docs
Product guide

Automations

Build multi-step automation pipelines that run actions across your endpoints - on a schedule, in response to events, or on demand.

What Are Automations#

An automation pipeline: actions run in sequence on the selected endpoints, started by a schedule, an event, or a manual run.
Example automation pipeline
Close image

Automations let you chain a series of actions into a single pipeline that runs across one or more endpoints. Instead of performing each task by hand - restarting a device, installing updates, sending a notification - you define the steps once and let the pipeline run them in order, automatically.

A pipeline is made up of actions. Each action does one thing - restart a device, run a script, send an email - and the pipeline runs them sequentially. You decide how the pipeline starts: on a schedule, when an event occurs, or only when someone runs it manually.

  • Build multi-step pipelines of actions that run on your endpoints
  • Mix device power control, security scans, app installs, and notifications in one flow
  • Trigger automatically on a schedule or in response to an event
  • Run manually from any endpoint detail page
  • Every step records its result so you can audit what happened

Creating a Pipeline#

Creating an automation pipeline takes a few minutes. You name the pipeline, add the actions you want in the order they should run, and choose how it should be triggered.

  1. 1
    Open the Automations section

    Navigate to Automations from the sidebar. You will see a list of existing automation pipelines with their status, action count, and trigger type.

  2. 2
    Create a new pipeline

    Click the "New Automation" button. Give the pipeline a name and an optional description so your team can recognize it later.

  3. 3
    Add actions to the pipeline

    Add one or more actions in the order you want them to run. Each action is added through a three-step wizard (see below). Actions run sequentially - the next one starts only after the previous one completes.

  4. 4
    Choose a trigger

    Decide how the pipeline should start: on a fixed schedule, in response to an event, or only when someone runs it manually. You can leave a pipeline without a trigger if it will always be started by hand.

  5. 5
    Save and activate

    Review the pipeline, then save it. Active pipelines run automatically according to their trigger. Archived pipelines are hidden and never run.

Order matters
Actions run one after another. The next action only starts once the previous one finishes successfully (or continues on failure if you allowed it). Put dependent steps in the right order - for example, install an app before notifying the user that it is ready.

Action Types#

There are 26 action types organized into five categories. When you add an action to a pipeline, a three-step wizard guides you through choosing the type, configuring it, and setting failure behavior.

The three-step wizard

  1. 1
    Choose Action Type

    Browse a category-grouped card grid. There are five categories - Device, Security, App Management, Notifications, and Other. Each card shows an icon, label, short description, and an "Agent" badge for actions that require the endpoint agent to be online.

  2. 2
    Configure Action

    Give the action a display name and fill in the type-specific fields. Depending on the action, these may include a script, package name, message text, URL, delay in minutes, scan type, SMS phone number, or email fields.

  3. 3
    Conditions & Failure Behavior

    Optionally add conditions that must be met for the action to run. Choose what happens on failure - for example, whether the pipeline should stop or continue - and set a retry count. A three-bar progress indicator at the top of the wizard shows which step you are on.

Edit mode skips step 1
When you edit an existing action, the wizard opens directly on step 2 because the action type cannot be changed after creation.

Category summary

CategoryActionsCovers
Device7Power control, group assignment, maintenance mode, tags, deletion
Security5OS updates, Defender update and scan, disk encryption
App Management5winget, Linux packages, Homebrew, Microsoft 365
Notifications5On-screen toast, approval gate, SMS, email, email report
Other4Run script, HTTP request, custom field, delay

Device

ActionRuns onDescription
Restart DeviceAgentRestart the device through the endpoint agent
Shutdown DeviceAgentShut down the device through the endpoint agent
Assign to GroupServer-sideMove the endpoint into a different group
Enable / Disable Maintenance ModeServer-sideToggle the endpoint maintenance mode on or off
Delete DeviceServer-sideRemove the endpoint from management entirely
Apply TagServer-sideAdd a tag to the endpoint
Remove TagServer-sideRemove a tag from the endpoint

Security

ActionRuns onDescription
Install OS UpdatesAgentInstall all available Windows, macOS, or Linux updates
Windows Defender UpdateAgentUpdate Defender signature definitions (Windows only)
Windows Defender ScanAgentRun a quick or full Defender scan (Windows only)
Enable / Disable Disk EncryptionAgentTurn BitLocker (Windows), FileVault (macOS), or LUKS (Linux) on or off
Rotate Disk Encryption KeyAgentGenerate a new disk encryption key

App Management

ActionRuns onDescription
Install / Uninstall / Upgrade winget PackageAgentManage Windows packages via winget
Install / Uninstall / Upgrade Linux PackageAgentManage Linux packages via apt, dnf, or yum
Install HomebrewAgentInstall the Homebrew package manager on macOS
Install / Upgrade / Uninstall Homebrew PackageAgentManage macOS packages via Homebrew
Install Microsoft 365AgentInstall Microsoft 365 on Windows or macOS

Notifications

ActionRuns onDescription
Notify UserAgentShow a toast message on the endpoint screen
User ApprovalServer-sidePause the pipeline until a person approves it
Send SMSServer-sideSend an SMS message to a phone number
Send EmailServer-sideSend an email notification
Send Email ReportServer-sideEmail an endpoint report to recipients

Other

ActionRuns onDescription
Run ScriptAgentRun a library or custom script on the endpoint
HTTP RequestAgentSend an HTTP request to an external webhook or API
Set Custom FieldServer-sideUpdate a custom field value on the endpoint
DelayServer-sideHold the pipeline for a specified duration before continuing
Operating-system limits
Some actions only work on a specific operating system. Windows Defender actions run only on Windows, Homebrew actions only on macOS, and Linux package actions only on Linux. Sending an incompatible action to an endpoint will fail.

Triggers#

A trigger decides how and when a pipeline starts. You can choose a schedule, an event, or no automatic trigger at all.

Trigger typeHow it works
ScheduleThe pipeline runs on a recurring schedule - for example, every day at a set time, or weekly on a chosen day. Use this for routine maintenance like installing updates or running security scans on a cadence.
Event-basedThe pipeline runs automatically in response to an event - such as a monitoring alert being triggered. This lets you react to conditions on your endpoints without manual intervention.
ManualThe pipeline has no automatic trigger and only runs when someone starts it by hand. You can run any automation manually from the endpoint detail page by selecting it and confirming "Run Now".
No trigger is fine
A pipeline without a trigger is perfectly valid - it simply waits to be started manually. This is useful for one-off maintenance tasks you only run on demand.

Run History#

Every time a pipeline runs - whether automatically or manually - the result is recorded. You can review what ran, when it started, how long it took, and whether each action succeeded or failed.

  • Each pipeline run is recorded with its start time, total duration, and overall status (success or failed).
  • Within a run, every individual action has its own result - running, succeeded, or failed - along with any output or error message.
  • On the endpoint detail page, the Automations tab shows available automations as a card grid and lists the 20 most recent runs for that endpoint.
  • The global Automations page lists runs across all endpoints, so you can see trends and spot recurring failures.
Per-endpoint view
On any endpoint detail page, the Automations tab shows the automations available for that endpoint and the 20 most recent runs on it. Click an automation card and confirm "Run Now" to trigger it immediately.

Edit Mode Behavior#

Editing an existing action is safe by design. Your changes are kept locally until you explicitly save them, so you can experiment without affecting the live pipeline.

  • When you edit an existing action, the wizard skips the type-selection step and opens directly on the configuration step, since the action type cannot be changed after creation.
  • Changes you make in edit mode are kept locally in your browser until you press "Done". Nothing is saved to the server until you explicitly confirm.
  • If you navigate away or close the wizard without pressing "Done", your local edits are discarded and the previously saved version remains unchanged.
  • Pressing "Done" commits all local changes at once, so the pipeline only ever reflects either the fully saved state or your complete set of edits.
Done saves everything at once
There is no partial save. Until you press "Done", the server still holds the previous version. After "Done", the pipeline reflects all your edits together.

FAQ#

Can I run an automation on a single endpoint without setting up a trigger?
Yes. Open the endpoint detail page and go to the Automations tab. You will see all available automations as cards. Click one and confirm "Run Now" to run it immediately on that endpoint, regardless of any trigger it may have.
What happens if an action in the middle of a pipeline fails?
It depends on the failure behavior you configured for that action. You can choose to stop the pipeline or let it continue to the next action. You can also set a retry count so the action is attempted again before it is marked as failed.
Do agent actions require the endpoint to be online?
Yes. Actions that run on the endpoint - such as restart, script execution, package install, or security scan - require the endpoint agent to be online so it can receive the command and report back. Server-side actions like email or group assignment do not need the agent.
Why does the edit wizard start on step 2 instead of step 1?
Because the action type is chosen when the action is first created and cannot be changed afterward. Editing an existing action only lets you adjust its configuration and failure behavior, so the wizard skips the type-selection step and opens directly on configuration.
Are my edits saved automatically?
No. Edits you make in the wizard are kept locally in your browser until you press "Done". If you navigate away or close the wizard without pressing "Done", your changes are discarded and the saved version remains unchanged.
What is the difference between an active and an archived automation?
Active automations run automatically according to their trigger and appear in the available automations list on endpoint pages. Archived automations are hidden, never run automatically, and cannot be run manually until they are restored.