Overview

Agentic Security helps you keep track of the AI agents on your devices. It finds the AI agents you have installed, flags settings that weaken their safeguards, and logs what agents like OpenClaw do on your machine. You can also monitor AI coding agents, such as Claude Code, Codex, and Cursor.

This guide walks you through setting up Agentic Security as a consumer user.

If your organization already uses Swif for device management (MDM), you can turn on Agentic Security from your existing Swif workspace, with no separate sign-up. See Getting Started with Agentic Security for Enterprise.

What you'll need

A computer running macOS, Windows, or Linux
An internet connection
About 5 minutes

Supported platforms

Platform	Installer	Architectures
macOS	Signed and notarized `.pkg` package (universal)	Apple silicon and Intel
Windows	Signed `.msi` package	x64 and ARM64
Linux	`.deb` package (Debian, Ubuntu) or `.rpm` package (RHEL, Rocky Linux, Fedora)	x64 and arm64

Standalone installers are also available: a .zip for macOS, an .exe for Windows, and a binary for Linux.

Set up Agentic Security

Step 1: Create your account

Global: app.swif.ai
EU: app.eu.swif.ai

During sign-up, you'll answer a few questions, such as your name and how you plan to use Agentic Security. Fill in the required fields, and click Continue.

Step 2: Install the Agentic Security agent

After the sign-up questions, you'll be asked to install the Agentic Security agent.

Select your installer. Choose your operating system from the dropdown menu.
Download the installer. Click Download Installer. The button is available once you've selected an installer.
Run the installer. It installs the agenticsecurity-install command on your device.

OS	How to install
macOS	Double-click the `.pkg` file, and follow the steps.
Windows	Double-click the `.msi` file, and follow the steps.
Linux	Run `sudo apt install ./<package>.deb`, or `sudo dnf install ./<package>.rpm`

Enroll your device. Open Terminal (macOS or Linux) or an administrator Command Prompt (Windows), and run the command below with your details.

Your Team ID is shown in Settings > Team.

macOS and Linux:
```
sudo agenticsecurity-install -teamId=<TEAM_ID> -name=<FIRST_NAME> -surname=<LAST_NAME> -email=<EMAIL>
```
Windows:
```
agenticsecurity-install -teamId=<TEAM_ID> -name=<FIRST_NAME> -surname=<LAST_NAME> -email=<EMAIL>
```
All four values are required. The command enrolls your device, and starts the Agentic Security agent in the background.
Confirm installation. Click I've Installed Agentic Security to continue.

Tip: To set up Agentic Security later, click Skip Setup to go to your dashboard. You can install the agent from the dashboard at any time.

Check that your device is enrolled

OS	Run this	You should see
macOS	`sudo launchctl print system/com.swifteam.agentsecurity-agent \| grep state`	`state = running`
Windows	`Get-Service STAgentSecurity` (in PowerShell)	`Running`
Linux	`systemctl status agentsecurity-agent`	`active (running)`

Your device also appears in your Agentic Security device list.

If enrollment fails

"team … is not valid": check your Team ID, then run the command again.
Already enrolled: running the installer again doesn't enroll the device twice, or create a duplicate.
Logs:
- macOS: /var/log/agentsecurity.log
- Windows: C:\Program Files\AgentSecurity\logs\log.txt
- Linux: /var/log/agentsecurity.log, or journalctl -u agentsecurity-agent

Step 3: Turn on your policies

Agentic Security uses policies to decide what to monitor on your device. In this step, you turn on the required policies with a toggle. No other setup is needed.

Policy	What it does
Swif Agent Inventory Policy	Finds autonomous AI agents on your device, such as OpenClaw, Hermes, and Grok Bot. Reports how they're set up, and uploads OpenClaw's event logs.
Swif OpenClaw Tools Access Policy	Controls which tools OpenClaw can use, starting with a minimal profile

Turn on the required policies, and click Continue. The button becomes available once the required policies are on.

Step 4: See your AI agents

Once your policies are on, open your device in the Agentic Security device list to see:

Installed AI agents: OpenClaw, Hermes, Grok Bot, Claude Code, Codex, and Cursor, with their versions, and whether they're running (on Windows, OpenClaw and Hermes don't show running status).
Risky settings: for example, an OpenClaw gateway exposed to the network with no authentication.
OpenClaw events: under Logs. Events upload about every minute. The Source column shows which part of OpenClaw each event came from, and you can filter events by keyword and date.
Device details: encryption status, OS version, and last connected time.

OpenClaw event logs come from the Log Agents setting in the Swif Agent Inventory Policy. It's set to openclaw when you turn on the policy during setup.

Optional: Monitor Claude Code, Codex, and Cursor

To also see what AI coding agents do on your device, such as their sessions, prompts, and tool use, turn on the Swif AI Agent Activity Monitoring Policy. It monitors:

Claude Code
Claude Desktop
Codex
Cursor
Cline
Warp

To turn it on:

Open your device in the Agentic Security device list.
Go to the Policies tab.
Find Swif AI Agent Activity Monitoring Policy, and click Create Policy.
Turn on Enabled, and save.

Sessions appear on the device within about 15 minutes.

Intel Macs: AI agent activity monitoring hasn't been verified on Intel Macs yet.

This policy collects prompts and responses. Session records can include your prompts, the agent's responses, project paths, and tool results. Turn it on only if you're comfortable with that data being stored in Swif. See the policy's article for details.

Manage your policies

You can create and assign policies from two places on a device.

From the Activity tab

Open an Agentic Security-enrolled device from the device list.
Go to the Activity tab.
If no policy exists, you'll see a Missing Policy state.
Click Create/Assign Inventory Policy.
The tab refreshes to show the new policy status.

From the Policies tab

Open an Agentic Security-enrolled device.
Go to the Policies tab.
If a policy shows Not Configured, click Create Policy.
The status changes to Not Assigned or Enabled, depending on your setup.

Status	Meaning	Action
Not Configured	The policy hasn't been created yet	Click Create Policy
Not Assigned	The policy exists, but isn't applied to this device	Click Assign Policy
Enabled	The policy is active	Click Configure Policy to change it

What you can do with Agentic Security

Feature	What it gives you
AI agent inventory	See which AI agents are installed, such as OpenClaw, Hermes, and Grok Bot, and how they're set up
Risky setting detection	Spot settings that weaken an agent's safeguards
OpenClaw event logs	Review what OpenClaw did, in near real time, filtered by keyword and date
AI coding agent monitoring	With the AI Agent Activity Monitoring Policy, review Claude Code, Codex, and Cursor sessions
Sensitive data redaction	API keys, passwords, and credentials are detected and hidden before they appear in your logs
CLI privilege protection	OpenClaw and other monitored CLI agents can't run with `sudo` or root. See below.
Device overview	Encryption status, OS version, and last connected time
Policy management	Create and manage policies from the Configuration tab

Automatic agent updates

The Agentic Security agent updates itself on macOS, Windows, and Linux. When a new version comes out, enrolled devices download and install it in the background, so you always have the latest protections and fixes.

Hands-free: you don't need to download or run a new installer.
Verified updates: update packages are checked and signed before they're installed, so only genuine Swif releases go onto your device.
No interruption: your policies and CLI privilege restrictions keep working after an update.

Check your agent version (macOS). Run this in Terminal:

/usr/local/agentsecurity/agentsecurity --version

If it still shows an older version, make sure the device is online. Then wait for the next update check, or restart the device.

CLI agent privilege restriction and root protection

Agentic Security stops CLI agents, such as openclaw, from running commands with root or sudo privileges.

Protection	What it means
Sudo and root blocking	Running monitored CLI agents with elevated privileges is blocked, including `sudo openclaw`, full binary paths (such as `sudo /opt/homebrew/bin/openclaw`), root shells (`sudo -s`), and preserved environments (`sudo -E`).
Normal use keeps working	AI agents keep working with standard user permissions, so your normal workflows aren't disrupted.
Only AI agents are restricted	Your other admin commands, such as `sudo ls` or `sudo cat /etc/hosts`, work as usual.
Persistent	Protection stays on across restarts and CLI updates.
Every user	Restrictions apply to all non-root accounts on the computer.
Logged	Blocked `sudo` attempts are logged on the device, in `/usr/local/agentsecurity/`, and shown on the device's Activity tab.

Remove the Agentic Security agent

To stop monitoring a device, without deleting it:

Open the Agentic Security device list.
Find the device, and open its action menu.
Click Remove Agentic Security Agent.
Confirm in the window, after checking the device name and serial number.
The device's Agent Status changes to Inactive.

Note: Uninstalling the installer package from the device removes only the agenticsecurity-install command. The agent keeps running. Use Remove Agentic Security Agent to stop monitoring.

Delete an Agentic Security device

To permanently remove a device from Agentic Security:

Open the Agentic Security device list.
Find the device, and open its action menu.
Click Delete Device.
Confirm the deletion.
The device is removed from the list, and no longer appears in search or filter results.

Frequently asked questions

Which AI tools does Agentic Security monitor?

Tools	Covered by
OpenClaw, Hermes, and Grok Bot	The Swif Agent Inventory Policy, which you turn on during setup. It also uploads OpenClaw's event logs.
Claude Code, Claude Desktop, Codex, Cursor, Cline, and Warp	The Swif AI Agent Activity Monitoring Policy, which you can turn on separately. See Optional: Monitor Claude Code, Codex, and Cursor.

What's the difference between removing the agent and deleting the device?

Removing the agent sets the device to Inactive, but keeps the device record. Deleting the device removes it from Agentic Security entirely.

Can I use Agentic Security on BYOD devices?

Yes. Agentic Security supports both company-managed and personal (BYOD) devices.

Can I skip the installation during setup?

Yes. Click Skip Setup in Step 2 to go to your dashboard, and install the agent later.

What happens if I don't turn on the policies?

Agentic Security doesn't inventory or monitor AI agents on your device. You can turn the policies on at any time from your dashboard's Configuration tab.

Why don't I see Claude Code or Cursor sessions?

These come from the Swif AI Agent Activity Monitoring Policy, which isn't turned on during setup. Turn it on, then wait about 15 minutes. See Optional: Monitor Claude Code, Codex, and Cursor.

Does Agentic Security rate each action as safe or risky?

No. Agentic Security flags risky settings, such as an exposed OpenClaw gateway, but doesn't rate individual events. Review events under Logs, and filter them by keyword and date.

Is Agentic Security the same as the Swif MDM agent?

No. Agentic Security is a separate product, focused on monitoring and securing AI agents. It doesn't include enterprise MDM features, such as remote lock, wipe, or fleet management. For those, upgrade to the full Swif workspace from the Plans page.

How do I install Agentic Security on other devices?

From your Agentic Security dashboard, click Setup Installer. You can also share the installation guide by entering people's email addresses.

Can I run OpenClaw or other CLI agents with sudo or root?

No. For your safety, Agentic Security stops monitored CLI agents from running with sudo or in root shells, to prevent unintended system changes. Run them as a standard user.

Will Agentic Security block my regular sudo commands?

No. The restriction applies only to monitored AI CLI tools. All your other sudo commands keep working.

Where can I see blocked privilege escalation attempts?

They're logged on your device, in /usr/local/agentsecurity/, and shown on the device's Activity tab.

Do I need to update the Agentic Security agent by hand?

No. The agent updates itself on macOS, Windows, and Linux. You only need to reinstall it if you removed it from the device.

Need help?

If you run into any issues during setup, contact Swif support at support@swif.ai.

Getting Started with Swif's Agentic Security (for Consumer)