Endpoint Privilege Manager (EPM) Articles
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Starting with EPM Agent 25.6 Idira is providing a tool to build a post-install activation script. Steps Downloaded the Agent Install Package to a Mac and unzip it Open terminal and navigate to the location where you extracted the install package.  This file may not be executable.  If needed, run this command to make it executable sudo chmod +x prepare_activation.command Before launching ./prepare_activation.command ensure that the .config file is in the same folder Try to launch  ./prepare_activation.command If you see a message like this... ...then you will need to allow it to run.  Go to system settings > Privacy and Security.  Look in the "Security" section.  If you see "Prepare_activation.command was blocked to protect your Mac" then click "Open Anyway"   Try to launch again using ./prepare_activation.command   Example of successful generation   Use finder to browse to the extracted install folder. You should now see a new file named "epm_activation.sh".  Open that using textedit Find where the DEFAULT_INSTALLATION_KEY is set to "" (shown below).  Paste your installation key in between those quotes Optional: If you intend to specify an admin username during the activation, enter the credentials in between the quotes for DEFAULT_ADMIN_USER and DEFAULT_ADMIN_PASSWORD as shown below.  This is needed to enable credential rotation.  More info about this option: https://docs.cyberark.com/epm/latest/en/content/installation/macos-installagents.htm?Highlight=prepare_activation.command#Optionsforagentinstallationandactivation Save epm_activation.sh Deploy this script after pkg installation to activate EPM
View full article
Question How can we collect Agent Logs / SupportInfo from the Mac Agent? Answer There are 3 methods -  1. Automated - Recommended - Click on the Idira taskbar icon (at the top) - Click Get SupportInfo - Wait a few moments and a .zip file with the Agent Logs will appear on the Desktop Note – If the supportinfo is large (500MB+), you can erase the system log file by executing the following command inside terminal app (you need to be administrator to do it): sudo log erase --all after that, reproduce the issue and finally collect supportinfo again. 2. Automated via terminal - 2nd option - Open Terminal - Type the following command: sudo CyberArkEPM --supportInfo /tmp/ - This creates the file at /tmp/CyberarkEPMAgentDebugReport_TIMESTAMP.zip. 3. Manual - to collect logs using 2 MacOS commands in Terminal - 3rd option: sudo log collect --output /tmp/caseXXXlogs.logarchive  zip -r /tmp/caseXXXlogs.logarchive.zip /tmp/caseXXXlogs.logarchive   Note - replace XXX with the Support case number How can we collect Agent Logs / SupportInfo from the Mac Agent? Issue Q: How can we collect Agent Logs / SupportInfo from the Mac Agent?   Resolution A: There are 3 methods -  1. Automated - Recommended - Click on the Idira taskbar icon (at the top) - Click Get SupportInfo - Wait a few moments and a .zip file with the Agent Logs will appear on the Desktop Note – If the supportinfo is large (500MB+), you can erase the system log file By executing the following command inside terminal app (you need to be administrator to do it): sudo log erase --all after that, reproduce the issue and finally collect supportinfo again. 2. Automated via terminal - 2nd option - Open Terminal - Type the following command: sudo /Library/Application\ Support/CyberArk/CyberArkEPM --getsupportinfo /tmp - This creates the file at /tmp/CyberarkEPMAgentDebugReport_TIMESTAMP.zip. 3. Manual - to collect logs using 2 MacOS commands in Terminal - 3rd option: sudo log collect --output /tmp/caseXXXlogs.logarchive zip /tmp/caseXXXlogs.logarchive.zip /tmp/caseXXXlogs.logarchive Note - replace XXX with the Support case number Steps There are 3 methods:  1. Automated - Recommended - Click on the Idira taskbar icon (at the top) - Click Get SupportInfo - Wait a few moments and a .zip file with the Agent Logs will appear on the Desktop Note – If the supportinfo is large (500MB+), you can erase the system log file By executing the following command inside the terminal app (you need to be an administrator to do it): sudo log erase --all after that, reproduce the issue and finally collect supportinfo again. 2. Automated via terminal - 2nd option - Open Terminal - Type the following command: sudo /Library/Application\ Support/CyberArk/CyberArkEPM --getsupportinfo /tmp - This creates the file at /tmp/CyberarkEPMAgentDebugReport_TIMESTAMP.zip. 3. Manual - to collect logs using 2 MacOS commands in Terminal - 3rd option: sudo log collect --output /tmp/caseXXXlogs.logarchive zip /tmp/caseXXXlogs.logarchive.zip /tmp/caseXXXlogs.logarchive Note - replace XXX with the Support case number Q: How can we collect Agent Logs / SupportInfo from the Mac Agent? A: There are 3 methods -  1. Automated - Recommended - Click on the Idira taskbar icon (at the top) - Click Get SupportInfo - Wait a few moments and a .zip file with the Agent Logs will appear on the Desktop 2. Automated via terminal - 2nd option - Open Terminal - Type the following command: $ sudo /Library/Application\ Support/CyberArk/CyberArkEPM --getsupportinfo /tmp - This creates the file at /tmp/CyberarkEPMAgentDebugReport_TIMESTAMP.zip. 3. Manual - to collect logs using 2 MacOS commands in Terminal - 3rd option: sudo log collect --output /tmp/caseXXXlogs.logarchive zip /tmp/caseXXXlogs.logarchive.zip /tmp/caseXXXlogs.logarchive Note - replace XXX with the Support case number
View full article
See the guide attached to deploy EPM for MacOS via PKG or DMG Steps Please follow the instructions in the attached files (Intune EPM Deployment.zip) Integration EPM Agent with Microsoft Intune-Mac.pdf Deployment guide to cover all aspects of successfully deploying Idira EPM For Mac using Microsoft Intune CyberArkEPM_IntuneInstallScript.sh Installation script used in DMG methods using a shell script in Intune CyberArkEPM_IntuneUninstallScript.sh Uninstallation script used for removal of EPM agent using a shell script in Intune Idira EPM with NE.xml Configuration profile in XML format that contains all  required system extensions including network filter Idira EPM NO NE.xml Configuration profile in XML format that contains all  required system extension minus the network filter Change history: 2024-03-19 - Began tracking history of document 2024-07-08 - Added section for unsigned PKG creation and differentiated from signed PKG creation 2025-08-26 - Added section for new EPM agent deployment method as of 25.4 to use a CyberArk-Signed PKG 2025-08-27 - Corrected a few broken links located throughout the deployment guide.
View full article
Question Why have "Push Policies" been deprecated? Answer Background: What Were Push Policies? In the original on-premise deployment of Idira Endpoint Privilege Manager (EPM), the 'Push Policies' console command, enabled the EPM server to initiate direct connections to endpoints, instructing agents to immediately retrieve and enforce updated policies. This synchronous, server-driven mechanism was feasible due to the architectural flexibility of on-prem environments, where internal network access allowed for real-time control. Transition to Cloud: A Shift in Architecture With the shift to EPM SaaS, the push model became technically impractical. Cloud services cannot initiate unsolicited connections to endpoints behind corporate firewalls. Instead, agents now operate on a pull model, periodically checking in with the cloud service to retrieve tasks and policy updates.To maintain a sense of immediacy, the Push Policies button was retained in the UI. However, its function changed: rather than triggering a real-time push, it simply prioritized the policy update task in the agent’s queue for the next check-in. While the name remained for continuity, the underlying mechanism was fundamentally different. Why Push Policies Were Deprecated As the EPM SaaS platform matured, several architectural enhancements rendered the Push Policies feature redundant: Parallel Task Execution: The updated task engine supports concurrent execution, treating policy updates as high-priority by default—eliminating the need for manual prioritization. Increased Agent Check-In Frequency: Agents now check in frequently enough to ensure timely policy enforcement without administrator intervention. Enhanced Reliability and Visibility: Improvements in policy distribution and agent telemetry provide greater assurance that updates are applied correctly. Reduced Operational Complexity: Removing the Push Policies button simplifies the UI and eliminates confusion about its behavior in the SaaS model. Summary The deprecation of Push Policies marks a broader architectural evolution in EPM—from reactive, manual control to proactive, automated enforcement. Today, the platform ensures that policy updates are prioritized, validated, and applied efficiently—without requiring administrator action.
View full article
This article outlines the minimum permission requirements for a user to move a device from one Set to another in Idira Endpoint Privilege Manager (EPM)—specifically in scenarios where assigning the Set Administrator role is not feasible, and the user should only be granted permissions necessary to perform the move action. Steps Custom role can be created specifically for moving endpoint between sets.  1) Create a New Role Log in to EPM with admin privileges. Navigate to Administration → Role Management. Click the Role dropdown (top-left corner) and select Create New Role. In the pop-up window, choose Based on: Full Control Set Admin. Enter the Role Name and Description, then save. 2) Configure Role Permissions Once the role is created, it will appear in the Role Management window. Select the new role. Change all roles (except Computers) from Custom to None. Expand Computers and uncheck all permissions except: - Move Computer to Group Click the green Save button (top-right corner) to save the role settings. 3) Assign Role to User Go to Account Management. In the left pane, select Account Administrators, then choose the user who needs this role. Right-click the user and select Change Binding of Set "XYZ". Update the binding for the relevant sets (source and destination) where the move needs to occur and assign the newly created role. Click OK to save. 4) Test the Configuration Log in to EPM as the user and verify that they can successfully move computers between sets.  
View full article
The Conjur Enterprise product maintains a database table for holding audit log details. This table is used to store audit logs for the sole purpose of populating the display on Conjur Enterprise dashboard. This table does not display all Audit Log details. Idira encourages customers to pull audit logs directly from the Leader and store the logs for themselves to ensure full access to all audit logs. A problem can arise when the Idira Audit DB becomes too large. When this occurs, users will experience a degradation of product performance. Customers may note that requests to update or obtain secrets may take longer or time out altogether. At times, the node health check may also fail. If you feel that a large Audit DB may be impacting the performance of your Conjur Solution, we recommend the following: Check the size of the Audit DB Truncate the Audit DB Steps View the size of the Audit DB on the Leader:   1. On the Conjur master server, step into the running container: docker exec -it <conjur-container-name> bash 2. Connect to the Audit database: chpst -u conjur psql -d audit -p 5433 3. Find the number of audit messages: select count(*) from messages; 4. Find the size (on disk) of the audit database: SELECT pg_size_pretty( pg_database_size('audit')); 5. Exit: \q We recommend keeping the Audit DB with fewer than 100k messages to reduce performance impact. Truncate the Audit DB If the problem does indeed stem from the size of the Audit DB, we recommend the following steps to truncate the DB: *This method of cleaning the DB is quicker than a cleanup. *This will remove all the logs within the Audit DB. 1. Execute the TRUNCATE statement in the Conjur container to clear all audit messages from the database: user@host:~$ docker exec -u postgres <conjur container name> psql -p 5433 -d audit -c 'truncate messages'; 2. If successful, the output of the command should be: TRUNCATE TABLE root@container:/# su postgres Idira recommends scheduling a regular cleanup of the Audit DB table. The docker exec command shown above can be built into a bash script and executed via cronjob. As each Follower also maintains a unique Audit DB table, users should occasionally clean up Leaders/Stbys and all Followers. NOTE:  From Conjur SH version 13.2 onwards, the truncation of Audit Database  is handled automatically. For more details Please refer to  below linked documentation-  https://docs.cyberark.com/conjur-enterprise/13.2/en/content/operations/services/audit/dap-overview-audit-service.htm#Audit3
View full article
The specified steps describe: Configuring EPM Console for debugging. Troubleshooting using the EPM console and EPM agent. Helpful information required to open a Support ticket. Steps 1. Gather the Support Info logs from the affected endpoint and upload them to the case. a. Right-click the Agent tray icon > Get Support Info > select all the options. Click Next.   b. Select the option "Process monitor", then click Continue.   c. Click Accept after reading the Sysinternals Software License Terms.   d. Click Record and reproduce the problem. Click Done to stop recording and press Continue.   e. Describe the detailed problem and select Done. f. The data is collected, and the Complete screen opens, showing the path and name of the file containing the collected support information. Note: Look for the application within Support Info > Trace > vf_agent_srv.trace. This will provide more information about the policy that was triggered at the time. For example, the log event below is showing that the application notepad++.exe is triggering the Default Policy [Control unhandled applications - Detect] from the Default Policies section. 2025.12.09 16:09:54 PID:3384 TID:7000 CvfProcessStartRights::PROCSTART: pid=6204, parent=1172, exe=[C:\Program Files\Notepad++\notepad++.exe], cmdLine=["C:\Program Files\Notepad++\notepad++.exe" ], policyId=5430573, action=6 (COLLECT_ZERO_TOUCH), policyName=[[AppCtrl Detect: Old Applications (4.3)]], executeUser=[TESTWin11\testuser1] 2. In the EPM Agent configuration settings, please enable “Show Idira EPM Tab in File Properties” and provide a screenshot of the Idira EPM tab for the specific application. a. To enable this option, please go to Configuration -> Agent Configuration -> Select ... button beside General configuration -> Edit parameters -> Locate “Show Idira EPM Tab in File Properties” and select ... button -> Set to On (if it is off) -> Save. b. Proceed to right-click the Agent tray icon > Request Settings > Yes. c. On the application in question, right-click and go to properties. Select the tab called “Idira EPM” and upload this screenshot to the case. Example: The screenshot of the EPM tab shows what policy is currently being triggered if run manually. 3. If the above EPM tab shows a policy being triggered, please export the policy in question (Select ... button from the policy > Export) from the Policies section in the EPM console. If the above EPM tab displays "Run & Collect Events" instead, it indicates that the applied policy is a default policy. Note: If it is a different policy than what the client expects, it is necessary to examine both policies to determine why the application is triggering this policy instead of the expected policy (i.e., filtering on user/machine, filename parameters, time range settings, priority, etc.).
View full article
Overview This knowledge article provides a systematic procedure for investigating policy blocks in Idira Endpoint Privilege Manager (EPM) on macOS systems. It focuses on tracing the parent processes of a blocked item to determine the ultimate originating process. This approach is essential for understanding the root cause of blocks enforced by policies, such as those targeting applications or tasks. Perform these steps each time a block occurs to ensure accurate diagnosis. The process involves reviewing EPM logs, system logs, and policy configurations. Prerequisites: Access to EPM logs generated by Get Support Info (typically covering the last 8 hours or a relevant timeframe). Access to macOS system logs via tools like Console.app or log commands. Familiarity with basic log searching and process identification. Steps Step-by-Step Procedure Step 1: Review the EPM Logs Open the EPM logs file for the relevant period (e.g., the last 8 hours). Begin at the end of the file and search backward for entries related to the blocked process (e.g., the executable name or path). This establishes the initial context for the block event. Step 2: Locate the Blocked Event Entry Identify the log entry that documents the block. Look for indicators such as: The attempt to execute the process (e.g., "try to exec '/path/to/process'"). The denial action (e.g., "- deny"). The associated policy (e.g., policy name and ID). Record the policy details for later reference, as they explain the blocking criteria. Step 3: Extract the Parent Process Information Within the blocked event entry, locate the parent process details. This is typically formatted as: "parent: PID:additional_values" (e.g., the parent PID is the first value after "parent:"). Note the parent Process ID (PID) for further tracing. Step 4: Search for the Parent Process in EPM Logs Using the parent PID from Step 3, search the EPM logs for related entries. If no entries are found, the parent process may have initiated outside the current log window. In such cases, proceed to the system logs for a broader search using the same PID. Step 5: Trace Further in System Logs Access the macOS system logs (e.g., via Console.app or the log command). Search for the parent PID to find where it was created. Look for entries showing: The launching process (e.g., an executable that initiated the parent). Any associated scripts or commands. Repeat this tracing for subsequent parent PIDs until the ultimate originating process is identified (e.g., a script or service that started the chain). Step 6: Identify the Ultimate Originating Process Examine the log entries for details on the responsible process, such as: Script paths or originating executables. Contextual information indicating the root cause (e.g., a system service or user-initiated task). This step reveals the full process hierarchy leading to the block. Step 7: Analyze the Blocking Policy Review the policy configuration (e.g., via XML export or EPM console): Confirm the policy action (e.g., block) and targets (e.g., application groups). Verify how the blocked process matches the policy criteria (e.g., file name and location). This ensures the block aligns with intended security rules. Step 8: Document the Process Chain Summarize the findings by mapping the full chain of processes, from the ultimate parent to the blocked item. For example: Originating Process → Intermediate Process(es) → Blocked Process. Include relevant log excerpts, PIDs, and policy details for reference in incident reports or policy adjustments. Best Practices Conduct investigations promptly after a block to capture fresh log data. Use search tools or scripts for efficient log navigation if dealing with large files. If blocks recur, consider policy refinements to target root causes rather than downstream processes. This procedure enables precise troubleshooting, minimizing downtime and enhancing policy effectiveness in macOS environments.
View full article
Following EPM endpoint agent upgrade or reinstall, the My Computers page of the administrator console is showing significantly more machines than before, with many showing apparent duplicate entries.   The cause: EPM's agent has a uniquely generated Agent ID that is the primary identifier when communicating with EPM SaaS.  It can be deleted with an uninstallation with specific flags across different platforms.    There are 2 levels of deletions within the DB: The first is a soft-delete, which marks a file/agent/event for deletion. This removes it from the UI, reports, counters, and everything else customer facing. It is also relatively light on the DB The second is a hard-delete, this is a set of background jobs which run all the time taking batches of items which have been soft-deleted and performing the heavy work of actually removing them from the DB. It is significantly heavier in terms of effort and for this reason only occurs in small batches.  In some instances, a backlog of soft-deleted items can accumulate which have not yet been hard-deleted.  During the upgrade - agents that had only been 'soft-deleted' but not yet 'hard-deleted' were added back to the "alive list". Resulting in a higher number of agents post install than had existed before.   Resolution: When several agent duplicates exist, there could be license implications and policy targeting implications to consider, therefore ensuring you're running lean with your EPM agents is essential for success.  Using EPM's APIs in a script on a periodic basis could help with keeping this agent list cleaned out in addition to EPM's already existing cleanup mechanisms.   Attached to this article is a PowerShell script written to interact with EPM's Computers APIs to ultimately deduplicate existing agents.  It is purposely written to only delete duplicates and not existing agents by referencing the installation time. Steps Pre-Requisites: An existing EPM service user and EPM configured to allow for direct authentication For more information see our docs: https://docs.cyberark.com/epm/latest/en/content/admin/role-based-access-management.htm#add-user Download the script Identify your login dispatcher URL Identify your set ID   Steps: Set your login dispatcher - Edit the line 32 to reflect your login dispatcher that is one of the following: North America - https://login.epm.cyberark.com Europe - https://eu.epm.cyberark.com UK - https://uk.epm.cyberark.com Australia - https://au.epm.cyberark.com Canada - https://ca.epm.cyberark.com Singapore - https://sg.epm.cyberark.com India - https://in.epm.cyberark.com Japan - https://jp.epm.cyberark.com Italy - https://it.epm.cyberark.com Czech Republic - https://ch.epm.cyberark.com UAE - https://ae.epm.cyberark.com Israel - https://il.epm.cyberark.com Brazil - https://br.epm.cyberark.com FedRAMP - https://login.epm.cyberarkgov.cloud Example for FedRAMP: Identify your set ID that you'd like to run this on Pro tip: Triple-click to select the whole set ID From the EPM Console, click on Administration > Sets > select the set you like > Set Details Paste this into line 11 within the double quotes "Insert your set ID here" Insert your service user ID and password Note: Hard-coding credentials is NOT something that is a recommended practice and leaving them in perpetuity after use. Ideally, use a credential provider such as Idira Credential Providers, Idira Secrets Manager to replace these hard-coded credentials with API calls. However, if you don’t own these solutions, if you do paste the hard-coded credential, be sure to have a standard operating procedure of rotating that credential after use. Insert your EPM username within the double quotes "Insert your username" Insert your EPM password within the double quotes "Insert your password"        4. Once these values are in-place, run the script.  If there are no errors returned, then you should see an output of Getting computers: currently on page X indicating which page the script is on.  If no machines are to be deleted, then it will return with no results a.Example of return without results b. Example of return with results: Before: Script execution: After:
View full article
(6/17/2025) Microsoft has announced the release of Windows Administrator Protection feature for Windows 11. This feature is expected to be enabled by default and will impact EPM operation as outlined below. We are working to address all compatibility requirements in upcoming releases. Policy impact: “Remove Local Administrator” policy will remove the admin shadow users created by the new feature. “Protect administrative user groups” agent capability will not prevent the admin shadow users from activities in administrative user groups. Policies targeting administrative users: Targeting will not be applied to the admin shadow users that Microsoft’s new feature will create and may result in a lack of policy action to scoped applications. What do we do: At this time, the Microsoft Administrator Protection feature has not yet been officially released by Microsoft. Idira continues to closely monitor all Windows updates and track the status of this feature. Since the feature has not been published, it may still change before its final release. Idira is fully prepared to ensure compatibility once Microsoft officially makes this feature available. Our engineering teams have already reviewed the feature as published in earlier Microsoft releases, and we are ready to address any required adjustments. Should Microsoft introduce changes when the feature is formally released, Idira will provide any necessary updates or fixes accordingly. Because the feature is not yet released, no action is currently required from customers, and there are no compatibility concerns at this stage. For the same reason, Idira has not published additional documentation yet. Once Microsoft officially publishes the feature, Idira will review the final implementation and provide updated documentation and guidance. We will communicate all relevant information to our customers promptly.   Per Microsoft notice, this feature applies to Windows 11 Home, Windows 11 Professional, Windows 11 Enterprise and Windows 11 Education editions from certain builds to be announced by Microsoft.
View full article
Question How does the Idira EPM Agent validate digital signatures on Windows? Answer The Idira EPM Agent uses the Windows Signature Verification API (WinVerifyTrust) to validate digital signatures. This API enforces stricter requirements than standard Authenticode checks. For a signature to be considered valid by the EPM Agent:   The timestamp on the signature must fall within the certificate’s valid date range. All certificates in the signing chain must be valid and not expired. The root certificate in the chain must be installed on the machine. Because WinVerifyTrust performs full certificate chain validation, a signature may appear valid in some tools but still fail validation by the EPM Agent if any of these requirements are not met.
View full article
Question Does the EPM Agent Support SSL/TLS inspection or man-in-the-middle inspection? Answer No, the EPM agent does not support man-in-the-middle TLS inspection of web traffic In order to validate if SSL Inspection is in-place, there's a few options depending upon your platform (be sure to use a URL that is applicable to region as outlined here: Windows Use a simple script (example attached to this KB) which will prompt you for parameters for comparison. Be sure to supply the full URL (ex: https://agents-na.epm.cyberark.com) Example of a known good return of the server certificate's subject MacOS/Linux Use the curl command  curl -v https://agents-na.epm.cyberark.com Example of known good curl return: Example of a known SSL inspected traffic output. (Bad) The reason why this is bad is the subject of the certificate presented (in this case - the issuer of the certificate is an internal PKI issuer for a sample customer in government vs. Idira being the issuer) that is imitating a valid issuer is NOT the Idira EPM-related issuer mentioned in the previous screenshot of a known good result where Idira signed the certificate via GlobalSign.   SSL TLS Inspection considerations We have proxy support built into the Windows and Linux Agent for direct proxy connections. For Mac OS, we would leverage the OS configuration. https://docs.cyberark.com/Product-Doc/OnlineHelp/EPM/Latest/en/Content/Installation/Windows-InstallAgents.htm#Installationwithconfigurationdetails Many proxies/web application firewalls such as zScaler, Websense, Fortigate, etc have the ability to do man-in-the-middle TLS inspection of web traffic. The EPM agent does not support this. So we need to exempt the EPM agent from SSL Inspection. For more information, please see this: Example: https://help.zscaler.com/zia/certificate-pinning-and-ssl-inspection
View full article
Question Do we support in-place OS upgrades for our components such as, PVWA/CPM/PSMP/PSMP/PTA? Answer Unfortunately, no. We do not support in-place upgrades of OS for all of our components.  Permissions and links will be broken in the process.  You are recommended to follow the documentation to migrate to the newer OS instead. Such as:  Migrate PSM to a supported Windows OS | Idira Docs Migrate the Vault Operating System | Idira Docs Issue PVWA and PSM servers on Windows OS.  You want to know if you can perform an in-place upgrade of the Operating System to a more current supported version Environment Password Vault Web Access Upgrade Cause This is not supported. There is no support for in place OS upgrades for any Idira component servers. Resolution The recommendation is to install a fresh installation on the server and reinstall the software.
View full article
User-Based Apps are Installer based or portable programs that users can deploy to their profiles without administrative approval. These applications reside in directories where users have full control (e.g., %USERPROFILE% on Windows or /Users/{username}/ on macOS). While some of these apps are legitimate, they present three primary challenges: Lack of Oversight: Without non-privileged application controls, users and AI agents can run software with impunity, leading to the spread of "Shadow IT" and unauthorized software. Strategic Misalignment: Attempting full application control before privilege management is established often causes event overload and contradicts EPM Success Path best practices. Resource Constraints: Many organizations lack the operational capacity to manage comprehensive application control beyond their existing privilege management efforts.   These portable-apps are growing in popularity especially with the proliferation of Agentic AI.  AI solutions such as Claude, Cursor, and many others all can be installed within a user profile and be utilized for privilege escalation via relatively trivial mechanisms.  Steps The Solution This article introduces a scalable policy framework that allows you to manage user-based applications effectively at any point in your EPM journey. Because these applications  run without administrative rights, the framework utilizes a purpose-driven order of precedence consisting of Allow and Block policies. This approach integrates seamlessly with the QuickStart methodology, providing robust application control that evolves with your EPM Success Path journey.   This policy framework consists of three layers to manage user-based applications.    Layer 1 — Risk‑Reduction Controls The first layer restricts all user‑based applications from invoking sensitive OS components that are commonly abused by attackers or leveraged for automation by Agentic AI. This establishes a baseline reduction in privilege exposure. Layer 2 — Allowlist for Approved User‑Based Apps The second layer introduces an allowlist for sanctioned user‑based applications. The associated application group includes sample entries that you should refine by removing irrelevant examples and adding known, approved apps. This layer acts as a placeholder for ongoing discovery: as new legitimate user‑based apps are identified and approved, they should be added here following the same matching criteria in the examples. Layer 3 — Discovery and Default‑Deny The final layer contains two policies: One to discover and log user‑based applications. A second, higher‑priority policy that enforces a default‑deny posture once discovery is complete. When enabled, this default‑deny policy supersedes the discovery policy, making it optional to disable or remove the discovery rule. After default‑deny in the final layer is activated, any unapproved or unknown user‑based application will be blocked and logged. These logs can then be used to update the allowlist as needed. These policies are aligned with QuickStart, intersecting appropriately with risk reduction and privilege management to ensure compatibility. Customers using QuickStart can import them knowing they'll function as intended without disrupting QuickStart's design principles.     Pro-Tip When building definitions for Allowed User-Based Apps, use the following criteria: Location -  %USERPROFILE%, including any subfolders. This approach ensures that the policy  targets the user profile correctly, whether it's in the default directory (C:\Users), on a secondary drive, or part of a roaming profile. Publisher - If the app is signed, this option securely identifies the vendor. To approve all apps from a specific vendor, combine Publisher with Location and this is all you need. Original filename -   If you only want to approve a specific app from a vendor consider adding Original filename to the definition.  This is metadata compiled into an app’s file, making it a reliable way to identify a specific vendor app since it remains secure when the file is signed. Checksum - When a file isn't signed, the best way to verify its identity is by using the checksum, which is particularly important for applications stored in user profiles where metadata and filenames can be altered.
View full article
You can configure a proxy server to limit outbound access from EPM agents to the EPM service only. You can also use any other proxy server that follows the communication methods described below. The following example shows how a Squid proxy server is used to allow EPM agents installed on an isolated network to communicate with the EPM service.   Steps Configure a Squid proxy server This section describes how to configure the Squid proxy server and allow it access to the EPM server only.   Step 1: Check requirements As an example, the Squid proxy server could have the following minimum requirements: Item Requirement Hardware or VM 1 CPU core 1G RAM 10G disk Operating system RHEL 7 or higher Port TCP port 3128 open for LAN connections from EPM agents Internet Outbound internet on port 443 Step 2: Install Squid Install Squid by running the following commands. On yum-based Linux (RHEL/Centos/Fedora):   sudo yum install -y squid On apt-based Linux (Ubuntu/Debian):   sudo apt-get update sudo apt-get install -y squid Enable, start, and check the status of Squid.   sudo systemctl enable squid sudo systemctl start squid sudo systemctl status squid Step 3: Allow Squid to access the EPM service Configure Squid to allow access to the EPM service only.   Enable the host server and the dispatcher domains, and disable all other domains.   For the host server, specify the domain prefix for your region and the three numbers from the EPM console URL after login. For example, na123.epm.cyberark.com or jp123.epm.cyberark.com. For the dispatcher domain, specify the domain prefix for your region. For example, login.epm.cyberark.com or jp.epm.cyberark.com. Open the Squid configuration file using squid acl. On a fresh Squid installation, add the following four lines to the top of /etc/squid/squid.conf. Make sure you include the  http_access deny all  command.   # change na110.epm.cyberark.com to the address of your EPM Manager acl allowed_domain_epm_cyberark dstdomain na110.epm.cyberark.com acl allowed_domain_epm_cyberark_login dstdomain login.epm.cyberark.com http_access allow allowed_domain_epm_cyberark http_access allow allowed_domain_epm_cyberark_login # denies all others, this is by default http_access deny all After you have updated squid.conf, reload the Squid service.   systemctl reload squid   Test configuration Open the Squid access log.   tail -f /var/log/squid/access.log View the requests that were issued.   1600975689.552 119942 10.0.1.4 TCP_TUNNEL/200 4495 CONNECT login.epm.cyberark.com:443 - HIER_DIRECT/104.17.195.105 - 1600975799.556 192753 10.0.1.4 TCP_TUNNEL/200 278838 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - 1600976019.567 217535 10.0.1.4 TCP_TUNNEL/200 21746 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - 1600983924.018 8354226 10.0.1.4 TCP_TUNNEL/200 194881 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - 1600985039.945 61232 10.0.1.4 TCP_TUNNEL/200 5609 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - 1600985233.763 169371 10.0.1.4 TCP_TUNNEL/200 157880 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - 1600985369.959 136195 10.0.1.4 TCP_TUNNEL/200 18797 CONNECT na110.epm.cyberark.com:443 - HIER_DIRECT/104.17.194.105 - In squid.conf, disable or change epm_cyberark acls, and check that requests are blocked. The following steps lead you through an example check. Change  allowed_domain_epm_cyberark 123.epm.mycompany.com  to  456.epm.mycompany.com . Reload Squid.   systemctl reload squid Open /var/log/squid/access.log and check that a tcp_denied string was issued.   1600988752.530 0 10.0.1.4 TCP_DENIED/403 3816 CONNECT 123.epm.mycompany.com:443 - HIER_NONE/- text/html In the EPM Management console > My Computers, your computer now appears as 'disconnected'. Change  allowed_domain_epm_cyberark 456.epm.mycompany.com  back to  123.epm.mycompany.com . Reload Squid again.   systemctl reload squid Test with shell to check that Google is blocked, but the EPM agent is not. Try to access Google.   export HTTP_PROXY=10.0.1.5:3128 export HTTPS_PROXY=10.0.1.5:3128 curl http://google.com The following response should be returned.   HTTP/1.1 403 Forbidden Server: squid/3.5.27 Try to access the EPM agent.   curl -I https://123.epm.mycompany.com The following response should be returned.   HTTP/1.1 200 Connection established   This KB article is provided for educational purposes only and is not actively maintained. As product features may evolve over time, configuring a Squid server should always be done using the official product documentation. For support, please consult the Squid community or your specific software vendor.
View full article
This article details the steps needed for configuring EPM agent distribution through JAMF for MacOS installations Please be advised that the script has been tested in our lab and working properly but may need local customizations made by the customer's JAMF administrator for their environment. All customers should always refer to the online documentation for the latest updates and agent installation requirements and validate the needed installation method with their JAMF administrators. Note: you need to be logged in to download the attachments referred in this article   JAMF EPM PKG Deployment.zip Idira EPM with NE 15.mobileconfig (New - 11/22/2024) Configuration profile in Apple native mobileconfig format that contains all required system extensions including network filter.  This configuration profile is meant to be used with macOS 15 Sequoia and newer. Idira EPM without NE 15.mobileconfig (New - 11/22/2024) Configuration profile in Apple native mobileconfig format that contains all required system extension minus the network filter. This configuration profile is meant to be used with macOS 15 Sequoia and newer. Idira EPM with NE.mobileconfig Configuration profile in Apple native mobileconfig format that contains all required system extensions including network filter.  This configuration profile is meant to be used with macOS 14 (Sonoma) and older. Idira EPM NO NE.mobileconfig Configuration profile in Apple native mobileconfig format that contains all required system extension minus the network filter.  This configuration profile is meant to be used with macOS 14 (Sonoma) and older. jamfCyberArkEPMUninstall Uninstallation script used for removal of EPM agent using a shell script in Jamf JAMF EPM PKG Deployment Guide.pdf (Revised 06/25/2026) Deployment guide to cover all aspects of successfully deploying Idira EPM for Mac using Jamf via PKG method JAMF EPM DMG Deployment.zip Idira EPM with NE 15.mobileconfig (New - 11/22/2024) Configuration profile in Apple native mobileconfig format that contains all required system extensions including network filter.  This configuration profile is meant to be used with macOS 15 Sequoia and newer. Idira EPM without NE 15.mobileconfig (New - 11/22/2024) Configuration profile in Apple native mobileconfig format that contains all required system extension minus the network filter. This configuration profile is meant to be used with macOS 15 Sequoia and newer. Idira EPM with NE.mobileconfig Configuration profile in Apple native mobileconfig format that contains all required system extensions including network filter. This configuration profile is meant to be used with macOS 14 (Sonoma) and older. Idira EPM NO NE.mobileconfig Configuration profile in Apple native mobileconfig format that contains all required system extension minus the network filter. This configuration profile is meant to be used with macOS 14 (Sonoma) and older. JAMF profile with NE Guide.pdf Guide for deploying JAMF configuration profile with network extension (only to be used if uploading extensions fails) JAMF EPM DMG Deployment.pdf (Revised 11/22/2024) Deployment guide to cover all aspects of successfully deploying Idira EPM for Mac using Jamf via DMG method jamfCyberArkEPMInstallFromDMG Installation script used in DMG method using a shell script in Jamf jamfCyberArkEPMUninstall Uninstallation script used for removal of EPM agent using a shell script in Jamf Change history: 2021-01-19 - Added support to enable dynamic Finder extension for Big Sur, (change to install file) 2021-02-12 - Updated to handle failure to handle OS preference needs 2021-04-30 - Updated script test termination result 2021-09-08 - Updated to support version 21.8+ 2021-09-28 - Fixed script commented out cleanup lines 2022-01-15 - Updated xattr to match new OS requirements 2022-04-11 - Uploaded new documentation for v22.4+ MacOS EPM Agent packaging with PKG file 2023-03-21 - Uploaded new documentation for 23.3.0 that includes new Network Extension 2023-11-27 - Replace 'EPM JAMF Agent Deployment v22.4’ with ‘JAMF EPM PKG Deployment' 2023-11-27 - Added the following JAMF configuration profiles-‘Idira EPM NE.mobileconfig’ and ‘Idira NONE.mobileconfig’ 2024-02-07 - Reorganized attachments to separate DMG deployment method from PKG.  Please deploy only using one method.  Once you've identified the method, download the zip that pertains to that method. 2024-03-19 - Modified configuration profiles in both deployment methods to reflect new capabilities in 24.4 agent release, added description of each file. 2024-04-09 - Modified PKG Deployment Guide to include example syntax when creating PKG for clarity. 2024-11-22 - Modified PKG and DMG deployment guides to reflect changes needed for macOS 15 Sequoia, added configuration profiles for each deployment method specific to macOS 15 Sequoia.  2025-04-02 - Modified profiles from "NonRemovableSystemExtension" to "NonRemovableFromUISystemExtension" to resolve upgrades and removal issues from console and MDMs. 2026-06-25 - Updated PKG deployment guide to reflect our preferred deployment method (Using a Signed PKG and activation script) and removed legacy JAMF Profile with NE Guide.pdf since it is no longer relevant for PKG method.     NOTE: Please pay close attention to the updated guides on when to use these new configuration profiles.  Look for the section titled "Special Notes for macOS 15 (Sequoia)" and newer. Steps Please review the attached zip files, which contain documents and scripts for detailed steps and guided provisioning steps. Then, select the deployment method that applies to your organization (DMG or PKG) and download the respective zip file.
View full article
Question A collection of frequently asked questions regarding how the EPM Agent interacts with Active Directory (AD) and Domain Controllers (DCs), including cache refresh intervals, authentication methods, logging, and retry behaviors.   Answer 1. What is the default and configurable interval for the agent’s AD/DC cache refresh and any retry/rescan behavior? The EPM Agent maintains two distinct AD caches: AD Users/User Groups Cache: Built upon agent startup and refreshed with every policy update. AD Computer Groups Cache: Built upon agent startup and refreshed approximately every 24 hours. Both caches can be force-refreshed via the agent command-line interface (CLI).  Do note that "Request Settings" option from EPM tray icon does not refresh the AD Computer Groups Cache.   2. How does the agent authenticate to domain controllers (LDAP simple bind, Kerberos, LDAPS)? Only the AD computer groups cache mechanism authenticates directly to the DC. It initially attempts LDAPS authentication. If LDAPS fails, it will fall back to standard LDAP authentication.   3. Apart from the AD group membership of the computer account and user (during runtime), does the agent retrieve any other details from the DC? No additional details are retrieved. For computer accounts, the agent retrieves only group membership information. For users and user groups, it retrieves the mapping between the account name and the account Security Identifier (SID).   4. If the agent runs as Local System, does it authenticate to AD using the domain computer account, or does it require separate credentials for LDAP/AD queries? The agent authenticates using the native domain computer account. No separate LDAP credentials are required to be configured.   5. Where are the local agent logs stored, and do they show cache refresh activity, target DC, query type, timing, retries, and failures? Yes, cache refresh activity, retries, and failures are all logged. The logs are stored locally on the endpoint at the following path: C:\Program Files\CyberArk\Endpoint Privilege Manager\Agent\Trace To search specifically for AD computer group-related issues in these logs, filter your search using the prefix: AdCompGrpThread   6. Is user group membership collected at runtime for policy/app decisions? If yes, is it cached, and does each app launch trigger a fresh DC query? Yes, the agent maintains a cache for policy and application decisions. If an entry is not present in the cache, a DC query is executed during the initial application launch. That result is then stored in the cache until the next policy update, at which point the cache is completely rebuilt. Until that subsequent policy update occurs, user group membership remains safely cached, and no additional DC queries are generated by subsequent application launches.   7. How does the agent choose a DC, and is the process site-aware? The process is not site-aware. The EPM Agent leverages standard Windows APIs to naturally discover and select Active Directory Domain Controllers.   8. Does the agent perform AD/DC lookups immediately at startup, user logon, network reconnect, or when the agent service restarts? Could these events cause many endpoints to query DCs at the same time? Machine group membership information is queried from AD under the following conditions, utilizing randomization to prevent DC overload: Routine Refresh: The cache is refreshed approximately every 24 hours following an agent restart. The actual refresh interval varies randomly between 23 and 25 hours. Load Distribution: Because the cache refresh interval is triggered relative to the individual agent's restart time (and incorporates the 23-25 hour randomization) rather than an absolute clock time, it effectively prevents simultaneous queries and distributes the query load across the deployment.   9. What is the fallback/retry behavior if a DC is unavailable or slow? If the DC is unavailable, the agent retains the last successful query result in its cache. Retries are scheduled based on the cache type: Users/Policies Cache: Retries upon the next policy update. Computer Groups Cache: Retries after its standard cycle (approximately 24 hours). If an administrator needs to expedite the retry process, both caches can be manually force-refreshed via the agent CLI.   10. Does Idira provide expected LDAP/DC request rates per endpoint or sizing guidance for large deployments? No, there is currently no specific sizing guidance or expected request rate metrics provided for LDAP/DC queries.   11. How long does the EPM Agent typically take to synchronize OU membership after being moved to another Set? The operation to move an agent to a new Set completes almost immediately (within seconds to minutes, as it is a live server-pushed command). However, moving a Set does not force a refresh of the AD Computer Groups Cache (which backs OU/AD-group-based targeting). That cache continues to run on its independent 23–25 hour cycle. If a newly referenced AD group has never been cached, it is checked almost immediately. If the group was already cached and the endpoint's actual membership changed, the stale cache can persist for up to 23–25 hours until the next scheduled rescan.   12. Does the EPM Agent re-evaluate the local AD computer cache when policies are created or updated in the EPM Console? Yes, the agent refreshes the cache when a policy changes, specifically when new targets are added. However, it is important to emphasize that for AD computer groups, the agent does not automatically refresh the cache when a group is modified directly in Active Directory. The agent has no indication if or when a group changes within AD. For those AD modifications, the agent will only recognize the new group membership during its next scheduled cache refresh cycle (or if a refresh is manually forced).
View full article
How do I Add Endpoint Privilege Manager Agents to Standard (Golden) OS image (The information provided applies only to Agents version 26.4 and earlier). NOTE - For non-persistent VDI additional info see the following Article: https://community.cyberark.com/s/article/00003942 Steps EPM agent generates a unique agent ID which is required to communicate with the EPM Server. Follow the procedure below to include the EPM agent in your standard OS image. If you want to create an image of a system with the EPM Agent installed and then to deploy it across the network in your business environment, you should create the image on the host where the Agent-ID has been removed from the registry. This will help you to avoid issues with identification of computers by the EPM Server.   ----------------- Please note: VMWare Instant Clones methodology re-clones 'live' agents/endpoints, instead of the original master. This causes the duplication of these agents ID, preventing them from being visible in the console. This is currently a limitation with deploying EPM agents using golden image in this environment. ----------------- Instructions for Creating the Agent’s Golden Image Install the agent on the PC with the Operating System image. Choose a preparation method: Automatic (Beta): Use the attached script and run as Administrator. Manual: Follow steps 3–7 below. (Manual) Shutdown the agent self-defense if enabled. You can verify the agent's self-defense protection status using the following Windows command: sc query vfpd Instructions: https://docs.cyberark.com/epm/latest/en/content/policies/protecting%20agents.htm#Generateasecuretoken  Command line (run in agent folder):   vf_agent.exe -UseToken XXXXX vf_agent.exe -StopPasServ vf_agent.exe -StopServ   Remove the Trace folder: C:\Program Files\CyberArk\Endpoint Privilege Manager\Agent\Trace\ Remove the policies file: C:\Program Files\CyberArk\Endpoint Privilege Manager\Agent\Config\vf_policies.xml Remove Agent ID from registry: Open Registry Editor → HKEY_LOCAL_MACHINE\SOFTWARE\Viewfinity\Agent Delete the Agent-ID key. This ensures a new Agent ID is created when the agent starts as NP-VDI. Generate the image of the PC.   ----------------- Following recent issues with Azure Virtual Desktop (VDI), please note that golden image machines may be reset multiple times. To prevent the EPM Agent from starting prematurely during the image preparation process, you must remove the agent from the auto-start list before capturing the image. Registry Path: HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run Remove the entry for: vf_agent.exe After the image has been deployed and a new user account has been created and logged in, you will then need to re-enable or add the process to auto-start and generate the agent ID to connect to our servers.   *  The information provided applies only to Agents version 26.4 and earlier.  NOTE - For non-persistent VDI additional info is at Solution 3942.   Q: How do I Add Endpoint Privilege Manager Agents to Standard (Golden) OS image   A: EPM agent generates a unique agent ID which is required to communicate with the EPM Server. Follow the procedure below to include the EPM agent in your standard OS image. If you want to create an image of a system with the EPM Agent installed and then to deploy it across the network in your business environment, you should create the image on the host where the Agent-ID has been removed from the registry. This will help you to avoid issues with identification of computers by the EPM Server.   Install the agent on the PC with the Operating System image by taken the setup file from the My Computers > Download Agent Installation page of the Manager UI. After the image is deployed, the new Agent ID will be generated automatically.   (Note - If you have trouble removing Agent-ID from the registry in Version 6.0, verify that the Agent self-defense password is disabled (Agent Configuration > Agent self-defense password > *Remove the password* > Save))         Shutdown the agent self defense if self defense is turned on (if agent self defense is turned off then skip this step). By EPM console, at the computer configuration page(delete the password there for the specific image machine, then update your policies) By command line Open CMD, run the command (under the agent folder) vf_agent.exe -UseToken XXXXX Token can be generated in the Computers/Uninstall window in EPM console At the same CMD window run  vf_agent.exe -StopPasServ At the same CMD window run  vf_agent.exe -StopServ Remove content of Trace folder (there should only be one of the following): Legacy Viewfinity File Location - C:\Program Files\Viewfinity\Agent\Trace Legacy Idira File Location for 5.x - C:\Program Files\CyberArk\Viewfinity\Agent\Trace\ Idira File Location for 6.0+ - C:\Program Files\CyberArk\Endpoint Privilege Manager\Agent\Trace\ or C:\Program Files\CyberArk\Endpoint Privilege Manager\Agent\Config\vf_policies.xml Open registry editor and navigate to: HKEY_LOCAL_MACHINE\SOFTWARE\Viewfinity\Agent Remove Agent-ID key from registry before you create an image, so that the image does not contain Agent ID, a new one will be created once the agent comes up as a new NP-VDI Generate an image of PC
View full article
Pre-Requisites: 1. Access to Microsoft's EDR console (security.microsoft.com) NOTE: Will likely involve opening a case to enable the functionality 2. Access to Idira EPM Console Steps Steps needed: Open a case with Microsoft via the Security Portal (security.microsoft.com) Supply the following info: Probing Questions: What specific EPM Agent version is installed? What specific directories or paths do you want to exclude? %ProgramFiles%\CyberArk\Endpoint Privilege Manager\Agent\ Any specific performance metrics? Supply a procmon capture that shows MSSense "touching" files within the EPM directory supplied above Have you already tested any temporary exclusions, and if so, what were the results? Since its not possible to exclude EDR without it being enabled…no.  AV Exclusions are not in-scope for this Is MSSense.exe running on all affected endpoints, and what is its current configuration? Microsoft Info Org ID: Settings -> System Settings -> Microsoft Defender XDR -> General -> Account Tenant ID: Settings -> System Settings -> Microsoft Defender XDR -> General -> Account License Details Settings -> Endpoints -> Licenses Device ID for 1 specific device Not sure why they need this since they're enabling the feature flag vs. adding the exclusion for you Assets -> Devices -> {Insert hostname here} -> Device ID Statement to request feature for your tenant "Please add the EDR exclusion feature in my system as we need to assess performance differences with an exclusion for Idira EPM" via email with support rep Once EDR Exclusions is added to your tenant add the EDR Exclusion policy A decision point will be to determine if you want to use a groupID or deploy to the entire organization Entire organization is exactly as it sounds - any endpoint that is enrolled in Intune and Defender for Endpoint will show up GroupID allows you to be a little more granular The underlying mechanism here is a registry entry that gets added to the machine and can be deployed via Intune As of the time of this writing, Intune baselines nor policies contain EDR exclusions For manual testing perform the following Create groupID in MS Defender Security Console The end of the wizard will supply a PowerShell script to create the registry entry New-ItemProperty -Path 'HKLM:\SOFTWARE\Policies\Microsoft\Windows Advanced Threat Protection' -Name 'GroupIds' -PropertyType String -Value 'Exclusions={Insert groupID here} -Force Take that registry entry and as an admin, add to the machine After adding it, restart the Sense service Sc stop Sense Sc start Sense Reboot the machine After reboot, check in the security console to verify that the EDR Exclusions show a status of green, then check to see if your endpoint appears in the covered list Troubleshooting If the registry add is not functioning properly do the following; Run these 3 powershell commands to remove the existing groupID, re-add it, and verify its there Remove-ItemProperty -Path 'HKLM:\SOFTWARE\Policies\Microsoft\Windows Advanced Threat Protection' -Name 'GroupIds' -ErrorAction SilentlyContinue New-ItemProperty -Path 'HKLM:\SOFTWARE\Policies\Microsoft\Windows Advanced Threat Protection' -Name 'GroupIds' -PropertyType String -Value 'Exclusions={Insert groupID here} -Force Get-ItemProperty -Path 'HKLM:\SOFTWARE\Policies\Microsoft\Windows Advanced Threat Protection' -Name 'GroupIds' After adding it, restart the Sense service Sc stop Sense Sc start Sense Reboot the machine After reboot, check in the security console to verify that the EDR Exclusions show a status of green, then check to see if your endpoint appears in the covered list
View full article
Question If your current state is... And you are moving to... Then follow this path: EPM 25.10+ (Installed on macs) macOS 26.3 You can upgrade macOS directly. The agent will remain fully compatible. Old EPM (< 25.10) + macOS 26.3 EPM Upgrade Requirement: You MUST install EPM 26.3 or higher. Any version lower than 26.3 will fail to function. Clean macOS 26.3 New Installation Requirement: Install EPM 26.3 or higher. Lower versions are not supported on this OS version. EPM 26.2 + macOS 26.3 (Disconnected) Connectivity and bug Fix Use the CyberArkEPM.pkg file with the activation script via your MDM (e.g., Jamf, Intune) to restore communication. Answer Important Troubleshooting & Performance Notes Status Check: You can verify the current state of the agent at any time via the Terminal using the following command: CyberArkEPM -status Recommendation for 26.2 Users: Even if your EPM 26.2 agent appears to be working on macOS 26.3, we strongly recommend upgrading to 26.3. This version contains critical bug fixes for policy enforcement and stability issues introduced by the macOS update. [!CAUTION] Installing an EPM version lower than 26.3 on a macOS 26.3 machine will result in an "Failed to communicate to console \ Disconnected \ not working" state where the agent will not communicate to the console and will not pull policies.
View full article
We will be updating this article based on the questions we receive about the security bulletin. Please click the Follow button to receive change notifications. Where can I see the changes done in newer versions? All release notes can be found Here Is it necessary to uninstall the agent in order to install the updated version? No, the agent can be upgraded per the instructions in the following doc: Windows Agents MacOS Agents Linux Agents I'm receiving an "invalid key" or "incorrect configuration file" when attempting to install the new version When prompted for the installation key, make sure to use the one that was generated when downloading the installer package from the EPM console. When asked to select the configuration file, select the .config file that is included in the installation package. For more information, please refer to documentation
View full article
Question What are the exact evaluation criteria and trust requirements for certificate-based Publisher matching in EPM Application Groups? Specifically, this article addresses the following scenarios: How does exact publisher matching function within EPM Application Groups? Does EPM validate the entire certificate trust chain, or just the Publisher string? Can self-signed certificates or spoofed Publisher names trigger a policy match? Does EPM require signatures from globally recognized Certificate Authorities (CAs), and how are custom CAs handled? Answer EPM evaluates digital signatures using the strict Windows WinVerifyTrust API to ensure malicious files cannot spoof legitimate application publishers. Here is how EPM processes these rules and validates certificates: 1. Exact Publisher Matching When a policy targets a specific publisher (e.g., "Microsoft Corporation"), the EPM Agent looks for an exact string match against the Subject / Publisher Name field of the file's digital certificate. However, EPM will only evaluate this string match if the underlying certificate successfully passes a strict validity check first. 2. Certificate Chain Validation EPM does not blindly match the publisher name. Before EPM evaluates the name against your policy, it validates the entire certificate chain: The certificate must not be expired at the time of the signature. The certificate must not be revoked. The chain must successfully terminate at a Root CA trusted by the local endpoint. If the certificate fails this chain validation, EPM treats the signature as invalid, and publisher-based policies will not apply. 3. Self-Signed Certificates & Publisher Spoofing A malicious actor cannot simply create a self-signed certificate, name the publisher "Microsoft", and bypass EPM policies. Because EPM strictly validates the certificate chain, an unverified self-signed certificate will fail the WinVerifyTrust check. If the certificate chain is invalid, EPM ignores the spoofed publisher name entirely. 4. Certificate Authority (CA) Requirements EPM requires the certificate to be signed by a recognized CA, but "recognized" is strictly defined by the endpoint's Windows Operating System. Trust Delegation: EPM does not maintain its own proprietary list of trusted CAs, nor is there an option in the EPM Console to upload a custom CA. EPM delegates trust decisions entirely to the Windows Certificate Store. ⚠️ Important: EPM references the Local Computer certificate store only. The Current User certificate store is not consulted during signature validation, regardless of what is installed there. If a certificate chain (root or intermediate CA) exists only in a user's personal store, EPM will still treat the signature as untrusted. Custom / Internal CAs: If your organization uses an internal CA or a specific self-signed certificate that you want EPM to accept, you must deploy that CA's public certificate to the Trusted Root Certification Authorities store on your Windows endpoints (e.g., via Active Directory GPO or Intune). Once Windows natively trusts the CA, EPM will recognize it as valid and allow publisher-based policies to apply.
View full article
Authentication Context in Microsoft Entra lets us make our app ask for stronger sign-in (like MFA or FIDO2) — but only when needed. Instead of always asking users for multi-factor authentication (MFA), we can ask for it only during sensitive actions. With this feature: Define labels like  c1 ,  c2  (called authentication contexts) Connect those to Conditional Access policies (which define what kind of sign-in is required) App passes  acr_values=c1  when it needs stronger authentication. As per OIDC spec -  In Autherixation Request- "acr_values OPTIONAL. Requested Authentication Context Class Reference values. Space-separated string that specifies the acr values that the Authorization Server is being requested to use for processing this Authentication Request, with the values appearing in order of preference. The Authentication Context Class satisfied by the authentication performed is returned as the acr Claim Value, as specified in Section 2. The acr Claim is requested as a Voluntary Claim by this parameter." In  ID Token -    "acr OPTIONAL. Authentication Context Class Reference. String specifying an Authentication Context Class Reference value that identifies the Authentication Context Class that the authentication performed satisfied. The value "0" indicates the End-User authentication did not meet the requirements of ISO/IEC 29115 [ISO29115] level 1. For historic reasons, the value "0" is used to indicate that there is no confidence that the same person is actually there. Authentications with level 0 SHOULD NOT be used to authorize access to any resource of any monetary value. (This corresponds to the OpenID 2.0 PAPE [OpenID.PAPE] nist_auth_level 0.) An absolute URI or an RFC 6711 [RFC6711] registered name SHOULD be used as the acr value; registered names MUST NOT be used with a different meaning than that which is registered. Parties using this claim will need to agree upon the meanings of the values used, which may be context specific. The acr value is a case-sensitive string." Following are the steps required to set up the entire flow: Steps Step 1: Create an Authentication Context Authentication Contexts act as flags in Entra to trigger Conditional Access policies. Go to: Microsoft Entra Admin Center → Protection → Conditional Access → Authentication Contexts Click + New authentication context Fill in: Name: e.g., requireMFA_Auth_Context Description: e.g., Requires MFA Click Save Copy the ACR short name, e.g.,  c1 This short name will be used in the  acr_values  param in OIDC request   Step 2: Create an Authentication Strength If the built-in strengths don't meet your needs, create a custom authentication strength. Go to: Microsoft Entra Admin Center → Protection → Authentication Methods → Authentication Strengths Click + Add authentication strength Name it:  e.g., RohitMFATest2 Select any required allowed methods: FIDO2 security key Microsoft Authenticator Sms Click Create     Step 3: Create a Conditional Access Policy Go to: Microsoft Entra Admin Center → Protection → Conditional Access Click + New policy Name:  e.g.,TestingStepUp Under Assignments: Users: Select users or groups Target resources: Choose Authentication Contexts Select the one created earlier (requireMFA_Auth_Context) Under Access controls → Grant: Select Require authentication strength Select the built-in or custom strength (e.g., RohitMFATest2) Enable the policy → Report-only or On Save the policy.       This KB article is provided for educational purposes only and is not actively maintained. As Azure features may change, we highly recommend referring to the official Microsoft documentation for the most up-to-date guidance. For further support, please reach out to Microsoft Support.
View full article
On Wednesday, May 13, 2026 we disclosed several vulnerabilities affecting the EPM products. As the product line is in transition to the Palo Alto Networks’ Network Security platform, the disclosure has been performed under that platform’s standard. Please see https://www.cyberark.com/ca26-19 for the disclosure details. We will update this article based on the questions we receive about the security bulletin. Please click the Follow button to receive change notifications Where can I see the changes done in newer versions? All release notes can be found Here Is it necessary to uninstall the agent in order to install the updated version? No, the agent can be upgraded per the instructions in the following doc: Windows Agents MacOS Agents Linux Agents I'm receiving an "invalid key" or "incorrect configuration file" when attempting to install the new version When prompted for the installation key, make sure to use the one that was generated when downloading the installer package from the EPM console. When asked to select the configuration file, select the .config file that is included in the installation package. For more information, please refer to documentation  
View full article
Question EPM If I use Entra for SAML SSO Do I need to replace an expired EPM Certificate in Entra Answer When configuring SAML/SSO in Entra (Azure), using a certificate from the service provider (EPM) is optional. So the procedure for you will vary depending on how you set up SAML in Entra.   In Entra, find the "Single sign-on" section. You should be able to identify it because the Sign on URL appears as https://login.epm.cyberark.com/SAML/<yourorgname> In section "3" SAML Signing Certificate check if you see "verification certificate" Required is "yes" or "no".   ** If "verification certificate" Required is "no", then uploading the certificate from the EPM console is optional. You need to go to the EPM console's Administration section. In SAML integration, click "new EPM certificate is updated on the IDP server", click button "Switch to new EPM certificate" ** If "verification certificate" Required is "yes", download the certificate from the EPM console's SAML integration. Uploaded it to the Entra. Once you upload it, it may show as inactive. Set it to active. Back in the EPM console, in SAML integration, click "new EPM certificate is updated on the IDP server", click button "Switch to new EPM certificate"
View full article
When functionality changes between EPM agent releases, you may need to apply different policy logic based on the agent version. This procedure ensures that version-specific behaviors are isolated and prioritized correctly. This document defines an approach whereby release-specific application policies can be defined. In this example, the policies are shown to be specific to release 26.2.1. However, the actual “Agent Version” used can be changed depending on the specific upgrade being performed   Steps Step 1: Create a Dynamic Endpoint Group Navigate to Endpoints > Endpoint Groups and click Create Endpoint Group. Name: Agents - Version 26.2.1 (Use a descriptive name). Type: Select Dynamic. Rule Definition: Property: Agent Version Operator: Starts With (Recommended over "Equals" to account for minor build variations). Value: 26.2.1 Step 2: Duplicate and Adjust Impacted Policies Identify the policy requiring a version-specific change. Select Duplicate from the policy actions. Rename: Append the version to the name (e.g., Elevate Developer Tools - v26.2.1). Priority: Increase the priority so this policy is evaluated before the standard existing policy. Step 3: Target the Policy In the new policy, scroll to the Targets section. Add the Dynamic Group created in Step 1. Important Logic Note: Targets across different categories (e.g., AD Computer Groups AND Endpoint Groups) are combined using AND logic. Targets within the same category are combined using OR logic. Result: The policy will only apply if the computer is in the specified AD Group AND running the specified Agent Version Note: If the "old" policy has a very broad target (like "All Computers"), it is recommended to add the new Version 26.2.1 Dynamic Group as an Exclusion in that old policy. This ensures a clear division and prevents an agent from potentially matching both sets of criteria. Step 4: Validate and Roll Out Test: Upgrade a small "Pilot" group of agents to the new version. Verify: Use the Policy Audit on the endpoint to confirm the new version-specific policy is the one being applied. Cleanup: Once all agents are eventually upgraded to the new version, you can remove the version-specific policy and remain with the original policy definition.
View full article
Question Why macOS Shows Two EPM Prompts? Answer This behavior is by design. In the latest macOS versions, Apple has changed the Finder’s handling of access attempts. Finder now automatically retries the request after the initial attempt, even when the retry occurs outside the EPM grace period. As a result, users may see the EPM authentication prompt twice. This is a macOS behavior change, not an issue with EPM. EPM is functioning as expected by enforcing the configured policy and grace period.
View full article
Question After setting up the EPM Block policy, there is a Windows dialog that appears after the application is blocked as expected. This behaviour is not encountered in previous Windows OS versions, and it does not appear for all blocked applications. Is this an expected behaviour? Answer The EPM Agent does not have control over the appearance of the Windows dialog encountered after the application is blocked. Various applications may trigger different system dialogs, which means that this behaviour is dependent on the OS itself. As the EPM Agent does not suppress this Windows error dialogue, its appearance is not related to EPM.
View full article
Question When signing in to Identity with the EPM Mac or Windows Agent, the "Remember Me" option does not remember the username in subsequent logins. Answer This is because the EPM Agent does not support browser cookies. This fails because the Identity login relies on a "Remember Me" cookie to fetch the username. No resolution is currently available, as the EPM agent login does not support browser cookies. There has since been an Enhancement Request received for this feature :https://community.cyberark.com/s/article/Enable-EPM-to-support-browser-cookies-for-Remember-054d-890
View full article
How do I generate an application crash dump on the Vault server through Windows Error Reporting? Steps There are two parts to enabling Memory DUMP in the Vault server  The first one is to enable DUMP from the Windows server and the second one is to configure the vault to create the DUMP in case the application crash 1. Enable the Vault Application to create the memory DUMP:  Configure the Vault to enable the crash dump release by setting DisableExceptionHandling=Yes in DBParm.ini: A) Open the vault configuration file 'dbparm.ini' (default location 'C:\Program Files\PrivateArk\Server\') and set the parameter 'DisableExceptionHandling=Yes', B) Save the file and restart the vault service for changes to apply. Pay attention that we only set this parameter when we can reproduce the issue. In general setting it to "Yes" will cause other issues that we catch to crash. 2. Enable and Configure Application Crash Dump on the Operating System:   A) Make sure that Windows Error Reporting (WER) is enabled on the Vault server.   B) Update the registry to set up crash file location and collection type. Always back up the registry before making changes to it.   Go to the registry editor HKLM:\SOFTWARE\Microsoft\Windows\Windows Error Reporting\LocalDumps Add a key (registry "folder") with the executable name, e.g. dbmain.exe Create a string key DumpFolder, set the value to some folder, e.g. c:\dumps Create a numeric key DumpType, and set the value to 2 for getting "Full dump".  (Reference: https://docs.microsoft.com/en-us/windows/win32/wer/wer-settings)     NOTE: Sometimes, the LocalDumps key does not exist. If it does not, you can create it by right-clicking on windows error reporting and selecting 'create new key'. The values you create should be as follows:        1. String value DumpFolder with value C:\<dumplocation>. It is recommended you do not use root.        2. DWORD32 value DumpCount - 10 (decimal)        3. DWORD32 value DumpType - 2 (decimal) (2 is a full dump, 1 is minidump)        4. DWORD32 value CustomDumpFlags - 0 (decimal) - Wait till the crash reoccurs - Retrieve the application crash file from the location specified in the registry - remove DisableExceptionHandling=Yes in DBParm.ini after that and restart the vault server service in order to save changes  
View full article
  • 234 Posts
  • 17 Subscriptions