Agentic Endpoint Security (AES)
AI agents, AI coding tools, MCP servers, IDE extensions, browser plugins, and code packages such as npm and pip create a non-binary endpoint attack surface that traditional antivirus tools do not cover. By enabling this capability, the Cortex XDR agent discovers agentic software running on the endpoint and remediates risks based on your AES policy. Learn more about Agentic Endpoint Security with Koi.
Browser exploits protection
Browsers can be subject to exploitation attempts from malicious web pages and exploit kits that are embedded in compromised websites. By enabling this capability, the Cortex XDR agent automatically protects browsers from common exploitation attempts.
|  |  | — | — | — | |Logical exploits protection
Attackers can use existing mechanisms in the operating system—such as DLL-loading processes or built in system processes—to execute malicious code. By enabling this capability, the Cortex XDR agent automatically protects endpoints from attacks that try to leverage common operating system mechanisms for malicious purposes.
|  |  | — | — | — | |Known vulnerable processes protection
Common applications in the operating system, such as PDF readers, Office applications, and even processes that are a part of the operating system itself can contain bugs and vulnerabilities that an attacker can exploit. By enabling this capability, the Cortex XDR agent protects these processes from attacks which try to exploit known process vulnerabilities.
|  |  |  | — | — | |Exploit protection for additional processes
To extend protection to third-party processes that are not protected by the default policy from exploitation attempts, you can add additional processes to this capability.
|  |  |  | — | — | |Operating system exploit protection
Attackers commonly leverage the operating system itself to accomplish a malicious action. By enabling this capability, the Cortex XDR agent protects operating system mechanisms such as privilege escalation and prevents them from being used for malicious purposes.
|  |  |  | — | — | |Unpatched vulnerabilities protection
If you have Windows endpoints in your network that are unpatched and exposed to a known vulnerability, Palo Alto Networks strongly recommends that you upgrade to the latest Windows Update that has a fix for that vulnerability. If you choose not to patch the endpoint, the Unpatched Vulnerabilities Protection capability allows the Cortex XDR agent to apply a workaround to protect the endpoints from the known vulnerability.
|  | — | — | — | — | | Malware security profiles | | | | | | |Behavioral threat protection
Prevents sophisticated attacks that leverage built-in OS executables and common administration utilities by continuously monitoring endpoint activity for malicious causality chains.
|  |  |  | — | — | |Credential gathering protection
Targets attempts to access and harvest passwords and credentials.
|  |  |  | — | — | |Anti webshell protection
Prevents web shell attacks by continuously monitoring endpoints for processes that try to drop malicious files.
|  |  |  | — | — | |Financial malware threat protection
Targets attempts to access or steal financial or banking information.
|  |  |  | — | — | |Cryptominers protection
Prevents cryptomining by monitoring for processes which attempt to locate or steal cryptocurrencies.
|  |  |  | — | — | |In-process shellcode protection
Targets attempts to run in-process shellcodes that load malicious code.
|  | — | — | — | — | |Ransomware protection
Targets encryption based activity associated with ransomware to analyze and halt ransomware before any data loss occurs.
|  |  | — | — | — | |Prevent malicious child process execution
Prevents script-based attacks used to deliver malware by blocking known targeted processes from launching child processes commonly used to bypass traditional security approaches.
|  |  |  | — | — | |Portable executables and DLLs examination
Analyzes and prevents malicious executable and DLL files from running.
|  |  |  | — | — | |ELF files examination
Analyzes and prevents malicious ELF files from being executed or written to disk.
| — | — |  | — | — | |Local file threat examination
Analyzes and quarantines malicious PHP files arriving from the web server.
| — | — |  | — | — | |Office files examination
Analyzes and prevents malicious macros embedded in Microsoft Office files from running.
|  | — | — | — | — | |JScript files examination
Analyzes and prevent malicious JScript files from being executed or written to disk.
|  | — | — | — | — | |Mach-O files examination
Analyzes and prevents malicious mach-o files from loading and running.
| — |  |  | — | — | |DMG files examination
Analyzes and prevents malicious DMG files from running.
| — |  | — | — | — | |APK files examination
Analyzes and prevents malicious APK files from running.
| — | — | — |  | — | |Reverse shell protection
Detects suspicious or abnormal network activity from shell processes and terminate the malicious shell process.
| — | — |  | — | — | |Network packet inspection engine
Analyzes network packet data to detect malicious behavior.
|  |  | — | — | — | |Dynamic kernel protection
Protect the endpoint from kernel-level threats such as bootkits, rootkits, and susceptible drivers.
|  | — | — | — | — | |SMS and MMS malicious URL filtering
Filter, report, or block (iOS only) malicious URLs received in SMS/MMS messages.
| — | — | — | — |  | | **Spam reports** | — | — | — | — |  | | **Call and messages blocking** | — | — | — | — |  | | **Container-escaping attempts** | — | — |  | — | — | |Network URL filtering
URL filtering for supervised devices
| — | — | — | — |  | |Cryptocurrency wallets protection
Protection for cryptocurrency wallets stored on endpoints.
|  |  | — | — | — | |LDAP query protection
Analyze and act upon suspicious LDAP queries sent by the agent to a Domain Controller, to detect and block Active Directory reconnaissance attacks.
|  | — | — | — | — | |Malicious device protection
Protect your systems from unauthorized hardware attacks and malicious USB devices. The Malicious Device Prevention module identifies and blocks Human Interface Device (HID) tools, such as the "USB Rubber Ducky", that exploit device trust to inject unauthorized keystrokes and similar actions. This feature reduces the physical attack surface, and prevents hardware-based social engineering threats from compromising data.
| — |  | — | — | — | | Restrictions security profiles | | | | | | |Execution paths
Many attack scenarios are based on writing malicious executable files to certain folders such as the local temp or download folder and then running them. Use this capability to restrict the locations from which executable files can run.
|  | — | — | — | — | |Network locations
To prevent attack scenarios that are based on writing malicious files to remote folders, you can restrict access to all network locations except for those that you explicitly trust.
|  | — | — | — | — | |Removable media
To prevent malicious code from gaining access to endpoints using external media such as a removable drive, you can restrict the executable files, that users can launch from external drives attached to the endpoints in your network.
|  | — | — | — | — | |Optical drive
To prevent malicious code from gaining access to endpoints using optical disc drives (CD, DVD, and Blu-ray), you can restrict the executable files, that users can launch from optical disc drives connected to the endpoints in your network.
|  | — | — | — | — | | | | | | | | | | | | | | | **Endpoint protection modules** Each security profile applies multiple security modules to protect your endpoints from a wide range of attack techniques. While the settings for each security module are not configurable, the Cortex XDR agent activates a specific protection module depending on the type of attack, the configuration of your security policy, and the operating system of the endpoint. When a security event occurs, the Cortex XDR agent logs details about the event including the security module employed by the Cortex XDR agent to detect and prevent the attack based on the technique. To help you understand the nature of the attack, the alert identifies the protection module the Cortex XDR agent employed. The following table lists the modules and the platforms on which they are supported. A dash (—) indicates that the module is not supported. | Module | Windows | Mac | Linux | Android | | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | |Anti-Ransomware
Targets encryption-based activity associated with ransomware and have the ability to analyze and halt ransomware activity before any data loss occurs.
|  |  | — | — | |APC protection
Prevents attacks that change the execution order of a process by redirecting an asynchronous procedure call (APC) to point to the malicious shellcode.
|  | — | — | — | |Behavioral threat
Prevents sophisticated attacks that leverage built-in OS executables and common administration utilities by continuously monitoring endpoint activity for malicious causality chains.
|  |  |  | — | |Brute force protection
Prevents attackers from hijacking the process control flow by monitoring memory layout enumeration attempts.
| — | — |  | — | |Child process protection
Prevents script-based attacks that are used to deliver malware, such as ransomware, by blocking known targeted processes from launching child processes that are commonly used to bypass traditional security approaches.
|  |  |  | — | |Container escaping protection
Prevents container-escaping attempts
| — | — |  | — | |CPL protection
Protects against vulnerabilities related to the display routine for Windows Control Panel Library (CPL) shortcut images, which can be used as a malware infection vector.
|  | — | — | — | |Data Execution Prevention (DEP)
Prevents areas of memory defined to contain only data from running executable code.
|  | — | — | — | |DLL hijacking
Prevents DLL-hijacking attacks where the attacker attempts to load dynamic-link libraries on Windows operating systems from unsecured locations to gain control of a process.
|  | — | — | — | |DLL security
Prevents access to crucial DLL metadata from untrusted code locations.
|  | — | — | — | |Dylib hijacking
Prevents Dylib-hijacking attacks where the attacker attempts to load dynamic libraries on Mac operating systems from unsecured locations to gain control of a process.
| — |  | — | — | |Exploit kit fingerprint
Protects against the fingerprinting technique used by browser exploit kits to identify information: such as the OS or applications which run on an endpoint—that attackers can leverage when launching an attack to evade protection capabilities.
|  | — | — | — | |Font protection
Prevents improper font handling, a common target of exploits.
|  | — | — | — | |Gatekeeper enhancement
Enhances the macOS gatekeeper functionality that allows apps to run based on their digital signature. This module provides an additional layer of protection by extending gatekeeper functionality to bundles and child processes so you can enforce the signature level of your choice.
| — |  | — | — | |Hash exception
Halts execution of files that an administrator identified as malware regardless of the WildFire verdict.
|  |  |  |  | |Hot patch protection
Prevents the use of system functions to bypass DEP and address space layout randomization (ASLR).
|  | — | — | — | |Java deserialization
Blocks attempts to execute malicious code during the Java objects deserialization process on Java-based servers.
|  | — |  | — | |JIT
Prevents an attacker from bypassing the operating system's memory mitigations using just-in-time (JIT) compilation engines.
|  |  | — | — | |Kernel Integrity Monitor (KIM)
Prevents rootkit and vulnerability exploitation on Linux endpoints. On the first detection of suspicious rootkit behavior, the behavioral threat protection (BTP) module generates a Cortex XDR Agent alert. Cortex XDR stitches logs about the process that loaded the kernel module with other logs relating to the kernel module to aid in the alert investigation. When the Cortex XDR agent detects subsequent rootkit behavior, it blocks the activity.
| — | — |  | — | |LDAP query protection
Analyzes and acts upon suspicious LDAP queries received by the Domain Controller, to detect and block Active Directory reconnaissance attacks.
|  | — | — | — | |Local analysis
Examines hundreds of characteristics of an unknown executable file, DLL, or macro to determine if it is likely to be malware. The local analysis module uses a static set of pattern-matching rules that inspect multiple file features and attributes, and a statistical model that was developed using machine learning on WildFire threat intelligence.
|  |  |  | — | |Local Threat Evaluation Engine (LTEE)
Protects against malicious PHP files arriving from the web server.
| — | — |  | — | |Local privilege escalation protection
Prevents attackers from performing malicious activities that require privileges that are higher than those assigned to the attacked or malicious process.
|  |  |  | — | |Malicious device protection
Protects your systems from unauthorized hardware attacks and malicious Human Interface Devices (HIDs) such as malicious USB devices.
| — |  | — | — | |Master Boot Record (MBR) Model
Protects against malicious Master Boot Record (MBR) manipulations.
|  | — | — | — | |Network packet inspection engine
Analyze network packet data to detect malicious behavior already at the network level. The engine leverages both Palo Alto Networks NGFW content rules, and new Cortex XDR content rules created by the Research Team which are updated through the security content.
|  |  | — | — | |Null dereference
Prevents malicious code from mapping to address zero in the memory space, making null dereference vulnerabilities unexploitable.
|  | — | — | — | |Restricted execution - local path
Prevents unauthorized execution from a local path.
|  | — | — | — | |Restricted execution - network location
Prevents unauthorized execution from a network path.
|  | — | — | — | |Restricted execution - removable media
Prevents unauthorized execution from removable media.
|  | — | — | — | |Reverse shell protection
Blocks malicious activity where an attacker redirects standard input and output streams to network sockets.
| — | — |  | — | |ROP
Protects against the use of return-oriented programming (ROP) by protecting APIs used in ROP chains.
|  |  |  | — | |SEH
Prevents hijacking of the structured exception handler (SEH), a commonly exploited control structure that can contain multiple SEH blocks that form a linked list chain, which contains a sequence of function records.
|  | — | — | — | |Shellcode protection
Reserves and protects certain areas of memory commonly used to house payloads using heap spray techniques.
| — | — |  | — | |ShellLink
Prevents shell-link logical vulnerabilities.
|  | — | — | — | |SO hijacking protection
Prevents dynamic loading of libraries from unsecured locations to gain control of a process.
| — | — |  | — | |SysExit
Prevents using system calls to bypass other protection capabilities.
|  | — | — | — | |UASLR
Improves or altogether implements ASLR (address space layout randomization) with greater entropy, robustness, and strict enforcement.
|  | — | — | — | |UEFI BTP
Reinforces the malware protection from pre-boot attacks.
|  | — | — | — | |Vulnerable drivers protection
Detect attempts to load vulnerable drivers.
|  | — | — | — | |WildFire
Leverages WildFire for threat intelligence to determine whether a file is malware. In the case of unknown files, Cortex XDR can forward samples to WildFire for in-depth analysis.
|  |  |  |  | |WildFire post-detection (malware and grayware)
Identifies a file that was previously allowed to run on an endpoint that is now determined to be malware. Post-detection events provide notifications for each endpoint on which the file is executed.
|  |  |  |  | --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://cortex-docs.paloaltonetworks.com/cortex-xdr-5.x/protect-your-endpoints/endpoint-protection/endpoint-protection-capabilities.md?ask=