* Do you work in a **cybersecurity company**? Do you want to see your **company advertised in HackTricks**? or do you want to have access to the **latest version of the PEASS or download HackTricks in PDF**? Check the [**SUBSCRIPTION PLANS**](https://github.com/sponsors/carlospolop)!
* Discover [**The PEASS Family**](https://opensea.io/collection/the-peass-family), our collection of exclusive [**NFTs**](https://opensea.io/collection/the-peass-family)
* Get the [**official PEASS & HackTricks swag**](https://peass.creator-spring.com)
* **Join the** [**💬**](https://emojipedia.org/speech-balloon/) [**Discord group**](https://discord.gg/hRep4RUj7f) or the [**telegram group**](https://t.me/peass) or **follow** me on **Twitter** [**🐦**](https://github.com/carlospolop/hacktricks/tree/7af18b62b3bdc423e11444677a6a73d4043511e9/\[https:/emojipedia.org/bird/README.md)[**@carlospolopm**](https://twitter.com/hacktricks\_live)**.**
* **Share your hacking tricks by submitting PRs to the** [**hacktricks repo**](https://github.com/carlospolop/hacktricks) **and** [**hacktricks-cloud repo**](https://github.com/carlospolop/hacktricks-cloud).
* **/Applications**: The installed apps should be here. All the users will be able to access them.
* **/bin**: Command line binaries
* **/cores**: If exists, it's used to store core dumps
* **/dev**: Everything is treated as a file so you may see hardware devices stored here.
* **/etc**: Configuration files
* **/Library**: A lot of subdirectories and files related to preferences, caches and logs can be found here. A Library folder exists in root and on each user's directory.
* **/private**: Undocumented but a lot of the mentioned folders are symbolic links to the private directory.
* **/.vol**: Running `stat a.txt` you obtain something like `16777223 7545753 -rw-r--r-- 1 username wheel ...` where the first number is the id number of the volume where the file exists and the second one is the inode number. You can access the content of this file through /.vol/ with that information running `cat /.vol/16777223/7545753`
* **System applications** are located under `/System/Applications`
* **Installed** applications are usually installed in `/Applications` or in `~/Applications`
* **Application data** can be found in `/Library/Application Support` for the applications running as root and `~/Library/Application Support` for applications running as the user.
* Third-party applications **daemons** that **need to run as root** as usually located in `/Library/PrivilegedHelperTools/`
* **Sandboxed** apps are mapped into the `~/Library/Containers` folder. Each app has a folder named according to the application’s bundle ID (`com.apple.Safari`).
* The **kernel** is located in `/System/Library/Kernels/kernel`
* **Apple's kernel extensions** are located in `/System/Library/Extensions`
* **Third-party kernel extensions** are stored in `/Library/Extensions`
### Firmlinks
The `Data` volume is mounted in **`/System/Volumes/Data`** (you can check this with `diskutil apfs list`).
The list of firmlinks can be found in the **`/usr/share/firmlinks`** file.
```bash
cat /usr/share/firmlinks
/AppleInternal AppleInternal
/Applications Applications
/Library Library
[...]
```
On the **left**, there is the directory path on the **System volume**, and on the **right**, the directory path where it maps on the **Data volume**. So, `/library` --> `/system/Volumes/data/library`
* **`$HOME/Library/Preferences/com.apple.LaunchServices.QuarantineEventsV2`**: Contains information about downloaded files, like the URL from where they were downloaded.
* **`/var/log/system.log`**: Main log of OSX systems. com.apple.syslogd.plist is responsible for the execution of syslogging (you can check if it's disabled looking for "com.apple.syslogd" in `launchctl list`.
* **Standard User:** The most basic of users. This user needs permissions granted from an admin user when attempting to install software or perform other advanced tasks. They are not able to do it on their own.
* **Admin User**: A user who operates most of the time as a standard user but is also allowed to perform root actions such as install software and other administrative tasks. All users belonging to the admin group are **given access to root via the sudoers file**.
This is a way to obtain **Alternate Data Streams in MacOS** machines. You can save content inside an extended attribute called **com.apple.ResourceFork** inside a file by saving it in **file/..namedfork/rsrc**.
The files `/System/Library/CoreServices/CoreTypes.bundle/Contents/Resources/System` contains the risk associated to files depending on the file extension.
The possible categories include the following:
* **LSRiskCategorySafe**: **Totally****safe**; Safari will auto-open after download
* **LSRiskCategoryNeutral**: No warning, but **not auto-opened**
* **LSRiskCategoryUnsafeExecutable**: **Triggers** a **warning** “This file is an application...”
* **LSRiskCategoryMayContainUnsafeExecutable**: This is for things like archives that contain an executable. It **triggers a warning unless Safari can determine all the contents are safe or neutral**.
printf "\nThe following services are OFF if '0', or ON otherwise:\nScreen Sharing: %s\nFile Sharing: %s\nRemote Login: %s\nRemote Mgmt: %s\nRemote Apple Events: %s\nBack to My Mac: %s\n\n" "$scrShrng" "$flShrng" "$rLgn" "$rmMgmt" "$rAE" "$bmM";
_**Gatekeeper**_ is designed to ensure that, by default, **only trusted software runs on a user’s Mac**. Gatekeeper is used when a user **downloads** and **opens** an app, a plug-in or an installer package from outside the App Store. Gatekeeper verifies that the software is **signed by** an **identified developer**, is **notarised** by Apple to be **free of known malicious content**, and **hasn’t been altered**. Gatekeeper also **requests user approval** before opening downloaded software for the first time to make sure the user hasn’t been tricked into running executable code they believed to simply be a data file.
In order for an **app to be notarised by Apple**, the developer needs to send the app for review. Notarization is **not App Review**. The Apple notary service is an **automated system** that **scans your software for malicious content**, checks for code-signing issues, and returns the results to you quickly. If there are no issues, the notary service generates a ticket for you to staple to your software; the notary service also **publishes that ticket online where Gatekeeper can find it**.
When the user first installs or runs your software, the presence of a ticket (either online or attached to the executable) **tells Gatekeeper that Apple notarized the software**. **Gatekeeper then places descriptive information in the initial launch dialog** indicating that Apple has already checked for malicious content.
Upon download of an application, a particular **extended file attribute** ("quarantine flag") can be **added** to the **downloaded****file**. This attribute **is added by the application that downloads the file**, such as a **web****browser** or email client, but is not usually added by others like common BitTorrent client software.\
**Checking** the **validity** of code signatures is a **resource-intensive** process that includes generating cryptographic **hashes** of the code and all its bundled resources. Furthermore, checking certificate validity involves doing an **online check** to Apple's servers to see if it has been revoked after it was issued. For these reasons, a full code signature and notarization check is **impractical to run every time an app is launched**.
Should malware make its way onto a Mac, macOS also includes technology to remediate infections. The _Malware Removal Tool (MRT)_ is an engine in macOS that remediates infections based on updates automatically delivered from Apple (as part of automatic updates of system data files and security updates). **MRT removes malware upon receiving updated information** and it continues to check for infections on restart and login. MRT doesn’t automatically reboot the Mac. (From [here](https://support.apple.com/en-gb/guide/security/sec469d47bd8/web))
Apple issues the **updates for XProtect and MRT automatically** based on the latest threat intelligence available. By default, macOS checks for these updates **daily**. Notarisation updates are distributed using CloudKit sync and are much more frequent.
**TCC (Transparency, Consent, and Control)** is a mechanism in macOS to **limit and control application access to certain features**, usually from a privacy perspective. This can include things such as location services, contacts, photos, microphone, camera, accessibility, full disk access, and a bunch more.
From a user’s perspective, they see TCC in action **when an application wants access to one of the features protected by TCC**. When this happens the user is prompted with a dialog asking them whether they want to allow access or not. This response is then stored in the TCC database.
The TCC database is just a **sqlite3 database**, which makes the task of investigating it much simpler. There are two different databases, a global one in `/Library/Application Support/com.apple.TCC/TCC.db` and a per-user one located in `/Users/<username>/Library/Application Support/com.apple.TCC/TCC.db`. The first database is **protected from editing with SIP**(System Integrity Protection), but you can read them by granting terminal(or your editor) **full disk access**.
By default an access via **SSH** will have **"Full Disk Access"**. In order to disable this you need to have it listed but disabled (removing it from the list won't remove those privileges):
MacOS Sandbox (initially called Seatbelt) makes applications run inside the sandbox **need to request access to resources outside of the limited sandbox**. This helps to ensure that **the application will be accessing only expected resources** and if it wants to access anything else it will need to ask for permissions to the user.
Any app with the **entitlement**`com.apple.security.app-sandbox` will be executed inside the sandbox. In order to publish inside the App Store, **this entitlement is mandatory**. So most applications will be executed inside the sandbox.
Some important components of the Sandbox are:
* The **kernel extension**`/System/Library/Extensions/Sandbox.kext`
* The **private framework**`/System/Library/PrivateFrameworks/AppSandbox.framework`
* A **daemon** running in userland `/usr/libexec/sandboxd`
* The **containers**`~/Library/Containers`
Inside the containers folder you can find **a folder for each app executed sanboxed** with the name of the bundle id:
```bash
ls -l ~/Library/Containers
total 0
drwx------@ 4 username staff 128 May 23 20:20 com.apple.AMPArtworkAgent
drwx------@ 4 username staff 128 May 23 20:13 com.apple.AMPDeviceDiscoveryAgent
drwx------@ 4 username staff 128 Mar 24 18:03 com.apple.AVConference.Diagnostic
drwx------@ 4 username staff 128 Mar 25 14:14 com.apple.Accessibility-Settings.extension
drwx------@ 4 username staff 128 Mar 25 14:10 com.apple.ActionKit.BundledIntentHandler
[...]
```
Inside each bundle id folder you can find the **plist** and the **Data directory** of the App:
```bash
cd /Users/username/Library/Containers/com.apple.Safari
ls -la
total 104
drwx------@ 4 username staff 128 Mar 24 18:08 .
drwx------ 348 username staff 11136 May 23 20:57 ..
-rw-r--r-- 1 username staff 50214 Mar 24 18:08 .com.apple.containermanagerd.metadata.plist
drwx------ 13 username staff 416 Mar 24 18:05 Data
ls -l Data
total 0
drwxr-xr-x@ 8 username staff 256 Mar 24 18:08 CloudKit
drwx------ 2 username staff 64 Mar 24 18:02 SystemData
drwx------ 2 username staff 64 Mar 24 18:02 tmp
```
{% hint style="danger" %}
Note that even if the symlinks are there to "escape" from the Sandbox and access other folders, the App still needs to **have permissions** to access them. These permissions are inside the **`.plist`**.
Important **system services** also run inside their own custom **sandbox** such as the mdnsresponder service. You can view these custom **sandbox profiles** written in a language called Sandbox Profile Language (SBPL) inside the **`/usr/share/sandbox`** directory. Other sandbox profiles can be checked in [https://github.com/s7ephen/OSX-Sandbox--Seatbelt--Profiles](https://github.com/s7ephen/OSX-Sandbox--Seatbelt--Profiles).
* [https://desi-jarvis.medium.com/office365-macos-sandbox-escape-fcce4fa4123c](https://desi-jarvis.medium.com/office365-macos-sandbox-escape-fcce4fa4123c) (they are able to write files outside the sandbox whose name starts with `~$`).
This protection was enabled to **help keep root level malware from taking over certain parts** of the operating system. Although this means **applying limitations to the root user** many find it to be worthwhile trade off.\
The most notable of these limitations are that **users can no longer create, modify, or delete files inside** of the following four directories in general:
Note that there are **exceptions specified by Apple**: The file **`/System/Library/Sandbox/rootless.conf`** holds a list of **files and directories that cannot be modified**. But if the line starts with an **asterisk** it means that it can be **modified** as **exception**.\
Means that `/usr`**cannot be modified****except** for the **3 allowed** folders allowed.
The final exception to these rules is that **any installer package signed with the Apple’s certificate can bypass SIP protection**, but **only Apple’s certificate**. Packages signed by standard developers will still be rejected when trying to modify SIP protected directories.
Note that if **a file is specified** in the previous config file **but** it **doesn't exist, it can be created**. This might be used by malware to obtain stealth persistence. For example, imagine that a **.plist** in `/System/Library/LaunchDaemons` appears listed but it doesn't exist. A malware may c**reate one and use it as persistence mechanism.**
Running a `ls -lO` you can find the directories protected by SIP because of the **`restricted`** flag. Moreover, directories with the **`sunlnk`** flag cannot be deleted (although files can be created and deleted inside of it).
```bash
ls -lO /
drwxr-xr-x@ 10 root wheel restricted 320 Feb 9 10:39 System
ls -lO /usr/
drwxr-xr-x 8 root wheel sunlnk 256 Apr 8 00:49 local
**SIP** handles a number of **other limitations as well**. Like it **doesn't allows for the loading of unsigned kexts**. SIP is also responsible for **ensuring** that no OS X **system processes are debugged**. This also means that Apple put a stop to dtrace inspecting system processes.
For more **information about SIP** read the following response: [https://apple.stackexchange.com/questions/193368/what-is-the-rootless-feature-in-el-capitan-really](https://apple.stackexchange.com/questions/193368/what-is-the-rootless-feature-in-el-capitan-really)
This post about a **SIP bypass vulnerability** is also very interesting: [https://www.microsoft.com/security/blog/2021/10/28/microsoft-finds-new-macos-vulnerability-shrootless-that-could-bypass-system-integrity-protection/](https://www.microsoft.com/security/blog/2021/10/28/microsoft-finds-new-macos-vulnerability-shrootless-that-could-bypass-system-integrity-protection/)
**More bypasses** in [https://jhftss.github.io/CVE-2022-26712-The-POC-For-SIP-Bypass-Is-Even-Tweetable/](https://jhftss.github.io/CVE-2022-26712-The-POC-For-SIP-Bypass-Is-Even-Tweetable/)
When checking some **malware sample** you should always **check the signature** of the binary as the **developer** that signed it may be already **related** with **malware.**
In the previous output it's possible to see that **macOS System volume snapshot is sealed** (cryptographically signed by the OS). SO, if SIP is bypassed and modifies it, the **OS won't boot anymore**.
It's also possible to verify that seal is enabled by running:
```bash
csrutil authenticated-root status
Authenticated Root status: enabled
```
Moreover, it's mounted as **read-only**:
```
mount
/dev/disk3s1s1 on / (apfs, sealed, local, read-only, journaled)
# will print all the running services under that particular user domain.
launchctl print gui/<usersUID>
# will print all the running services under root
launchctl print system
# will print detailed information about the specific launch agent. And if it’s not running or you’ve mistyped, you will get some output with a non-zero exit code: Could not find service “com.company.launchagent.label” in domain for login
An **ASEP** is a location on the system that could lead to the **execution** of a binary **without****user****interaction**. The main ones used in OS X take the form of plists.
**`launchd`** is the **first****process** executed by OX S kernel at startup and the last one to finish at shut down. It should always have the **PID 1**. This process will **read and execute** the configurations indicated in the **ASEP****plists** in:
When a user logs in the plists located in `/Users/$USER/Library/LaunchAgents` and `/Users/$USER/Library/LaunchDemons` are started with the **logged users permissions**.
The **main difference between agents and daemons is that agents are loaded when the user logs in and the daemons are loaded at system startup** (as there are services like ssh that needs to be executed before any user access the system). Also agents may use GUI while daemons need to run in the background.
There are cases where an **agent needs to be executed before the user logins**, these are called **PreLoginAgents**. For example, this is useful to provide assistive technology at login. They can be found also in `/Library/LaunchAgents`(see [**here**](https://github.com/HelmutJ/CocoaSampleCode/tree/master/PreLoginAgents) an example).
New Daemons or Agents config files will be **loaded after next reboot or using**`launchctl load <target.plist>` It's **also possible to load .plist files without that extension** with `launchctl -F <file>` (however those plist files won't be automatically loaded after reboot).\
To **ensure** that there isn't **anything** (like an override) **preventing** an **Agent** or **Daemon****from****running** run: `sudo launchctl load -w /System/Library/LaunchDaemos/com.apple.smdb.plist`
There you can find the regular **cron****jobs**, the **at****jobs** (not very used) and the **periodic****jobs** (mainly used for cleaning temporary files). The daily periodic jobs can be executed for example with: `periodic daily`.
These tasks differ from cron in that **they are one time tasks** t**hat get removed after executing**. However, they will **survive a system restart** so they can’t be ruled out as a potential threat.
By **default** they are **disabled** but the **root** user can **enable****them** with:
Apple introduced a logging mechanism called **emond**. It appears it was never fully developed, and development may have been **abandoned** by Apple for other mechanisms, but it remains **available**.
This little-known service may **not be much use to a Mac admin**, but to a threat actor one very good reason would be to use it as a **persistence mechanism that most macOS admins probably wouldn't know** to look for. Detecting malicious use of emond shouldn't be difficult, as the System LaunchDaemon for the service looks for scripts to run in only one place:
```bash
ls -l /private/var/db/emondClients
```
{% hint style="danger" %}
**As this isn't used much, anything in that folder should be suspicious**
After placing a new directory in one of these two locations, **two more items** need to be placed inside that directory. These two items are a **rc script****and a plist** that holds a few settings. This plist must be called “**StartupParameters.plist**”.
Configuration profiles can force a user to use certain browser settings, DNS proxy settings, or VPN settings. Many other payloads are possible which make them ripe for abuse.
* **`/private/var/vm/swapfile0`**: This file is used as a **cache when physical memory fills up**. Data in physical memory will be pushed to the swapfile and then swapped back into physical memory if it’s needed again. More than one file can exist in here. For example, you might see swapfile0, swapfile1, and so on.
***`/private/var/vm/sleepimage`**: When OS X goes into **hibernation**, **data stored in memory is put into the sleepimage file**. When the user comes back and wakes the computer, memory is restored from the sleepimage and the user can pick up where they left off.
In order to dump the memory in a MacOS machine you can use [**osxpmem**](https://github.com/google/rekall/releases/download/v1.5.1/osxpmem-2.1.post4.zip).
**Note**: The following instructions will only work for Macs with Intel architecture. This tool is now archived and the last release was in 2017. The binary downloaded using the instructions below targets Intel chips as Apple Silicon wasn't around in 2017. It may be possible to compile the binary for arm64 architecture but you'll have to try for yourself.
sudo osxpmem.app/osxpmem --format raw -o /tmp/dump_mem
#Dump aff4 format
sudo osxpmem.app/osxpmem -o /tmp/dump_mem.aff4
```
If you find this error: `osxpmem.app/MacPmem.kext failed to load - (libkern/kext) authentication failure (file ownership/permissions); check the system/kernel logs for errors or try kextutil(8)` You can fix it doing:
```bash
sudo cp -r osxpmem.app/MacPmem.kext "/tmp/"
sudo kextutil "/tmp/MacPmem.kext"
#Allow the kext in "Security & Privacy --> General"
sudo osxpmem.app/osxpmem --format raw -o /tmp/dump_mem
[**Scripts like this one**](https://gist.github.com/teddziuba/3ff08bdda120d1f7822f3baf52e606c2) or [**this one**](https://github.com/octomagon/davegrohl.git) can be used to transform the hash to **hashcat****format**.
The attacker still needs to gain access to the system as well as escalate to **root** privileges in order to run **keychaindump**. This approach comes with its own conditions. As mentioned earlier, **upon login your keychain is unlocked by default** and remains unlocked while you use your system. This is for convenience so that the user doesn’t need to enter their password every time an application wishes to access the keychain. If the user has changed this setting and chosen to lock the keychain after every use, keychaindump will no longer work; it relies on an unlocked keychain to function.
It’s important to understand how Keychaindump extracts passwords out of memory. The most important process in this transaction is the ”**securityd**“ **process**. Apple refers to this process as a **security context daemon for authorization and cryptographic operations**. The Apple developer libraries don’t say a whole lot about it; however, they do tell us that securityd handles access to the keychain. In his research, Juuso refers to the **key needed to decrypt the keychain as ”The Master Key“**. A number of steps need to be taken to acquire this key as it is derived from the user’s OS X login password. If you want to read the keychain file you must have this master key. The following steps can be done to acquire it. **Perform a scan of securityd’s heap (keychaindump does this with the vmmap command)**. Possible master keys are stored in an area flagged as MALLOC\_TINY. You can see the locations of these heaps yourself with the following command:
**Keychaindump** will then search the returned heaps for occurrences of 0x0000000000000018. If the following 8-byte value points to the current heap, we’ve found a potential master key. From here a bit of deobfuscation still needs to occur which can be seen in the source code, but as an analyst the most important part to note is that the necessary data to decrypt this information is stored in securityd’s process memory. Here’s an example of keychain dump output.
Based on this comment [https://github.com/juuso/keychaindump/issues/10#issuecomment-751218760](https://github.com/juuso/keychaindump/issues/10#issuecomment-751218760) it looks like these tools aren't working anymore in Big Sur.
[**Chainbreaker**](https://github.com/n0fate/chainbreaker) can be used to extract the following types of information from an OSX keychain in a forensically sound manner:
* Hashed Keychain password, suitable for cracking with [hashcat](https://hashcat.net/hashcat/) or [John the Ripper](https://www.openwall.com/john/)
* Internet Passwords
* Generic Passwords
* Private Keys
* Public Keys
* X509 Certificates
* Secure Notes
* Appleshare Passwords
Given the keychain unlock password, a master key obtained using [volafox](https://github.com/n0fate/volafox) or [volatility](https://github.com/volatilityfoundation/volatility), or an unlock file such as SystemKey, Chainbreaker will also provide plaintext passwords.
Without one of these methods of unlocking the Keychain, Chainbreaker will display all other available information.
The **kcpassword** file is a file that holds the **user’s login password**, but only if the system owner has **enabled automatic login**. Therefore, the user will be automatically logged in without being asked for a password (which isn't very secure).
The password is stored in the file **`/etc/kcpassword`** xored with the key **`0x7D 0x89 0x52 0x23 0xD2 0xBC 0xDD 0xEA 0xA3 0xB9 0x1F`**. If the users password is longer than the key, the key will be reused.\
The code of **dyld is open source** and can be found in [https://opensource.apple.com/source/dyld/](https://opensource.apple.com/source/dyld/) and cab be downloaded a tar using a **URL such as** [https://opensource.apple.com/tarballs/dyld/dyld-852.2.tar.gz](https://opensource.apple.com/tarballs/dyld/dyld-852.2.tar.gz)
> This is a colon separated **list of dynamic libraries** to l**oad before the ones specified in the program**. This lets you test new modules of existing dynamic shared libraries that are used in flat-namespace images by loading a temporary dynamic shared library with just the new modules. Note that this has no effect on images built a two-level namespace images using a dynamic shared library unless DYLD\_FORCE\_FLAT\_NAMESPACE is also used.
This technique may be also **used as an ASEP technique** as every application installed has a plist called "Info.plist" that allows for the **assigning of environmental variables** using a key called `LSEnvironmental`.
* Existence of `__RESTRICT/__restrict` section in the macho binary.
* The software has entitlements (hardened runtime) without [`com.apple.security.cs.allow-dyld-environment-variables`](https://developer.apple.com/documentation/bundleresources/entitlements/com\_apple\_security\_cs\_allow-dyld-environment-variables) entitlement or [`com.apple.security.cs.disable-library-validation`](https://developer.apple.com/documentation/bundleresources/entitlements/com\_apple\_security\_cs\_disable-library-validation).
In more updated versions you can find this logic at the second part of the function **`configureProcessRestrictions`.** However, what is executed in newer versions is the **beginning checks of the function** (you can remove the ifs related to iOS or simulation as those won't be used in macOS.
You can check if a binary has **hardenend runtime** with `codesign --display --verbose <bin>` checking the flag runtime in **`CodeDirectory`** like: **`CodeDirectory v=20500 size=767 flags=0x10000(runtime) hashes=13+7 location=embedded`**
Remember that **previous restrictions also apply** to perform Dylib hijacking attacks.
{% endhint %}
As in Windows, in MacOS you can also **hijack dylibs** to make **applications****execute****arbitrary****code**.\
However, the way **MacOS** applications **load** libraries is **more restricted** than in Windows. This implies that **malware** developers can still use this technique for **stealth**, but the probably to be able to **abuse this to escalate privileges is much lower**.
First of all, is **more common** to find that **MacOS binaries indicates the full path** to the libraries to load. And second, **MacOS never search** in the folders of the **$PATH** for libraries.
The **main** part of the **code** related to this functionality is in **`ImageLoader::recursiveLoadLibraries`** in `ImageLoader.cpp`.
However, there are **2 types of dylib hijacking**:
* **Missing weak linked libraries**: This means that the application will try to load a library that doesn't exist configured with **LC\_LOAD\_WEAK\_DYLIB**. Then, **if an attacker places a dylib where it's expected it will be loaded**.
* The fact that the link is "weak" means that the application will continue running even if the library isn't found.
* The **code related** to this is in the function `ImageLoaderMachO::doGetDependentLibraries` of `ImageLoaderMachO.cpp` where `lib->required` is only `false` when `LC_LOAD_WEAK_DYLIB` is true.
* **Find weak liked libraries** in binaries with (you have later an example on how to create hijacking libraries): 
* ```bash
otool -l </path/to/bin> | grep LC_LOAD_WEAK_DYLIB -A 5 cmd LC_LOAD_WEAK_DYLIB
cmdsize 56
name /var/tmp/lib/libUtl.1.dylib (offset 24)
time stamp 2 Wed Jun 21 12:23:31 1969
current version 1.0.0
compatibility version 1.0.0
```
* **Configured with @rpath**: Mach-O binaries can have the commands **`LC_RPATH`** and **`LC_LOAD_DYLIB`**. Base on the **values** of those commands, **libraries** are going to be **loaded** from **different directories**.
* **`LC_RPATH`** contains the paths of some folders used to load libraries by the binary. 
* **`LC_LOAD_DYLIB`** contains the path to specific libraries to load. These paths can contain **`@rpath`**, which will be **replaced** by the values in **`LC_RPATH`**. If there are several paths in **`LC_RPATH`** everyone will be used to search the library to load. Example:
* If **`LC_LOAD_DYLIB`** contains `@rpath/library.dylib` and **`LC_RPATH`** contains `/application/app.app/Contents/Framework/v1/` and `/application/app.app/Contents/Framework/v2/`. Both folders are going to be used to load `library.dylib`**.** If the library doesn't exist in `[...]/v1/` and attacker could place it there to hijack the load of the library in `[...]/v2/` as the order of paths in **`LC_LOAD_DYLIB`** is followed.
* **Find rpath paths and libraries** in binaries with: `otool -l </path/to/binary> | grep -E "LC_RPATH|LC_LOAD_DYLIB" -A 5`
{% hint style="info" %}
**`@executable_path`**: Is the **path** to the directory containing the **main executable file**. 
**`@loader_path`**: Is the **path** to the **directory** containing the **Mach-O binary** which contains the load command. 
* When used in an executable, **`@loader_path`** is effectively the **same** as **`@executable_path`**. 
* When used in a **dylib**, **`@loader_path`** gives the **path** to the **dylib**.
{% endhint %}
The way to **escalate privileges** abusing this functionality would be in the rare case that an **application** being executed **by****root** is **looking** for some **library in some folder where the attacker has write permissions.**
{% hint style="success" %}
A nice **scanner** to find **missing libraries** in applications is [**Dylib Hijack Scanner**](https://objective-see.com/products/dhs.html) or a [**CLI version**](https://github.com/pandazheng/DylibHijack).\
A nice **report with technical details** about this technique can be found [**here**](https://www.virusbulletin.com/virusbulletin/2015/03/dylib-hijacking-os-x).
* When path **does not contain a slash character** (i.e. it is just a leaf name), **dlopen() will do searching**. If **`$DYLD_LIBRARY_PATH`** was set at launch, dyld will first **look in that director**y. Next, if the calling mach-o file or the main executable specify an **`LC_RPATH`**, then dyld will **look in those** directories. Next, if the process is **unrestricted**, dyld will search in the **current working directory**. Lastly, for old binaries, dyld will try some fallbacks. If **`$DYLD_FALLBACK_LIBRARY_PATH`** was set at launch, dyld will search in **those directories**, otherwise, dyld will look in **`/usr/local/lib/`** (if the process is unrestricted), and then in **`/usr/lib/`**.
1.`$DYLD_LIBRARY_PATH`
2.`LC_RPATH`
3.`CWD`(if unrestricted)
4.`$DYLD_FALLBACK_LIBRARY_PATH`
5.`/usr/local/lib/` (if unrestricted)
6.`/usr/lib/`
* When path **looks like a framework** path (e.g. /stuff/foo.framework/foo), if **`$DYLD_FRAMEWORK_PATH`** was set at launch, dyld will first look in that directory for the framework partial path (e.g. foo.framework/foo). Next, dyld will try the **supplied path as-is** (using current working directory for relative paths). Lastly, for old binaries, dyld will try some fallbacks. If **`$DYLD_FALLBACK_FRAMEWORK_PATH`** was set at launch, dyld will search those directories. Otherwise, it will search **`/Library/Frameworks`** (on macOS if process is unrestricted), then **`/System/Library/Frameworks`**.
1.`$DYLD_FRAMEWORK_PATH`
2. supplied path (using current working directory for relative paths)
* When path **contains a slash but is not a framework path** (i.e. a full path or a partial path to a dylib), dlopen() first looks in (if set) in **`$DYLD_LIBRARY_PATH`** (with leaf part from path ). Next, dyld **tries the supplied path** (using current working directory for relative paths (but only for unrestricted processes)). Lastly, for older binaries, dyld will try fallbacks. If **`$DYLD_FALLBACK_LIBRARY_PATH`** was set at launch, dyld will search in those directories, otherwise, dyld will look in **`/usr/local/lib/`** (if the process is unrestricted), and then in **`/usr/lib/`**.
1.`$DYLD_LIBRARY_PATH`
2. supplied path (using current working directory for relative paths if unrestricted)
3.`$DYLD_FALLBACK_LIBRARY_PATH`
4.`/usr/local/lib/` (if unrestricted)
5.`/usr/lib/`
Note: If the main executable is a **set\[ug]id binary or codesigned with entitlements**, then **all environment variables are ignored**, and only a full path can be used.
#### Check paths
Lets check all the options with the following code:
Most of the interesting information is going to be in **blob**. So you will need to **extract** that content and **transform** it to **human****readable** or use **`strings`**. To access it you can do:
for i in $(sqlite3 ~/Library/Group\ Containers/group.com.apple.notes/NoteStore.sqlite "select Z_PK from ZICNOTEDATA;"); do sqlite3 ~/Library/Group\ Containers/group.com.apple.notes/NoteStore.sqlite "select writefile('body1.gz.z', ZDATA) from ZICNOTEDATA where Z_PK = '$i';"; zcat body1.gz.Z ; done
It's a scripting language used for task automation **interacting with remote processes**. It makes pretty easy to **ask other processes to perform some actions**. **Malware** may abuse these features to abuse functions exported by other processes.\
For example, a malware could **inject arbitrary JS code in browser opened pages**. Or **auto click** some allow permissions requested to the user;
Find more info about malware using applescripts [**here**](https://www.sentinelone.com/blog/how-offensive-actors-use-applescript-for-attacking-macos/).
Apple scripts may be easily "**compiled**". These versions can be easily "**decompiled**" with `osadecompile`
However, there are still some tools that can be used to understand this kind of executables, [**read this research for more info**](https://labs.sentinelone.com/fade-dead-adventures-in-reversing-malicious-run-only-applescripts/)). The tool [**applescript-disassembler**](https://github.com/Jinmo/applescript-disassembler) with [**aevt\_decompile**](https://github.com/SentineLabs/aevt\_decompile) will be very useful to understand how the script works.
Red Teaming in **environments where MacOS** is used instead of Windows can be very **different**. In this guide you will find some interesting tricks for this kind of assessments:
* Do you work in a **cybersecurity company**? Do you want to see your **company advertised in HackTricks**? or do you want to have access to the **latest version of the PEASS or download HackTricks in PDF**? Check the [**SUBSCRIPTION PLANS**](https://github.com/sponsors/carlospolop)!
* Discover [**The PEASS Family**](https://opensea.io/collection/the-peass-family), our collection of exclusive [**NFTs**](https://opensea.io/collection/the-peass-family)
* Get the [**official PEASS & HackTricks swag**](https://peass.creator-spring.com)
* **Join the** [**💬**](https://emojipedia.org/speech-balloon/) [**Discord group**](https://discord.gg/hRep4RUj7f) or the [**telegram group**](https://t.me/peass) or **follow** me on **Twitter** [**🐦**](https://github.com/carlospolop/hacktricks/tree/7af18b62b3bdc423e11444677a6a73d4043511e9/\[https:/emojipedia.org/bird/README.md)[**@carlospolopm**](https://twitter.com/hacktricks\_live)**.**
* **Share your hacking tricks by submitting PRs to the** [**hacktricks repo**](https://github.com/carlospolop/hacktricks) **and** [**hacktricks-cloud repo**](https://github.com/carlospolop/hacktricks-cloud).
