jamf/aftermath

About

Aftermath is a Swift-based, open-source incident response framework.

Aftermath can be leveraged by defenders in order to collect and subsequently analyze the data from the compromised host. Aftermath can be deployed from an MDM (ideally), but it can also run independently from the infected user's command line.

Aftermath first runs a series of modules for collection. The output of this will either be written to the location of your choice, via the -o or --output option, or by default, it is written to the /tmp directory.

Once collection is complete, the final zip/archive file can be pulled from the end user's disk. This file can then be analyzed using the --analyze argument pointed at the archive file. The results of this will be written to the /tmp directory. The administrator can then unzip that analysis directory and see a parsed view of the locally collected databases, a timeline of files with the file creation, last accessed, and last modified dates (if they're available), and a storyline which includes the file metadata, database changes, and browser information to potentially track down the infection vector.

Build

To build Aftermath locally, clone it from the repository

git clone https://github.com/jamf/aftermath.git

cd into the Aftermath directory

cd <path_to_aftermath_directory>

Build using Xcode

xcodebuild -scheme "aftermath"

cd into the Release folder

cd build/Release

Run aftermath

sudo ./aftermath

Usage

Aftermath needs to be root, as well as have full disk access (FDA) in order to run. FDA can be granted to the Terminal application in which it is running.

The default usage of Aftermath runs

sudo ./aftermath

To specify certain options

sudo ./aftermath [option1] [option2]

Examples

sudo ./aftermath -o /Users/user/Desktop --deep

sudo ./aftermath --analyze <path_to_collection_zip>

External Unified Log Predicates

As of v1.2.0, users have the ability to pass Aftermath a text file of unified log predicates using the --logs or -l arguments. The file being passed to Aftermath is required to be a text file and each predicate needs to be newline-separated. In addition, each line item will be a dictionary object. The key in the dictionary will whatever the user desires to call this predicate. For example, if you want to see all login events, we will create a predicate and title it login_events.

login_events: processImagePath contains "loginwindow" and eventMessage contains "com.apple.sessionDidLogin
tcc: process == "tccd"

Releases

There is an Aftermath.pkg available under Releases. This pkg is signed and notarized. It will install the aftermath binary at /usr/local/bin/. This would be the ideal way to deploy via MDM. Since this is installed in bin, you can then run aftermath like

sudo aftermath [option1] [option2]

Uninstall

To uninstall the aftermath binary, run the AftermathUninstaller.pkg from the Releases. This will uninstall the binary and also run aftermath --cleanup to remove aftermath directories. If any aftermath directories reside elsewhere, from using the --output command, it is the responsibility of the user/admin to remove said directories.

Help Menu

--analyze -> analyze the results of the Aftermath results
     usage: --analyze <path_to_aftermath_collection_file>
--collect-dirs -> specify locations of (space-separated) directories to dump those raw files
    usage: --collect-dirs <path_to_dir> <path_to_another_dir>
--deep or -d -> perform a deep scan of the file system for modified and accessed timestamped metadata
    WARNING: This will be a time-intensive, memory-consuming scan.
--logs -> specify an external text file with unified log predicates (as dictionary objects) to parse
    usage: --logs /Users/<USER>/Desktop/myPredicates.txt
-o or --output -> specify an output location for Aftermath collection results (defaults to /tmp)
     usage: -o Users/user/Desktop
--pretty -> colorize Terminal output
--cleanup -> remove Aftermath folders from default locations ("/tmp", "/var/folders/zz/) 

Contributors

• Stuart Ashenbrenner
• Jaron Bradley
• Maggie Zirnhelt
• Matt Benyo
• Ferdous Saljooki

Thank You

This project leverages the open source TrueTree project, written and licensed by Jaron Bradley.