A Repository of curated datasets from various attacks to:

• Easily develop detections without having to build an environment from scratch or simulate an attack.
• Test detections, specifically Splunks Security Content
• Replay into streaming pipelines for validating your detections in your production SIEM

Installation

GitHub LFS is used in this project. For Mac users git-lfs can be derived with homebrew (for another OS click here):

brew install git-lfs

Then you need to install it. I would recommend using the --skip-smudge parameter, which will avoid that all Git LFS files are downloaded during git clone. You can install it with the following command:

git lfs install --skip-smudge

Download the repository with this command:

git clone https://github.com/splunk/attack_data

Fetch all or select attack data sets

# This pulls all data - Warning >9Gb of data
git lfs pull

# This pulls one data set directory
git lfs pull --include=datasets/attack_techniques/T1003.001/atomic_red_team/

# Or pull just one log like this
git lfs pull --include=datasets/attack_techniques/T1003.001/atomic_red_team/windows-sysmon.log

Anatomy of a Dataset 🧬

Datasets

Datasets are defined by a common YML structure. The structure has the following fields:

field	description
id	UUID of dataset
name	name of author
date	last modified date
dataset	array of URLs where the hosted version of the dataset is located
description	describes the dataset as detailed as possible
environment	markdown filename of the environment description see below
technique	array of MITRE ATT&CK techniques associated with dataset
references	array of URLs that reference the dataset
sourcetypes	array of sourcetypes that are contained in the dataset

For example

id: 405d5889-16c7-42e3-8865-1485d7a5b2b6
author: Patrick Bareiss
date: '2020-10-08'
description: 'Atomic Test Results: Successful Execution of test T1003.001-1 Windows
  Credential Editor Successful Execution of test T1003.001-2 Dump LSASS.exe Memory
  using ProcDump Return value unclear for test T1003.001-3 Dump LSASS.exe Memory using
  comsvcs.dll Successful Execution of test T1003.001-4 Dump LSASS.exe Memory using
  direct system calls and API unhooking Return value unclear for test T1003.001-6
  Offline Credential Theft With Mimikatz Return value unclear for test T1003.001-7
  LSASS read with pypykatz '
environment: attack_range
technique:
- T1003.001
dataset:
- https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1003.001/atomic_red_team/windows-powershell.log
- https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1003.001/atomic_red_team/windows-security.log
- https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1003.001/atomic_red_team/windows-sysmon.log
- https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1003.001/atomic_red_team/windows-system.log
references:
- https://attack.mitre.org/techniques/T1003/001/
- https://github.com/redcanaryco/atomic-red-team/blob/master/atomics/T1003.001/T1003.001.md
- https://github.com/splunk/security-content/blob/develop/tests/T1003_001.yml
sourcetypes:
- XmlWinEventLog:Microsoft-Windows-Sysmon/Operational
- WinEventLog:Microsoft-Windows-PowerShell/Operational
- WinEventLog:System
- WinEventLog:Security

Environments

Environments are a description of where the dataset was collected. At this moment there are no specific restrictions, although we do have a simple template a user can start with here. The most common environment for most datasets will be the attack_range since this is the tool that used to generate attack data sets automatically.

Replay Datasets 📼

Most datasets generated will be raw log files. There are two main simple ways to ingest it.

Into Splunk

using replay.py

pre-requisite, clone, create virtual env and install python deps:

git clone git@github.com:splunk/attack_data.git
cd attack_data
pip install virtualenv
virtualenv venv
source venv/bin/activate
pip install -r bin/requirements.txt

0. Download dataset
1. configure bin/replay.yml
2. run python bin/replay.py -c bin/replay.yml

using UI

0. Download dataset
1. In Splunk enterprise , add data -> Files & Directories -> select dataset
2. Set the sourcetype as specified in the YML file
3. Explore your data

See a quick demo 📺 of this process here.

Into DSP

To send datasets into DSP the simplest way is to use the scloud command-line-tool as a requirement.

1. Download the dataset
2. Ingest the dataset into DSP via scloud command `cat attack_data.json | scloud ingest post-events --format JSON
3. Build a pipeline that reads from the firehose and you should see the events.

Contribute Datasets 🥰

1. Generate a dataset
2. Under the corresponding MITRE Technique ID folder create a folder named after the tool the dataset comes from, for example: atomic_red_Team
3. Make PR with <tool_name_yaml>.yml file under the corresponding created folder, upload dataset into the same folder.

See T1003.002 for a complete example.

Note the simplest way to generate a dataset to contribute is to launch your simulations in the attack_range, or manually attack the machines and when done dump the data using the dump function.

See a quick demo 📺 of the process to dump a dataset here.

To contribute a dataset simply create a PR on this repository, for general instructions on creating a PR see this guide.

Automatically generated Datasets ⚙️

This project takes advantage of automation to generate datasets using the attack_range. You can see details about this service on this sub-project folder attack_data_service.

Author

• Patrick Bareiß
• Jose Hernandez

License

Copyright 2023 Splunk Inc.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.