CB-Threat-Hunting

Report Bug · Request Feature

About The Project

Initially started by @0xAnalyst, I will try to update this repo with my humble contributions.
It is meant to provide security analysts a 'cheatsheet' or 'guideline' to be used alongside CarbonBlack Response

This repository will contain mainly 2 things:

• Detections: Search queries to help security analysts finding malicious activities.
• LiveResponse: Response actions via CB Live Response feature to help security analysts performing incident response activities.
• Scripts: Scripts leveraging the the CB Response API to perform various actions.

Please don't hesitate to propose any additional queries and response actions for inclusion in this repository.

Roadmap

• Redefine and document the git structure
• Define rule format
• Create a script to automate watchlist creation via API
• Document usage guide
• Update old detections content and format
• Script to generate MITRE ATT&CK Navigator based on yaml watchlist (in progress)
• Update createWatchlist script to automatically tick action boxes in CB (email, alert, syslog)
• Create a LiveQuery section GH-1
• Create new content (continuous)

See the open issues for a full list of proposed features (and known issues).

Detection

Detection Engineering Tips

1. In the sensor group settings, under "Advanced", make sure that the "Retention Maximization" setting is set to "Minimum Retention", this would ensure that all processes are available for search. Of course, that is only if you can afford this settings, this means that you should send the telemetry to a central logging server (SIEM for example) to ensure proper retention.
2. Paths are sometimes difficult to search for in Solr backend, here are some important examples:\
   • In path, filemod, modload or regmod fields, you can use a forward slash / or double backslash \\, such as:
     regmod:software/microsoft/windows/currentversion/run*
   • In cmdline field, the forward slash doesn't work, only the double backslash \\ works. However, as opposed to the previous bullet point, because of the way the tokenization works, you can't use the wildcard at the end, you will have to combine multiple search fields such as:
     cmdline:"Software\\Microsoft\\Windows\\CurrentVersion\\Run" OR cmdline:"Software\\Microsoft\\Windows\\CurrentVersion\\RunOnce"
3. For some of the detections to work, especially those around Powershell activites, you will need to check the box "Fileless script loads" under the sensor group settings.

Detection rule format

The following describes the rule format used in this repository to document detection rules.
YAML is the format and it is meant to be easy to use. It is also mapped with the API /v1/watchlist so that it can be easily automated via cbapi python module or for example a SOAR.

Filename

The rule filename must adhere to the following rules:

• Does not contain space (use _)
• Must contain the tactic name, followed by
• ...the technique ID, followed by
• ...the name of the watchlist, and
• ...must ends by .yaml (or .yml)

Examples

• No existing technique
  • Try at least to pick the tactic relevant to your detection such as: c2_unsigned_process_running_from_programdata_with_network_connection.yaml
• One tactic, one technique
  • defense_evasion_t1202_usage_of_winrs.yaml
• Multiple tactics and techniques
  • persistence_t1505.003_discovery_t1082_webshell_detection_via_commands_executed_from_web_server_process.yaml

Format

name: 
index_type: 
description: 
references:
    -
    - 
tags: 
search_query: 
on_hit: 

name

String API argument when creating a watchlist.
This represents the name of the watchlist.

index_type

String API argument when creating a watchlist.
This represents the index in which to search for this watchlist, its value must either be events (for processes) or modules (for binaries).

Values:

• events
• modules

description

String API argument when creating a watchlist.
Description of the watchlist. The more precise the better. It should also include things like false positives.

references

This is not an API argument. However it's recommended to include it as part of the description field.
Reference to links and documents. For example is the rule is a translation from a Sigma rule, the sigma rule reference should be documented here.

tags

This is not an API argument. However it's recommended to include it as part of the description field.
This represents the tags that can be added to the watchlist. MITRE ATT&CK tactics, techniques and subtechniques should be added here.

search_query

String API argument when creating a watchlist.
This represents the query that will search into the index_type specified.

on_hit

List API argument when creating the watchlist.
This represents the different checkbox that can be checked to define the action trigger to fire when the query hits.

Values:

• email
• alert
• syslog

This field can also be used to understand the maturity of the rule.

• If it contains only syslog, this means the rule has to be fine tuned or it is meant to be a hunt
• If it contains at least alert, this mean that the rule has been tested in production environment and is not prone to (too much) false positives

Example

c2_defense_evasion_t1197_t1105_file_download_via_bitsadmin.yaml

name: C2, T1105, Defense Evasion T1197 File Download via Bitsadmin Usage
index_type: events
description: Detects the usage of the Windows built-in bitsadmin to download a file (especially with the transfer argument).
references:
    - https://github.com/SigmaHQ/sigma/blob/master/rules/windows/process_creation/proc_creation_win_bitsadmin_download.yml, https://attack.mitre.org/techniques/T1197/, https://attack.mitre.org/techniques/T1105/
tags: c2, t1105, defense_evasion, t1197, s1090
search_query: process_name:"bitsadmin.exe" AND cmdline:"/transfer"
on_hit: alert, syslog

LiveResponse

You will find here the cheatsheet/guideline that will help security analysts on how to perform various response actions through the CarbonBlack EDR Live Response feature.
The response actions categories are inspired from the ones defined in the RE&CT framework.
Here they are in alphabetic order:

• 📧 Email
• 📄 File
• 🔑 Identity
• 🌐 Network
• 👾 Process
• 💻 System (Configuration in RE&CT)

Scripts

Detection automation

In CarbonBlack world, a detection is called a watchlist.
Once you've selected and downloaded the watchlists of your interest, place them all in a single folder. You can then use the script watchlist_create.py to "bulk" create all the selected watchlists in your CB Response instance via API.

The script will concatenate the fields 'references' and 'tags' into the 'description' field and call the /v1/watchlist API endpoint to create the watchlists selected.
You can read about those fields in Detection rule format.

Thanks to @mahmoudawni88, the script can now automatically leverage the on_hit field in order to check the action triggers of the watchlist (email, alert, syslog).\

Script result for 2 yaml files

POST requests

Acknowledgments

• @mahmoudawni88
• @0xAnalyst

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement".

License

Distributed under the MIT License. See LICENSE.txt for more information.