Sigma Filters

When deploying Sigma rules out to your organization, you may find that some Sigma rules are noisier than others, or that a Sigma rule needs to be updated to only detect upon a given set of host computers, or user accounts.

Sigma Filters are a new way to build lists of these "exclusions" without having to modify the Sigma rule itself – meaning keeping each Sigma rule can now remain in its original state. This makes updating from Sigma's build-in rule-packs even easier.

Filters are in ALPHA

In 2024, Sigma Introduced a new feature called Filters. This feature is still in ALPHA. However, filters should work in most backends out of the box, without backend-specific support required.

If you find any issues with Filters, please open an issue here.

Overview

Sigma Filters resemble Sigma rules in their structure, but use the filter keyword in place of detection.

title: Filter Out Domain Controllers
description: Filter out events from Domain Controllers
logsource:
    product: windows
filter: 
  rules:
    - files_added_to_an_archive_using_rar_exe
    - login_on_jump_host
  selection:
    ComputerName|startswith: "DC-"
  condition: not selection

Writing Sigma Filters is nearly identical to writing Sigma, but allow you to split out your exclusions into separate files – as well as target multiple rules using the same filter.

Filters are used to define a set of conditions that can be applied to Sigma rules to exclude or include specific events based on the conditions defined in the filter.

Usage

When applying Filters in your "detection-as-code" strategy, filters can be applied recursively when converting your Sigma rules for use in your SIEM.

./filters/win_filter_admins.yml

title: Filter Out Administrator accounts
description: Filters out administrator accounts that start with adm_
logsource:
  category: process_creation
  product: windows
filter:
  rules:
    - proc_creation_win_sc_create_service
  selection:
    User|startswith: "adm_"
  condition: not selection

./rules/windows/process_creation/proc_creation_win_sc_create_service.yml

title: New Service Creation Using Sc.EXE
name: proc_creation_win_sc_create_service
description: Detects the creation of a new service using the "sc.exe" utility.
author: Timur Zinniatullin, Daniil Yugoslavskiy, oscd.community
logsource:
  category: process_creation
  product: windows
detection:
  selection:
    Image|endswith: '\sc.exe'
    CommandLine|contains|all:
      - "create"
      - "binPath"
  condition: selection
falsepositives:
  - Legitimate administrator or user creates a service for legitimate reasons.
  - Software installation
level: low

In the example above, the win_filter_admins.yml filter is applied to the proc_creation_win_sc_create_service.yml Sigma rule. The filter will exclude any events where the User field starts with adm_. The filter references the Sigma rule in the rules list under filter section, using the name field within the Sigma rule.

To apply a filter when converting a Sigma rule, use the --filter option in Sigma CLI, followed by the path to the filter folder or file. You can apply multiple filters by specifying a directory containing multiple filter files.

Here's an example of running the Sigma rule with the filter:

sigma convert -t splunk --pipeline splunk_windows \
  --filter ./filters/win_filter_admins.yml \ // [!code highlight]
  ./rules/windows/process_creation/proc_creation_win_sc_create_service.yml

The resulting Splunk query will include the filter condition:

Image="*\\sc.exe" \
  CommandLine="*create*" \
  CommandLine="*binPath*" \
  NOT User="adm_*"

Filters and Field Mapping

Every field that is used in Filters also gets mapped according to any supplied pipelines.

Filters support referencing the initial rule using the id or name field.

Design

If you've been using Sigma for detections, it's likely that you've come across this pattern to filter out known false-positives at the detection stage.

detection:
  selection:
    Image|endswith: '\rundll32.exe'
  registry:
    TargetObject|contains: '\Control Panel\Desktop\SCRNSAVE.EXE'
    Details|endswith: ".scr"
  filter:
    Details|contains:
      - 'C:\Windows\System32\'
      - 'C:\Windows\SysWOW64\'
  condition: selection and registry and not filter
falsepositives:
  - Legitimate use of screen saver

During the conversion process, Sigma Filters complete a similar process, which will append to each referenced Sigma rule's condition field, allowing you to exclude or include specific events based on the conditions defined in the filter.

detection:
    ...
    condition: selection and (filter)
#              |             |
#              | Sigma Rule  |
#                            | Sigma Filter

This allows Sigma Filters to either be used to exclude specific events – such as to tune your Sigma rule for false-positives or known-safe alerts, or to include only specific events.

The advantage of using Sigma Filters is that you can now build a set of exclusions that can be applied to multiple Sigma rules, without having to modify the Sigma rule itself – making it easier to update from Sigma's build-in rule-packs.

Example of Exclusion:

./filters/exclude_something.yml

filter:
    exclude_these_matches:
        field: value
    # This will exclude events where field equals value
    condition: not exclude_these_matches

Example of Inclusion:

./filters/include_something.yml

filter:
    include_these_matches:
        field: value
    # This will scope the detection to only include events where field equals value
    condition: include_these_matches

Filters Logsource Matching

Filters can be applied to any Sigma rule, but only where the logsource field is an exact match the logsource field in the filter.

Logsource Matching Behaviour

Currently, you can specify a logsource of a lower specificity in the filter than in the Sigma rule, but this behaviour isn't guaranteed and may change in the coming releases.

This is to facilitate the use of filters across multiple Sigma rules of "similar" log sources – such as all Windows logs, or all logs from a specific product.

# Filter
logsource:
  product: windows

# Sigma Rule
logsource:
  category: process_creation
  product: windows

The SigmaHQ team will be assessing this going forward to determine if this behaviour is beneficial or not.