What is Sigma
See description at the Sigma Github repository
Sigma in Timesketch
Since early 2020 Timesketch has Sigma support implemented. Sigma can be used as an analyzer. The other option is to use Sigma via the API and the API client or the Web interface.
Sigma rules are exposed to the Web Interface as part of a sketch.
To list all Sigma rules, visit :
This will show a table with all Sigma rules installed on a system. You can search for:
- Rule ID
- OpenSearch Search query (ES Query)
- File Name
So if you want to search for ZMap related rules, you can search for
Title, File Name and it will show you the pre installed rule.
If you have run the Sigma Analyzer on a sketch and a rule has produced hits, the following fields will be added to the event:
ts_sigma_rulewill store the rule name that produced hits on an event.
ts_ttpif a rule had ATT&CK(r) tags added, they will be added to this array
To query all rules that had Sigma rules matched in an analyzer run, query for:
E.g. an event might have the following attributes:
ts_sigma_rule:[ "av_password_dumper.yml" ] ts_ttp:[ "attack.t1003.002", "attack.t1003.001", "attack.credential_access", "attack.t1558", "attack.t1003" ]
From that table, there are small icons to copy the values or explore the sketch with the given value. For example if you click the small lens icon next to the ES Query from the found rule
(data_type:("shell\:zsh\:history" OR "bash\:history\:command" OR "apt\:history\:line" OR "selinux\:line") AND "*apt\-get\ install\ zmap*") it will open an explore view for this sketch with this query pre filled for you to explore the data.
If you click the rule ID
5266a592-b793-11ea-b3de-0242ac130004 a detail view for that rule will open up.
In this detail view all key and values of that rule that has been parsed by Timesketch are exposed.
Timesketch deliberately does not provide a set of Sigma rules, as those would add complexity to maintain. To use the official community rules you can clone github.com/Neo23x0/sigma to /data/sigma. This directory will not be caught by git.
cd data git clone https://github.com/Neo23x0/sigma
The rules then will be under
Sigma Rules Blocklist file
data/sigma_blocklist.csv is where Timesketch maintains a list of rules.
Each rule can have one of the following status values:
exploratoy rules will be shown in the UI but ignored in the Analyzer. So this status can be used to test rules. By default each rule is considered
good rules will be used in the Sigma analyzer.
bad will be ignored and not shown in the UI or used in the Sigma analyzer.
It is good practice to add new rules in this file if they are tested and verified to not be compatible.
Each method that reads Sigma rules from the a folder is checking if part of the full path of a rule is mentioned in the
For example a file at
/etc/timesketch/data/sigma/rules-unsupported/foo/bar.yml would not be parsed as a line in
/rules-unsupported/,bad,Sigma internal folder name,2021-11-19,
The header for that file are:
The windows rules are stored in
The linux rules are stored in
Timesketch config file
There are multiple sigma related config variables in
# Sigma Settings SIGMA_RULES_FOLDERS = ['/etc/timesketch/sigma/rules/'] SIGMA_CONFIG = '/etc/timesketch/sigma_config.yaml' SIGMA_TAG_DELAY = 5 SIGMA_BLOCKLIST_CSV = '/etc/timesketch/sigma_blocklist.csv'
SIGMA_RULES_FOLDERS points to the folder(s) where Sigma rules are stored. The folder is the local folder of the Timesketch server (celery worker and webserver). For a distributed system, mounting network shares is possible.
SIGMA_TAG_DELAYcan be used to throttle the Sigma analyzer. If Timesketch is running on a less powerful machine (or docker-dev) a sleep timer of 15 seconds will help avoid OpenSearch Search exceptions for to many requests to the ES backend in a to short timerange. For more powerful Timesketch installations, this value can be set to 0.
In the config file
There is a section with mappings, most mappings where copied from HELK configuration. If you find a mapping missing, feel free to add and create a PR.
Some adjustments verified:
You can run the Sigma analyzer providing sample data:
python3 test_tools/analyzer_run.py --test_file test_tools/test_events/sigma_events.jsonl timesketch/lib/analyzers/sigma_tagger.py RulesSigmaPlugin
If you want to test that feature, get some evtx files from the following links and parse it via plaso
Compose new rules
In the Sigma Tab in a sketch there is a toggle called
Compose Sigma rule.
If turned on it will show a text area that takes the yaml text of a Sigma rule.
Once you are happy with your rule, click
Parse and the rule will be parsed as if it is installed on Timesketch.
This feature can be helpful if you want to test out field mapping.
From the parse result you can copy the
es_query value and paste it in a new window where you have the explore of a Sketch open.
You need to remember to copy your rule when you are ready and create a new file on your Timesketch server to store the rule and make it available to others. The text from the compose area will be resetted with each reload of the page.
When writing Rules specific for Timesketch first and foremost you should the guide from one of the creators of Sigma: How to Write Sigma Rules.
When setting the
date field in your rule, stick to
Number of or
On top of that, it is recommended to avoid a large chain or
detection: keywords: - 'value1' - 'value2' - 'value3' ... - 'value20' condition: keywords
such a query would look like:
*value1* or *value2* or *value3* ... or *value20* because that results in a very expensive / long query to execute on the dataset.
Instead it is recommended to splice it into multiple rules:
detection: keywords: - 'value1' - 'value2' - 'value3' ... - 'value10' condition: keywords
detection: keywords: - 'value11' - 'value12' - 'value13' ... - 'value20' condition: keywords
That will create two queries:
*value1* or *value2* or *value3* ... or *value10* and
*value11* or *value12* or *value13* ... or *value20*.
The Sigma analyzer is designed to batch and throttle execution of queries which is benefitial for such rule structure.
Reduce the haystack
If you can, define the haystack OpenSearch has to query. This can be achieved by adding a check for
Deploying rules that can not be parsed by Sigma can cause problems on analyst side as well as Timesketch operator side. The analyst might not be able to see the logs and the errors might only occur when running the analyzer.
This is why a standalone tool can be used from:
This tool takes the following options:
usage: sigma_verify_rules.py [-h] [--config_file PATH_TO_TEST_FILE] PATH_TO_RULES sigma_verify_rules.py: error: the following arguments are required: PATH_TO_RULES
And could be used like the following to verify your rules would work:
sigma_verify_rules.py --config_file ../data/sigma_config.yaml ../data/sigma/rules
If any rules in that folder is causing problems it will be shown:
sigma_verify_rules.py --config_file ../data/sigma_config.yaml ../timesketch/data/sigma/rules ERROR:root:reverse_shell.yaml Error generating rule in file ../timesketch/data/sigma/rules/linux/reverse_shell.yaml you should not use this rule in Timesketch: No condition found ERROR:root:recon_commands.yaml Error generating rule in file ../timesketch/data/sigma/rules/data/linux/recon_commands.yaml you should not use this rule in Timesketch: No condition found You should NOT import the following rules ../timesketch/data/sigma/rules/linux/reverse_shell.yaml ../timesketch/data/sigma/rules/linux/recon_commands.yaml
How to find issues
In the celery logs, while running the sigma analyzer, you will see something like that:
result: Applied 0 tags * win_apt_carbonpaper_turla.yml: 0 ... * win_syskey_registry_access.yml: 0 Problematic rules: XXXX
The XXX here is the "problem" and you should note those rules. Once you note and identified those rules, it is recommended to take the id and attempt a API call like the following:
from timesketch_api_client import config ts = config.get_client() rule = ts.get_sigma_rule("c0478ead-5336-46c2-bd5e-b4c84bc3a36e") print(rule.es_query)
Where the ID is the id of your problematic rule. This will hopefully give you more insight from the web server logs of what caused the problem. E.g. "Aggregations not implemented for this backend" It is then recommended to move those rules to a separate folder, maybe even creating a small shell script that does that for you once you pull upstream rules from the Sigma repository.
How to verify issues
Timesketch API / logs
If you have doubt if a rule does work, take the uuid and run python code mentioned above.
Another option is to run the rule against the official sigma client with the Timesketch sigma mapping file.
For our example from above:
sigma/tools/sigma$ python3 sigmac.py -t es-qs --config ../../../sigma_config.yaml ../../rules/windows/image_load/sysmon_mimikatz_inmemory_detection.yml An unsupported feature is required for this Sigma rule (../../rules/windows/image_load/sysmon_mimikatz_inmemory_detection.yml): Aggregations not implemented for this backend Feel free to contribute for fun and fame, this is open source :) -> https://github.com/Neo23x0/sigma
What to do with problematic rules
To reduce load on the system it is recommended to not keep the problematic rules in the directory, as it will cause the exception every time the rules folders are parsed (a lot!).
The parser is made to ignore "deprecated" folders. So you could move the problematic rules to your rules folder in a subfolder /deprecated/.
If the rules do not contain any sensitive content, you could also open an issue in the timesketch project and or in the upstream sigma project and explain your issue (best case: provide your timesketch sigma config and the rule file so it can be verified).