Panther Analysis Tool Commands

Use PAT to manage your Panther content

Overview

You can manage your Panther detection content using the Panther Analysis Tool (PAT). PAT lets you upload, test, and delete assets, among other actions.

Each of the PAT commands accepts certain options. For example, you can use --filter with several of the commands to narrow the scope of the action.

PAT commands

See the full list of available PAT commands in the following codeblock. Beneath it, find additional information about several of the commands.

PAT commands can be run using panther_analysis_tool or pat. Learn more about these aliases here.

To understand which Panther permissions you need to execute each PAT command, see Permissions required per command.

% panther_analysis_tool -h
usage: panther_analysis_tool [-h] [--version] [--debug] {release,test,publish,upload,delete,update-custom-schemas,test-lookup-table,validate,zip,check-connection,sdk,benchmark,enrich-test-data} ...

Panther Analysis Tool: A command line tool for managing Panther policies and rules.

positional arguments:
  {release,test,publish,upload,delete,update-custom-schemas,test-lookup-table,validate,zip,check-connection,sdk,benchmark,enrich-test-data}
    release             Create release assets for repository containing panther detections. Generates a file called panther-analysis-all.zip and optionally generates panther-analysis-all.sig
    test                Validate analysis specifications and run policy and rule tests.
    publish             Publishes a new release, generates the release assets, and uploads them. Generates a file called panther-analysis-all.zip and optionally generates panther-analysis-all.sig
    upload              Upload specified policies and rules to a Panther deployment.
    delete              Delete policies, rules, or saved queries from a Panther deployment
    update-custom-schemas
                        Update or create custom schemas on a Panther deployment.
    test-lookup-table   Validate a Lookup Table spec file.
    validate            Validate your bulk uploads against your panther instance
    zip                 Create an archive of local policies and rules for uploading to Panther.
    check-connection    Check your Panther API connection
    sdk                 Perform operations using the Panther SDK exclusively (pass sdk --help for more)
    benchmark           Performance test one rule against one of its log types. The rule must be the only item in the working directory or specified by --path, --ignore-files, and --filter. This feature is an extension of Data Replay and is subject to the same limitations.
    enrich-test-data    Enrich test data with additional enrichments from the Panther API.

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --debug

test: Running unit tests

Use PAT to load the defined specification files and evaluate unit tests locally:

panther_analysis_tool test --path <folder-name>

To filter rules or policies based on certain attributes:

panther_analysis_tool test --path <folder-name> --filter RuleID=Category.Behavior.MoreInfo

Running pat test for correlation rules and Simple Detections requires an API token. See Authenticating with an API token for more information.

benchmark: Evaluating rule performance

You can use benchmark to test the performance of an existing or draft rule against one hour of data, for one log type. It can be particularly useful to iterate on a rule that is timing out. This is a long-running command intended to be used manually, as needed—not in a regular CI/CD workflow.

The API token used with this command must be granted the "Read Panther Metrics" (also known as SummaryRead) and "Manage Rules" (also known as RuleModify) permissions. Because benchmark is an extension of Data Replay, it is subject to the same limitations.

You must provide a single rule to benchmark, either by having just one rule in the working directory (./ or --path), or through the use of --ignore-files or --filter.

If you do not specify a certain hour of data, the system will select the historical hour with the highest volume of data. To specify a specific hour to run against, use --hour. (Most common time formats are supported, e.g., 2023-07-31T09:00:00-7:00—minutes, seconds, etc. will be truncated). For example:

panther_analysis_tool benchmark --hour <datetime>

For a rule with multiple log types, one must be specified using --log-type. For example:

panther_analysis_tool benchmark --log-type <log-type>

The output of benchmark will be written to both stdout, and to the directory indicated by the --out option.

enrich-test-data: Enriching test data with Enrichment content

Use enrich-test-data to enrich the test content of your Rules and Scheduled Rules with data from connected Enrichment providers and custom Lookup Tables. This allows you to build more sophisticated test cases for detections that rely on enrichment content.

enrich-test-data is simple to use, but may introduce substantial changes to your analysis YAML files. The command will modify files based on the following criteria:

  • If the Rule or Scheduled Rule does not have test cases, the YAML file will not be modified.

  • If the log type does not support enrichment, the YAML file will not be modified.

  • If the log type supports enrichment and there are test cases:

    • Test cases represented as inline JSON content will be reformatted into YAML.

    • The YAML file will be formatted according to common YAML conventions, using two spaces for indentation.

Similar to other commands, enrich-test-data works from the current directory, recursively. If you run the command at the root directory of your panther-analysis copy, it will attempt to enrich all Rules and Scheduled Rules. To enrich content in a single directory, navigate to that directory before running the command.

You can run enrich-test-data in PAT versions 0.26 and beyond using the following command:

panther_analysis_tool enrich-test-data

The output of the command will be written to stdout, including a list of any Rules or Scheduled Rules that were enriched.

validate: Ensuring detection content is ready to be uploaded

The validate command verifies your detection content is ready to be uploaded to your Panther instance by running the same checks that happen during the upload process. Because some of these checks require configuration information in your Panther instance, validate makes an API call.

To validate your detections against your Panther instance using PAT:

  1. Run the following command:

    panther_analysis_tool validate --path <path-to-your-detections> --api-token <your-api-token> --api-host https://api.<your-panther-instance-name>.runpanther.net/public/graphql
    • You may exclude the --api-token and --api-host options if you are setting configuration values another way, i.e., by using environment variables or a configuration file.

zip: Creating a package to upload to the Panther Console

To create a package for uploading manually to the Panther Console, run the following command:

$ panther_analysis_tool zip --path tests/fixtures/valid_policies/ --out tmp
[INFO]: Testing analysis packs in tests/fixtures/valid_policies/

AWS.IAM.MFAEnabled
	[PASS] Root MFA not enabled fails compliance
	[PASS] User MFA not enabled fails compliance

[INFO]: Zipping analysis packs in tests/fixtures/valid_policies/ to tmp
[INFO]: <current working directory>/tmp/panther-analysis-2020-03-23T12-48-18.zip

Uploading content in the Panther Console

  1. In the lefthand side of the Panther Console, click Detections.

  2. Click the Upload button in the upper right corner.

  3. Drag and drop your .zip file onto the page, or click Select file.

upload: Uploading packages to Panther directly

Starting with PAT version 0.22.0, if you have authenticated with an API token and execute the upload command, an asynchronous bulk upload will automatically be performed, to prevent timeout issues.

If you did not use an API token to authenticate, you can use the --batch option. The --batch option is only available in versions of PAT after 0.19.0.

The upload command uploads your detection content to your Panther instance.

To use upload:

  1. Run panther_analysis_tool test to ensure your unit tests are passing.

  2. Run the following command: panther_analysis_tool upload --path <path-to-your-detections> --api-token <your-api-token> --api-host https://api.<your-panther-instance-name>.runpanther.net/public/graphql

    • You may exclude the --api-token and --api-host options if you are setting configuration values another way, i.e., by using environment variables or a configuration file.

When using upload, detections and Lookup Tables with existing IDs are overwritten. Locally deleted detections will not automatically be deleted in your Panther instance on upload—they must be removed with the delete command (or manually deleted in your Panther Console). When using the CLI workflow, it's recommended to set a detection's Enabled property to false, instead of deleting.

If you update the ID of an entity (i.e., the value of RuleId or PolicyId) and use upload—but do not also use delete to manually remove the old entity, both versions will exist in your Panther instance. If you intend to merely update the ID without creating a duplicate detection, use delete with the old ID.

delete: Deleting Rules, Policies, or Saved Queries

While panther_analysis_tool upload --path <directory> will upload everything from <directory>, it will not delete anything in your Panther instance if you simply remove a local file from <directory>. Instead, you can use the panther_analysis_tool delete command to explicitly delete detections from your Panther instance. To delete a specific detection, you can run the following command:

panther_analysis_tool delete --analysis-id MyRuleId

This will interactively ask you for a confirmation before it deletes the detection. If you would like to delete without confirming, you can use the following command:

panther_analysis_tool delete --analysis-id MyRuleId --no-confirm

You can delete up to 1000 detections at once with PAT.

update-custom-schemas: Creating or updating custom schemas

Use update-custom-schemas to create or update custom schemas.

After using this command to create a schema, wait at least 15 minutes before using upload to upload detections that reference the new schema.

Permissions required per command

Below is a mapping of permissions required for each command.

PAT command options (sub commands)

See the options for each of the PAT commands in the codeblock below.

$ panther_analysis_tool release -h
usage: panther_analysis_tool release [-h] [--api-token API_TOKEN] [--api-host API_HOST] [--aws-profile AWS_PROFILE] [--filter KEY=VALUE [KEY=VALUE ...]]
                                     [--ignore-files IGNORE_FILES [IGNORE_FILES ...]] [--kms-key KMS_KEY] [--minimum-tests MINIMUM_TESTS] [--out OUT] [--path PATH]
                                     [--skip-tests] [--skip-disabled-tests] [--available-destination AVAILABLE_DESTINATION] [--sort-test-results]
                                     [--ignore-table-names]

optional arguments:
  -h, --help            show this help message and exit
  --api-token API_TOKEN
                        The Panther API token to use. See: https://docs.panther.com/api-beta
  --api-host API_HOST   The Panther API host to use. See: https://docs.panther.com/api-beta
  --aws-profile AWS_PROFILE
                        The AWS profile to use when updating the AWS Panther deployment.
  --filter KEY=VALUE [KEY=VALUE ...]
  --ignore-files IGNORE_FILES [IGNORE_FILES ...]
                        Relative path to files in this project to be ignored by panther-analysis tool, space separated. Example ./foo.yaml ./bar/baz.yaml
  --kms-key KMS_KEY     The key id to use to sign the release asset.
  --minimum-tests MINIMUM_TESTS
                        The minimum number of tests in order for a detection to be considered passing. If a number greater than 1 is specified, at least one True and
                        one False test is required.
  --out OUT             The path to store output files.
  --path PATH           The relative path to Panther policies and rules.
  --skip-tests          Skip testing before uploading
  --skip-disabled-tests Skip testing disabled detections before uploading
  --available-destination AVAILABLE_DESTINATION
                        A destination name that may be returned by the destinations function. Repeat the argument to define more than one name.
  --sort-test-results   Sort test results by whether the test passed or failed (passing tests first), then by rule ID
  --ignore-table-names  Allows skipping of table names from schema validation. Useful when querying non-Panther or non-Snowflake tables
  --valid-table-names   VALID_TABLE_NAMES [VALID_TABLE_NAMES ...]
                        Fully qualified table names that should be considered valid during schema validation (in addition to standard Panther/Snowflake tables), space
                        separated. Accepts '*' as wildcard character matching 0 or more characters. Example foo.bar.baz bar.baz.* foo.*bar.baz baz.* *.foo.*

--filter: Filtering PAT commands

The test, zip, upload, and release commands all support filtering. Filtering works by passing the --filter argument with a list of filters specified in the format KEY=VALUE1,VALUE2. The keys can be any valid field in a policy or rule. When using a filter, only analysis that matches each filter specified will be considered.

For example, the following command will test only items with the AnalysisType of policy AND the severity of High:

panther_analysis_tool test --path tests/fixtures/valid_policies --filter AnalysisType=policy Severity=High
[INFO]: Testing analysis packs in tests/fixtures/valid_policies

AWS.IAM.BetaTest
	[PASS] Root MFA not enabled fails compliance
	[PASS] User MFA not enabled fails compliance

The following command will test items with the AnalysisType policy OR rule, AND the severity High:

panther_analysis_tool test --path tests/fixtures/valid_policies --filter AnalysisType=policy,rule Severity=High
[INFO]: Testing analysis packs in tests/fixtures/valid_policies

AWS.IAM.BetaTest
	[PASS] Root MFA not enabled fails compliance
	[PASS] User MFA not enabled fails compliance

AWS.CloudTrail.MFAEnabled
	[PASS] Root MFA not enabled fails compliance
	[PASS] User MFA not enabled fails compliance

When writing policies or rules that refer to the global analysis types, be sure to include them in your filter. You can include an empty string as a value in a filter, and it will mean the filter is only applied if the field exists.

The following command will return an error, because the policy in question imports a global but the global does not have a severity so it is excluded by the filter:

panther_analysis_tool test --path tests/fixtures/valid_policies --filter AnalysisType=policy,global Severity=Critical
[INFO]: Testing analysis packs in tests/fixtures/valid_policies

AWS.IAM.MFAEnabled
	[ERROR] Error loading module, skipping

Invalid: tests/fixtures/valid_policies/example_policy.yml
	No module named 'panther'

[ERROR]: [('tests/fixtures/valid_policies/example_policy.yml', ModuleNotFoundError("No module named 'panther'"))]

For this query to work as expected, you need to allow for the severity field to be absent:

panther_analysis_tool test --path tests/fixtures/valid_policies --filter AnalysisType=policy,global Severity=Critical,""
[INFO]: Testing analysis packs in tests/fixtures/valid_policies

AWS.IAM.MFAEnabled
	[PASS] Root MFA not enabled fails compliance
	[PASS] User MFA not enabled fails compliance

Filters work for the zip, upload, and release commands in the same way they work for the test command.

--minimum-tests: Requiring a certain number of unit tests

You can set a minimum number of unit tests with the --minimum-tests flag. Detections that don't have the minimum number of tests will be considered failing, and if --minimum-tests is set to 2 or greater it will also enforce that at least one test must return True and one must return False.

In the example below, even though the rules passed all their tests, they're still considered failing because they do not have the correct test coverage:

panther_analysis_tool test --path tests/fixtures/valid_policies --minimum-tests 2
% panther_analysis_tool test --path okta_rules --minimum-tests 2
[INFO]: Testing analysis packs in okta_rules

Okta.AdminRoleAssigned
	[PASS] Admin Access Assigned

Okta.BruteForceLogins
	[PASS] Failed login

Okta.GeographicallyImprobableAccess
	[PASS] Non Login
	[PASS] Failed Login

--------------------------
Panther CLI Test Summary
	Path: okta_rules
	Passed: 0
	Failed: 3
	Invalid: 0

--------------------------
Failed Tests Summary
	Okta.AdminRoleAssigned
		['Insufficient test coverage, 2 tests required but only 1 found.', 'Insufficient test coverage: expected at least one passing and one failing test.']

	Okta.BruteForceLogins
		['Insufficient test coverage, 2 tests required but only 1 found.', 'Insufficient test coverage: expected at least one passing and one failing test.']

	Okta.GeographicallyImprobableAccess
		['Insufficient test coverage: expected at least one passing and one failing test.']

Last updated