• Restrictions
• Jobs
• Scan execution policy editor
• Scan execution policies schema
• Scan execution policy schema
• pipeline rule type
• schedule rule type
  • Cadence
  • agent schema
    • Policy example
• scan action type
  • Scanner behavior
  • DAST profiles
  • Scan settings
• CI/CD variables
• Policy scope schema
• Example security policy project
• Example for scan execution policy editor
• Avoiding duplicate scans
• Experimental features

Scan execution policies

Use scan execution policies to enforce security scans, either as part of the pipeline or on a specified schedule. The security scans run with multiple project pipelines if you define the policy at a group or subgroup level.

Scan execution policies are enforced for all applicable projects. For projects without a .gitlab-ci.yml file, or where AutoDevOps is disabled, security policies create the .gitlab-ci.yml file implicitly. This ensures policies enabling execution of secret detection, static analysis, or other scanners that do not require a build in the project, are still able to run and be enforced.

This feature has some overlap with compliance pipelines, as we have not unified the user experience for these two features. For details on the similarities and differences between these features, see Enforce scan execution.

• For a video walkthrough, see How to set up Security Scan Policies in GitLab.
• For an overview, see Enforcing scan execution policies on projects with no GitLab CI/CD configuration.

Restrictions

• You can assign a maximum of five rules to each policy.
• You can assign a maximum of five scan execution policies to each security policy project.

Jobs

Policy jobs for scans, other than DAST scans, are created in the test stage of the pipeline. If you remove the test stage from the default pipeline, jobs run in the scan-policies stage instead. This stage is injected into the CI/CD pipeline at evaluation time if it doesn’t exist. If the build stage exists, it is injected just after the build stage, otherwise it is injected at the beginning of the pipeline. DAST scans always run in the dast stage. If this stage does not exist, then a dast stage is injected at the end of the pipeline.

To avoid job name conflicts, a hyphen and a number is appended to the job name. The number is unique per policy action.

Scan execution policy editor

Use the scan execution policy editor to create or edit a scan execution policy.

Prerequisites:

• Only group, subgroup, or project Owners have the permissions to select Security Policy Project.

Once your policy is complete, save it by selecting Configure with a merge request at the bottom of the editor. You are redirected to the merge request on the project’s configured security policy project. If one does not link to your project, a security policy project is automatically created. Existing policies can also be removed from the editor interface by selecting Delete policy at the bottom of the editor.

Most policy changes take effect as soon as the merge request is merged. Any changes that do not go through a merge request and are committed directly to the default branch may require up to 10 minutes before the policy changes take effect.

Scan execution policies schema

The YAML file with scan execution policies consists of an array of objects matching scan execution policy schema nested under the scan_execution_policy key. You can configure a maximum of 5 policies under the scan_execution_policy key. Any other policies configured after the first 5 are not applied.

When you save a new policy, GitLab validates its contents against this JSON schema. If you’re not familiar with how to read JSON schemas, the following sections and tables provide an alternative.

Scan execution policy schema

pipeline rule type

This rule enforces the defined actions whenever the pipeline runs for a selected branch.

1. You must specify only one of branches or branch_type.

schedule rule type

Use the schedule rule type to run security scanners on a schedule.

A scheduled pipeline:

• Runs only the scanners defined in the policy, not the jobs defined in the project’s .gitlab-ci.yml file.
• Runs according to the schedule defined in the cadence field.
• Runs under a security_policy_bot user account in the project, with the Guest role and permissions to create pipelines and read the repository’s content from a CI/CD job. This account is created when the policy is linked to a group or project.

1. You must specify only one of branches, branch_type, or agents.

Cadence

Use the cadence field to schedule when you want the policy’s actions to run. The cadence field uses cron syntax, but with some restrictions:

• Only the following types of cron syntax are supported:
  • A daily cadence of once per hour around specified time, for example: 0 18 * * *
  • A weekly cadence of once per week on a specified day and around specified time, for example: 0 13 * * 0
• Use of the comma (,), hyphens (-), or step operators (/) are not supported for minutes and hours. Any scheduled pipeline using these characters is skipped.

Consider the following when choosing a value for the cadence field:

• Timing is based on UTC for GitLab SaaS and on the GitLab host’s system time for GitLab self-managed. When testing new policies, it may appear pipelines are not running properly when in fact they are scheduled in your server’s time zone.
• A scheduled pipeline starts around the time mentioned in the policy, when the resources become available to create it. In other words, the pipeline may not begin precisely at the timing specified in the policy.

When using the schedule rule type with the agents field:

• The GitLab agent for Kubernetes checks every 30 seconds to see if there is an applicable policy. When a policy is found, the scans are executed according to the cadence defined.
• The cron expression is evaluated using the system time of the Kubernetes-agent pod.

When using the schedule rule type with the branches field:

• The cron worker runs on 15 minute intervals and starts any pipelines that were scheduled to run during the previous 15 minutes. Therefore, scheduled pipelines may run with an offset of up to 15 minutes.
• If a policy is enforced on a large number of projects or branches, the policy is processed in batches, and may take some time to create all pipelines.

agent schema

Use this schema to define agents objects in the schedule rule type.

Policy example

- name: Enforce Container Scanning in cluster connected through my-gitlab-agent for default and kube-system namespaces
  enabled: true
  rules:
  - type: schedule
    cadence: '0 10 * * *'
    agents:
      <agent-name>:
        namespaces:
        - 'default'
        - 'kube-system'
  actions:
  - scan: container_scanning

The keys for a schedule rule are:

• cadence (required): a Cron expression for when the scans are run.
• agents:<agent-name> (required): The name of the agent to use for scanning.
• agents:<agent-name>:namespaces (optional): The Kubernetes namespaces to scan. If omitted, all namespaces are scanned.

scan action type

This action executes the selected scan with additional parameters when conditions for at least one rule in the defined policy are met.

Scanner behavior

Some scanners behave differently in a scan action than they do in a regular CI/CD pipeline-based scan.

• Static Application Security Testing (SAST): Runs only if the repository contains files supported by SAST).
• Secret detection:
  • Only rules with the default ruleset are supported. Custom rulesets are not supported. Instead, you can configure a remote configuration file and set the SECRET_DETECTION_RULESET_GIT_REFERENCE variable.
  • For scheduled scan execution policies, secret detection by default runs first in historic mode (SECRET_DETECTION_HISTORIC_SCAN = true). All subsequent scheduled scans run in default mode with SECRET_DETECTION_LOG_OPTIONS set to the commit range between last run and current SHA. You can override this behavior by specifying CI/CD variables in the scan execution policy. For more information, see Full history pipeline secret detection.
  • For triggered scan execution policies, secret detection works just like regular scan configured manually in the .gitlab-ci.yml.
• Container scanning: A scan that is configured for the pipeline rule type ignores the agent defined in the agents object. The agents object is only considered for schedule rule types. An agent with a name provided in the agents object must be created and configured for the project.

DAST profiles

The following requirements apply when enforcing Dynamic Application Security Testing (DAST):

• For every project in the policy’s scope the specified site profile and scanner profile must exist. If these are not available, the policy is not applied and a job with an error message is created instead.
• When a DAST site profile or scanner profile is named in an enabled scan execution policy, the profile cannot be modified or deleted. To edit or delete the profile, you must first set the policy to Disabled in the policy editor or set enabled: false in the YAML mode.
• When configuring policies with a scheduled DAST scan, the author of the commit in the security policy project’s repository must have access to the scanner and site profiles. Otherwise, the scan is not scheduled successfully.

Scan settings

The following settings are supported by the scan_settings parameter:

CI/CD variables

Variables defined in a scan execution policy follow the standard CI/CD variable precedence.

Preconfigured values are used for the following CI/CD variables in any project on which a scan execution policy is enforced. Their values can be overridden, but only if they are declared in a policy. They cannot be overridden by group or project CI/CD variables:

DS_EXCLUDED_PATHS: spec, test, tests, tmp
SAST_EXCLUDED_PATHS: spec, test, tests, tmp
SECRET_DETECTION_EXCLUDED_PATHS: ''
SECRET_DETECTION_HISTORIC_SCAN: false
SAST_EXCLUDED_ANALYZERS: ''
DS_EXCLUDED_ANALYZERS: ''

In GitLab 16.9 and earlier:

• If the CI/CD variables suffixed _EXCLUDED_PATHS were declared in a policy, their values could be overridden by group or project CI/CD variables.
• If the CI/CD variables suffixed _EXCLUDED_ANALYZERS were declared in a policy, their values were ignored, regardless of where they were defined: policy, group, or project.

Policy scope schema

To customize policy enforcement, you can define a policy’s scope to either include, or exclude, specified projects, groups, or compliance framework labels. For more details, see Scope.

Example security policy project

You can use this example in a .gitlab/security-policies/policy.yml file stored in a security policy project:

---
scan_execution_policy:
- name: Enforce DAST in every release pipeline
  description: This policy enforces pipeline configuration to have a job with DAST scan for release branches
  enabled: true
  rules:
  - type: pipeline
    branches:
    - release/*
  actions:
  - scan: dast
    scanner_profile: Scanner Profile A
    site_profile: Site Profile B
- name: Enforce DAST and secret detection scans every 10 minutes
  description: This policy enforces DAST and secret detection scans to run every 10 minutes
  enabled: true
  rules:
  - type: schedule
    branches:
    - main
    cadence: "*/10 * * * *"
  actions:
  - scan: dast
    scanner_profile: Scanner Profile C
    site_profile: Site Profile D
  - scan: secret_detection
    scan_settings:
      ignore_default_before_after_script: true
- name: Enforce Secret Detection and Container Scanning in every default branch pipeline
  description: This policy enforces pipeline configuration to have a job with Secret Detection and Container Scanning scans for the default branch
  enabled: true
  rules:
  - type: pipeline
    branches:
    - main
  actions:
  - scan: secret_detection
  - scan: sast
    variables:
      SAST_EXCLUDED_ANALYZERS: brakeman
  - scan: container_scanning

In this example:

• For every pipeline executed on branches that match the release/* wildcard (for example, branch release/v1.2.1)
  • DAST scans run with Scanner Profile A and Site Profile B.
• DAST and secret detection scans run every 10 minutes. The DAST scan runs with Scanner Profile C and Site Profile D.
• Secret detection, container scanning, and SAST scans run for every pipeline executed on the main branch. The SAST scan runs with the SAST_EXCLUDED_ANALYZER variable set to "brakeman".

Example for scan execution policy editor

You can use this example in the YAML mode of the scan execution policy editor. It corresponds to a single object from the previous example.

name: Enforce Secret Detection and Container Scanning in every default branch pipeline
description: This policy enforces pipeline configuration to have a job with Secret Detection and Container Scanning scans for the default branch
enabled: true
rules:
  - type: pipeline
    branches:
      - main
actions:
  - scan: secret_detection
  - scan: container_scanning

Avoiding duplicate scans

Scan execution policies can cause the same type of scanner to run more than once if developers include scan jobs in the project’s .gitlab-ci.yml file. This behavior is intentional as scanners can run more than once with different variables and settings. For example, a developer may want to try running a SAST scan with different variables than the one enforced by the security and compliance team. In this case, two SAST jobs run in the pipeline, one with the developer’s variables and one with the security and compliance team’s variables.

If you want to avoid running duplicate scans, you can either remove the scans from the project’s .gitlab-ci.yml file or disable your local jobs by setting SAST_DISABLED: "true". Disabling jobs this way does not prevent the security jobs defined by scan execution policies from running.

Experimental features

This experiment has concluded and will not continue. After receiving feedback within this experiment, we will be focusing our efforts on a new policy type for enforcement of custom CI. The experiment will be removed in 17.3.

Learn more about the pipeline execution policy.