Static Application Security Testing (SAST) (ULTIMATE)

Introduced in GitLab Ultimate 10.3.

NOTE: 4 of the top 6 attacks were application based. Download our whitepaper, "A Seismic Shift in Application Security" to learn how to protect your organization.

Overview

If you are using GitLab CI/CD, you can analyze your source code for known vulnerabilities using Static Application Security Testing (SAST).

You can take advantage of SAST by doing one of the following:

Including the SAST template in your existing .gitlab-ci.yml file.
Implicitly using Auto SAST provided by Auto DevOps.

GitLab checks the SAST report, compares the found vulnerabilities between the source and target branches, and shows the information right on the merge request.

The results are sorted by the priority of the vulnerability:

Critical
High
Medium
Low
Unknown
Everything else

Use cases

Your code has a potentially dangerous attribute in a class, or unsafe code that can lead to unintended code execution.
Your application is vulnerable to cross-site scripting (XSS) attacks that can be leveraged to unauthorized access to session data.

Requirements

To run a SAST job, by default, you need GitLab Runner with the docker or kubernetes executor running in privileged mode. If you're using the shared Runners on GitLab.com, this is enabled by default.

Privileged mode is not necessary if you've disabled Docker in Docker for SAST

CAUTION: Caution: Our SAST jobs currently expect a Linux container type. Windows containers are not yet supported.

CAUTION: Caution: If you use your own Runners, make sure that the Docker version you have installed is not 19.03.0. See troubleshooting information for details.

Supported languages and frameworks

The following table shows which languages, package managers and frameworks are supported and which tools are used.

Language (package managers) / framework	Scan tool	Introduced in GitLab Version
.NET	Security Code Scan	11.0
Any	Gitleaks and TruffleHog	11.9
Apex (Salesforce)	pmd	12.1
C/C++	Flawfinder	10.7
Elixir (Phoenix)	Sobelow	11.10
Go	Gosec	10.7
Groovy (Ant, Gradle, Maven and SBT)	SpotBugs with the find-sec-bugs plugin	11.3 (Gradle) & 11.9 (Ant, Maven, SBT)
Java (Ant, Gradle, Maven and SBT)	SpotBugs with the find-sec-bugs plugin	10.6 (Maven), 10.8 (Gradle) & 11.9 (Ant, SBT)
JavaScript	ESLint security plugin	11.8
Kubernetes manifests	Kubesec	12.6
Node.js	NodeJsScan	11.1
PHP	phpcs-security-audit	10.8
Python (pip)	bandit	10.3
React	ESLint react plugin	12.5
Ruby on Rails	brakeman	10.3
Scala (Ant, Gradle, Maven and SBT)	SpotBugs with the find-sec-bugs plugin	11.0 (SBT) & 11.9 (Ant, Gradle, Maven)
TypeScript	TSLint config security	11.9

NOTE: Note: The Java analyzers can also be used for variants like the Gradle wrapper, Grails and the Maven wrapper.

Contribute your scanner

The Security Scanner Integration documentation explains how to integrate other security scanners into GitLab.

Configuration

NOTE: Note: You don't have to configure SAST manually as shown in this section if you're using Auto SAST provided by Auto DevOps.

For GitLab 11.9 and later, to enable SAST you must include the SAST.gitlab-ci.yml template that is provided as a part of your GitLab installation. For GitLab versions earlier than 11.9, you can copy and use the job as defined that template.

Add the following to your .gitlab-ci.yml file:

include:
  - template: SAST.gitlab-ci.yml

The included template will create a sast job in your CI/CD pipeline and scan your project's source code for possible vulnerabilities.

The results will be saved as a SAST report artifact that you can later download and analyze. Due to implementation limitations, we always take the latest SAST artifact available. Behind the scenes, the GitLab SAST Docker image is used to detect the languages/frameworks and in turn runs the matching scan tools.

Customizing the SAST settings

The SAST settings can be changed through environment variables by using the variables parameter in .gitlab-ci.yml.

In the following example, we include the SAST template and at the same time we set the SAST_GOSEC_LEVEL variable to 2:

include:
  - template: SAST.gitlab-ci.yml

variables:
  SAST_GOSEC_LEVEL: 2

Because the template is evaluated before the pipeline configuration, the last mention of the variable will take precedence.

Overriding the SAST template

CAUTION: Deprecation: Beginning in GitLab 13.0, the use of only and except is no longer supported. When overriding the template, you must use rules instead.

If you want to override the job definition (for example, change properties like variables or dependencies), you need to declare a sast job after the template inclusion and specify any additional keys under it. For example:

include:
  - template: SAST.gitlab-ci.yml

sast:
  variables:
    CI_DEBUG_TRACE: "true"

Using environment variables to pass credentials for private repositories

Some analyzers require downloading the project's dependencies in order to perform the analysis. In turn, such dependencies may live in private Git repositories and thus require credentials like username and password to download them. Depending on the analyzer, such credentials can be provided to it via custom environment variables.

Using a variable to pass username and password to a private Maven repository

If you have a private Maven repository which requires login credentials, you can use the MAVEN_CLI_OPTS environment variable.

Read more on how to use private Maven repos.

Disabling Docker in Docker for SAST

You can avoid the need for Docker in Docker by running the individual analyzers. This does not require running the executor in privileged mode. For example:

include:
  - template: SAST.gitlab-ci.yml

variables:
  SAST_DISABLE_DIND: "true"

This will create individual <analyzer-name>-sast jobs for each analyzer that runs in your CI/CD pipeline.

By removing Docker-in-Docker (DIND), GitLab relies on Linguist to start relevant analyzers depending on the detected repository language(s) instead of the orchestrator. However, there are some differences in the way repository languages are detected between DIND and non-DIND. You can observe these differences by checking both Linguist and the common library. For instance, Linguist looks for *.java files to spin up the spotbugs image, while orchestrator only looks for the existence of pom.xml, build.xml, gradlew, grailsw, or mvnw. GitLab uses Linguist to detect new file types in the default branch. This means that when introducing files or dependencies for a new language or package manager, the corresponding scans won't be triggered in the MR and will only run on the default branch once the MR is merged. This will be addressed by #211702.

NOTE: Note: With the current language detection logic, any new languages or frameworks introduced within the context of a merge request don't trigger a corresponding scan. These scans only occur once the code is committed to the default branch.

Enabling kubesec analyzer

Introduced in GitLab Ultimate 12.6.

When Docker in Docker is disabled, you will need to set SCAN_KUBERNETES_MANIFESTS to "true" to enable the kubesec analyzer. In .gitlab-ci.yml, define:

include:
  - template: SAST.gitlab-ci.yml

variables:
  SAST_DISABLE_DIND: "true"
  SCAN_KUBERNETES_MANIFESTS: "true"

Pre-compilation

If your project requires custom build configurations, it can be preferable to avoid compilation during your SAST execution and instead pass all job artifacts from an earlier stage within the pipeline. This is the current strategy when requiring a before_script execution to prepare your scan job.

To pass your project's dependencies as artifacts, the dependencies must be included in the project's working directory and specified using the artifacts:path configuration. If all dependencies are present, the -compile=false flag can be provided to the analyzer and compilation will be skipped:

image: maven:3.6-jdk-8-alpine

stages:
 - build
 - test

include:
  - template: SAST.gitlab-ci.yml

variables:
  SAST_DISABLE_DIND: "true"

build:
  stage: build
  script:
    - mvn package -Dmaven.repo.local=./.m2/repository
  artifacts:
    paths:
      - .m2/
      - target/

spotbugs-sast:
  dependencies:
    - build
  script:
    - /analyzer run -compile=false
  variables:
    MAVEN_REPO_PATH: ./.m2/repository
  artifacts:
    reports:
      sast: gl-sast-report.json

NOTE: Note: The path to the vendored directory must be specified explicitly to allow the analyzer to recognize the compiled artifacts. This configuration can vary per analyzer but in the case of Java above, MAVEN_REPO_PATH can be used. See Analyzer settings for the complete list of available options.

Available variables

SAST can be configured using environment variables.

Custom Certificate Authority

To trust a custom Certificate Authority, set the ADDITIONAL_CA_CERT_BUNDLE variable to the bundle of CA certs that you want to trust within the SAST environment.

Docker images

The following are Docker image-related variables.

Environment variable	Description
`SAST_ANALYZER_IMAGE_PREFIX`	Override the name of the Docker registry providing the default images (proxy). Read more about customizing analyzers.
`SAST_ANALYZER_IMAGE_TAG`	DEPRECATED: Override the Docker tag of the default images. Read more about customizing analyzers.
`SAST_DEFAULT_ANALYZERS`	Override the names of default images. Read more about customizing analyzers.
`SAST_DISABLE_DIND`	Disable Docker in Docker and run analyzers individually.

Vulnerability filters

Some analyzers make it possible to filter out vulnerabilities under a given threshold.

Environment variable	Default value	Description
`SAST_EXCLUDED_PATHS`	-	Exclude vulnerabilities from output based on the paths. This is a comma-separated list of patterns. Patterns can be globs, or file or folder paths (for example, `doc,spec` ). Parent directories will also match patterns.
`SAST_BANDIT_EXCLUDED_PATHS`	-	comma-separated list of paths to exclude from scan. Uses Python's `fnmatch` syntax
`SAST_BRAKEMAN_LEVEL`	1	Ignore Brakeman vulnerabilities under given confidence level. Integer, 1=Low 3=High.
`SAST_FLAWFINDER_LEVEL`	1	Ignore Flawfinder vulnerabilities under given risk level. Integer, 0=No risk, 5=High risk.
`SAST_GITLEAKS_ENTROPY_LEVEL`	8.0	Minimum entropy for secret detection. Float, 0.0 = low, 8.0 = high.
`SAST_GOSEC_LEVEL`	0	Ignore gosec vulnerabilities under given confidence level. Integer, 0=Undefined, 1=Low, 2=Medium, 3=High.
`SAST_GITLEAKS_COMMIT_FROM`	-	The commit a gitleaks scan starts at.
`SAST_GITLEAKS_COMMIT_TO`	-	The commit a gitleaks scan ends at.
`SAST_GITLEAKS_HISTORIC_SCAN`	false	Flag to enable a historic gitleaks scan.

Docker-in-Docker orchestrator

The following variables configure the Docker-in-Docker orchestrator.

Environment variable	Default value	Description
`SAST_ANALYZER_IMAGES`		Comma-separated list of custom images. Default images are still enabled. Read more about customizing analyzers. Not available when Docker-in-Docker is disabled.
`SAST_PULL_ANALYZER_IMAGES`	1	Pull the images from the Docker registry (set to 0 to disable). Read more about customizing analyzers. Not available when Docker-in-Docker is disabled.
`SAST_DOCKER_CLIENT_NEGOTIATION_TIMEOUT`	2m	Time limit for Docker client negotiation. Timeouts are parsed using Go's `ParseDuration`. Valid time units are "ns", "us" (or "µs"), "ms", "s", "m", "h". For example, "300ms", "1.5h" or "2h45m".
`SAST_PULL_ANALYZER_IMAGE_TIMEOUT`	5m	Time limit when pulling the image of an analyzer. Timeouts are parsed using Go's `ParseDuration`. Valid time units are "ns", "us" (or "µs"), "ms", "s", "m", "h". For example, "300ms", "1.5h" or "2h45m".
`SAST_RUN_ANALYZER_TIMEOUT`	20m	Time limit when running an analyzer. Timeouts are parsed using Go's `ParseDuration`. Valid time units are "ns", "us" (or "µs"), "ms", "s", "m", "h". For example, "300ms", "1.5h" or "2h45m".

NOTE: Note: Timeout variables are not applicable for setups with disabled Docker In Docker.

Analyzer settings

Some analyzers can be customized with environment variables.

Environment variable	Analyzer	Description
`SCAN_KUBERNETES_MANIFESTS`	kubesec	Set to `"true"` to scan Kubernetes manifests when Docker in Docker is disabled.
`ANT_HOME`	spotbugs	The `ANT_HOME` environment variable.
`ANT_PATH`	spotbugs	Path to the `ant` executable.
`GRADLE_PATH`	spotbugs	Path to the `gradle` executable.
`JAVA_OPTS`	spotbugs	Additional arguments for the `java` executable.
`JAVA_PATH`	spotbugs	Path to the `java` executable.
`SAST_JAVA_VERSION`	spotbugs	Which Java version to use. Supported versions are `8` and `11`. Defaults to `8`.
`MAVEN_CLI_OPTS`	spotbugs	Additional arguments for the `mvn` or `mvnw` executable.
`MAVEN_PATH`	spotbugs	Path to the `mvn` executable.
`MAVEN_REPO_PATH`	spotbugs	Path to the Maven local repository (shortcut for the `maven.repo.local` property).
`SBT_PATH`	spotbugs	Path to the `sbt` executable.
`FAIL_NEVER`	spotbugs	Set to `1` to ignore compilation failure.

Custom environment variables

Introduced in GitLab Ultimate 12.5.

In addition to the aforementioned SAST configuration variables, all custom environment variables are propagated to the underlying SAST analyzer images if the SAST vendored template is used.

CAUTION: Caution: Variables having names starting with these prefixes will not be propagated to the SAST Docker container and/or analyzer containers: DOCKER_, CI, GITLAB_, FF_, HOME, PWD, OLDPWD, PATH, SHLVL, HOSTNAME.

Reports JSON format

CAUTION: Caution: The JSON report artifacts are not a public API of SAST and their format may change in the future.

The SAST tool emits a JSON report file. Here is an example of the report structure with all important parts of it highlighted:

{
  "version": "2.0",
  "vulnerabilities": [
    {
      "id": "9e96e0ab-23da-4d7d-a09e-0acbaa5e83ca",
      "category": "sast",
      "name": "Predictable pseudorandom number generator",
      "message": "Predictable pseudorandom number generator",
      "description": "The use of java.util.Random is predictable",
      "severity": "Medium",
      "confidence": "Medium",
      "scanner": {
        "id": "find_sec_bugs",
        "name": "Find Security Bugs"
      },
      "location": {
        "file": "groovy/src/main/groovy/com/gitlab/security_products/tests/App.groovy",
        "start_line": 47,
        "end_line": 47,
        "class": "com.gitlab.security_products.tests.App",
        "method": "generateSecretToken2",
        "dependency": {
          "package": {}
        }
      },
      "identifiers": [
        {
          "type": "find_sec_bugs_type",
          "name": "Find Security Bugs-PREDICTABLE_RANDOM",
          "value": "PREDICTABLE_RANDOM",
          "url": "https://find-sec-bugs.github.io/bugs.htm#PREDICTABLE_RANDOM"
        },
        {
          "type": "cwe",
          "name": "CWE-330",
          "value": "330",
          "url": "https://cwe.mitre.org/data/definitions/330.html"
        }
      ]
    },
    {
      "id": "e6dbf91f-4c07-46f7-a365-0169489c27d1",
      "category": "sast",
      "message": "Probable insecure usage of temp file/directory.",
      "severity": "Medium",
      "confidence": "Medium",
      "scanner": {
        "id": "bandit",
        "name": "Bandit"
      },
      "location": {
        "file": "python/hardcoded/hardcoded-tmp.py",
        "start_line": 10,
        "end_line": 10,
        "dependency": {
          "package": {}
        }
      },
      "identifiers": [
        {
          "type": "bandit_test_id",
          "name": "Bandit Test ID B108",
          "value": "B108",
          "url": "https://docs.openstack.org/bandit/latest/plugins/b108_hardcoded_tmp_directory.html"
        }
      ]
    },
  ],
  "remediations": []
}

CAUTION: Deprecation: Beginning with GitLab 12.9, SAST no longer reports undefined severity and confidence levels.

Here is the description of the report file structure nodes and their meaning. All fields are mandatory in the report JSON unless stated otherwise. Presence of optional fields depends on the underlying analyzers being used.

Report JSON node	Function
`version`	Report syntax version used to generate this JSON.
`vulnerabilities`	Array of vulnerability objects.
`vulnerabilities[].id`	Unique identifier of the vulnerability.
`vulnerabilities[].category`	Where this vulnerability belongs (SAST, Dependency Scanning etc.). For SAST, it will always be `sast`.
`vulnerabilities[].name`	Name of the vulnerability, this must not include the occurrence's specific information. Optional.
`vulnerabilities[].message`	A short text that describes the vulnerability, it may include the occurrence's specific information. Optional.
`vulnerabilities[].description`	A long text that describes the vulnerability. Optional.
`vulnerabilities[].cve`	(DEPRECATED - use `vulnerabilities[].id` instead) A fingerprint string value that represents a concrete occurrence of the vulnerability. It's used to determine whether two vulnerability occurrences are same or different. May not be 100% accurate. This is NOT a CVE.
`vulnerabilities[].severity`	How much the vulnerability impacts the software. Possible values: `Undefined` (an analyzer has not provided this information), `Info`, `Unknown`, `Low`, `Medium`, `High`, `Critical`.
`vulnerabilities[].confidence`	How reliable the vulnerability's assessment is. Possible values: `Undefined` (an analyzer has not provided this information), `Ignore`, `Unknown`, `Experimental`, `Low`, `Medium`, `High`, `Confirmed`.
`vulnerabilities[].solution`	Explanation of how to fix the vulnerability. Optional.
`vulnerabilities[].scanner`	A node that describes the analyzer used to find this vulnerability.
`vulnerabilities[].scanner.id`	Id of the scanner as a snake_case string.
`vulnerabilities[].scanner.name`	Name of the scanner, for display purposes.
`vulnerabilities[].location`	A node that tells where the vulnerability is located.
`vulnerabilities[].location.file`	Path to the file where the vulnerability is located. Optional.
`vulnerabilities[].location.start_line`	The first line of the code affected by the vulnerability. Optional.
`vulnerabilities[].location.end_line`	The last line of the code affected by the vulnerability. Optional.
`vulnerabilities[].location.class`	If specified, provides the name of the class where the vulnerability is located. Optional.
`vulnerabilities[].location.method`	If specified, provides the name of the method where the vulnerability is located. Optional.
`vulnerabilities[].identifiers`	An ordered array of references that identify a vulnerability on internal or external DBs.
`vulnerabilities[].identifiers[].type`	Type of the identifier. Possible values: common identifier types (among `cve`, `cwe`, `osvdb`, and `usn`) or analyzer-dependent ones (e.g., `bandit_test_id` for Bandit analyzer).
`vulnerabilities[].identifiers[].name`	Name of the identifier for display purposes.
`vulnerabilities[].identifiers[].value`	Value of the identifier for matching purposes.
`vulnerabilities[].identifiers[].url`	URL to identifier's documentation. Optional.

Secret detection

GitLab is also able to detect secrets and credentials that have been unintentionally pushed to the repository (for example, an API key that allows write access to third-party deployment environments).

This check is performed by a specific analyzer during the sast job. It runs regardless of the programming language of your app, and you don't need to change anything to your CI/CD configuration file to turn it on. Results are available in the SAST report.

GitLab currently includes Gitleaks and TruffleHog checks.

NOTE: Note: The secrets analyzer will ignore "Password in URL" vulnerabilities if the password begins with a dollar sign ($) as this likely indicates the password being used is an environment variable. For example, https://username:$password@example.com/path/to/repo will not be detected, whereas https://username:password@example.com/path/to/repo would be detected.

Security Dashboard

The Security Dashboard is a good place to get an overview of all the security vulnerabilities in your groups, projects and pipelines. Read more about the Security Dashboard.

Interacting with the vulnerabilities

Once a vulnerability is found, you can interact with it. Read more on how to interact with the vulnerabilities.

Vulnerabilities database update

For more information about the vulnerabilities database update, check the maintenance table.

Running SAST in an offline environment

For self-managed GitLab instances in an environment with limited, restricted, or intermittent access to external resources through the internet, some adjustments are required for the SAST job to successfully run. For more information, see Offline environments.

Requirements for offline SAST

To use SAST in an offline environment, you need:

Disable Docker-In-Docker
GitLab Runner with the docker or kubernetes executor.
Docker Container Registry with locally available copies of SAST analyzer images.

NOTE: Note: GitLab Runner has a default pull policy of always, meaning the runner will try to pull Docker images from the GitLab container registry even if a local copy is available. GitLab Runner's pull_policy can be set to if-not-present in an offline environment if you prefer using only locally available Docker images. However, we recommend keeping the pull policy setting to always as it will better enable updated scanners to be utilized within your CI/CD pipelines.

Make GitLab SAST analyzer images available inside your Docker registry

For SAST with all supported languages and frameworks, import the following default SAST analyzer images from registry.gitlab.com to your local "offline" registry:

registry.gitlab.com/gitlab-org/security-products/analyzers/bandit:2
registry.gitlab.com/gitlab-org/security-products/analyzers/brakeman:2
registry.gitlab.com/gitlab-org/security-products/analyzers/eslint:2
registry.gitlab.com/gitlab-org/security-products/analyzers/flawfinder:2
registry.gitlab.com/gitlab-org/security-products/analyzers/go-ast-scanner:2
registry.gitlab.com/gitlab-org/security-products/analyzers/gosec:2
registry.gitlab.com/gitlab-org/security-products/analyzers/kubesec:2
registry.gitlab.com/gitlab-org/security-products/analyzers/nodejs-scan:2
registry.gitlab.com/gitlab-org/security-products/analyzers/phpcs-security-audit:2
registry.gitlab.com/gitlab-org/security-products/analyzers/pmd-apex:2
registry.gitlab.com/gitlab-org/security-products/analyzers/secrets:2
registry.gitlab.com/gitlab-org/security-products/analyzers/security-code-scan:2
registry.gitlab.com/gitlab-org/security-products/analyzers/sobelow:2
registry.gitlab.com/gitlab-org/security-products/analyzers/spotbugs:2
registry.gitlab.com/gitlab-org/security-products/analyzers/tslint:2

The process for importing Docker images into a local offline Docker registry depends on your network security policy. Please consult your IT staff to find an accepted and approved process by which external resources can be imported or temporarily accessed. Note that these scanners are updated periodically with new definitions, so consider if you are able to make periodic updates yourself.

For details on saving and transporting Docker images as a file, see Docker's documentation on docker save, docker load, docker export, and docker import.

Set SAST CI job variables to use local SAST analyzers

Override SAST environment variables to use to your local container registry as the source for SAST analyzer images.

For example, assuming a local Docker registry repository of localhost:5000/analyzers:

include:
- template: SAST.gitlab-ci.yml

variables:
SAST_ANALYZER_IMAGE_PREFIX: "localhost:5000/analyzers"
SAST_DISABLE_DIND: "true"

The SAST job should now use local copies of the SAST analyzers to scan your code and generate security reports without requiring internet access.

Troubleshooting

Error response from daemon: error processing tar file: docker-tar: relocation error

This error occurs when the Docker version used to run the SAST job is 19.03.0. You are advised to update to Docker 19.03.1 or greater. Older versions are not affected. Read more in this issue.