GitLab offers security scanning and visualization directly via and on their platform.
One nice feature is direct insights on merge requests. However, this feature is only available with the Ultimate tier. To also use this feature on the free tier, one can build around it by taking the security tool output, converting it to the JUnit format, and uploading it as JUnit report.
To summarize, this tool is for you if:
- You use GitLab's free tier
- You use Gitlabs security templates
- You want to easily access security tool output in merge requests
If you are on the GitLabs Ultimate tier, just use their tooling! No need to mess up your .gitlab-ci.yml
file. 😄
All scanning types available under the free tier:
- Secret Scanning
- Static Application Security Testing (SAST)
- Container Scanning
- Infrastructure as Code Scanning
Procedure:
- Overwrite the existing job so that the report can be used by future jobs.
- Convert report
- Upload converted report as junit report
This example can be used as is.
stages:
- test
- convert
- include:
- template: Security/Secret-Detection.gitlab-ci.yml
secret_detection:
artifacts:
paths:
- gl-secret-detection-report.json
when: always
secret_convert:
stage: convert
dependencies:
- secret_detection
script:
- pip3 install SecScanner2JUnit
- ss2ju secrets gl-secret-detection-report.json gl-secret-detection-report.xml
artifacts:
reports:
junit: gl-secret-detection-report.xml
Since GitLab decides dynamically which scanners to use depending on project languages, it makes sense to first perform a testrun only including the template. This way one can see which jobs are executed and then overwrite them.
stages:
- test
- convert
- include:
- template: Security/SAST.gitlab-ci.yml
semgrep-sast:
after_script:
- cp gl-sast-report.json gl-sast-semgrep-report.json
artifacts:
paths:
- gl-sast-semgrep-report.json
when: always
brakeman-sast:
after_script:
- cp gl-sast-report.json gl-sast-brakeman-report.json
artifacts:
paths:
- gl-sast-brakeman-report.json
when: always
semgrep-sast-convert:
stage: convert
dependencies:
- semgrep-sast
script:
- pip3 install SecScanner2JUnit
- ss2ju sast gl-sast-semgrep-report.json gl-sast-semgrep-report.xml
artifacts:
reports:
junit: gl-sast-semgrep-report.xml
brakeman-sast-convert:
stage: convert
dependencies:
- brakeman-sast
script:
- pip3 install SecScanner2JUnit
- ss2ju sast gl-sast-brakeman-report.json gl-sast-brakeman-report.xml
artifacts:
reports:
junit: gl-sast-brakeman-report.xml
- include:
- template: Jobs/Build.gitlab-ci.yml #Build and push the container image
- template: Security/Container-Scanning.gitlab-ci.yml #Scan the built image
container_scanning:
artifacts:
paths:
- gl-container-scanning-report-format.json
when: always
container_scanning-convert:
stage: convert
dependencies:
- container_scanning
script:
- pip3 install SecScanner2JUnit
- ss2ju container_scanning gl-container-scanning-report.json gl-container-scanning-report.xml
artifacts:
reports:
junit: gl-container-scanning-report.xml
You can provide a file with suppression which will allow to ignore some vulnerabilities.
You have to create a file ss2ju-config.yml
f.e. in .gitlab
directory which includes:
sast:
suppressions:
- type: "cwe"
value: "2555"
- type: "find_sec_bugs_type"
value: "SPRING_ENDPOINT"
And now you can modify execution commands as follows:
- ss2ju sast gl-sast-semgrep-report.json gl-sast-semgrep-report.xml .gitlab/ss2ju-config.yml
For easier usage in CI, Secscanner2JUnit
is also shipped in a docker container: https://hub.docker.com/r/angrymeir/secscanner2junit
Its' usage is similar to the ways described above:
...
secret_convert:
stage: convert
image:
name: angrymeir/secscanner2junit:latest
entrypoint: [""]
dependencies:
- secret_detection
script:
- ss2ju secrets gl-secret-detection-report.json gl-secret-detection-report.xml
artifacts:
reports:
junit: gl-secret-detection-report.xml
- Implement IaC Scanning