guardia-ai/gitlab-componentEU AI Act compliance scanner for GitLab CI/CD pipelines — detects AI/ML libraries and posts risk classification as MR comments.
To make it easy for you to get started with GitLab, here's a list of recommended next steps.
Already a pro? Just edit this README.md and make it your own. Want to make it easy? Use the template at the bottom!
cd existing_repo
git remote add origin https://gitlab.com/guardia-ai/gitlab-component.git
git branch -M main
git push -uf origin main
Use the built-in continuous integration in GitLab.
When you're ready to make this README your own, just edit this file and use the handy template below (or feel free to structure it however you want - this is just a starting point!). Thanks to makeareadme.com for this template.
Every project is different, so consider which of these sections apply to yours. The sections used in the template are suggestions for most open source projects. Also keep in mind that while a README can be too long and detailed, too long is better than too short. If you think your README is too long, consider utilizing another form of documentation rather than cutting out information.
Choose a self-explaining name for your project.
Let people know what your project can do specifically. Provide context and add a link to any reference visitors might be unfamiliar with. A list of Features or a Background subsection can also be added here. If there are alternatives to your project, this is a good place to list differentiating factors.
On some READMEs, you may see small images that convey metadata, such as whether or not all the tests are passing for the project. You can use Shields to add some to your README. Many services also have instructions for adding a badge.
Depending on what you are making, it can be a good idea to include screenshots or even a video (you'll frequently see GIFs rather than actual videos). Tools like ttygif can help, but check out Asciinema for a more sophisticated method.
Within a particular ecosystem, there may be a common way of installing things, such as using Yarn, NuGet, or Homebrew. However, consider the possibility that whoever is reading your README is a novice and would like more guidance. Listing specific steps helps remove ambiguity and gets people to using your project as quickly as possible. If it only runs in a specific context like a particular programming language version or operating system or has dependencies that have to be installed manually, also add a Requirements subsection.
Use examples liberally, and show the expected output if you can. It's helpful to have inline the smallest example of usage that you can demonstrate, while providing links to more sophisticated examples if they are too long to reasonably include in the README.
Tell people where they can go to for help. It can be any combination of an issue tracker, a chat room, an email address, etc.
If you have ideas for releases in the future, it is a good idea to list them in the README.
State if you are open to contributions and what your requirements are for accepting them.
For people who want to make changes to your project, it's helpful to have some documentation on how to get started. Perhaps there is a script that they should run or some environment variables that they need to set. Make these steps explicit. These instructions could also be useful to your future self.
You can also document commands to lint the code or run tests. These steps help to ensure high code quality and reduce the likelihood that the changes inadvertently break something. Having instructions for running tests is especially helpful if it requires external setup, such as starting a Selenium server for testing in a browser.
Show your appreciation to those who have contributed to the project.
For open source projects, say how it is licensed.
If you have run out of energy or time for your project, put a note at the top of the README saying that development has slowed down or stopped completely. Someone may choose to fork your project or volunteer to step in as a maintainer or owner, allowing your project to keep going. You can also make an explicit request for maintainers.
Beyond detecting which AI libraries you use, the scanner reads your source and reports specific obligations at specific lines:
| Rule | What it looks for |
|---|---|
GA-ART50-001 | A user-facing endpoint that reaches a model, with no disclosure anywhere in the repository that responses are AI-generated |
GA-ART12-001 | A model invoked with no logging, audit or tracing call in scope |
Findings appear three ways: as a comment on the merge request, as markers on the merge request diff via the Code Quality report, and — with an API key — as a record in your Guardia dashboard that tracks what you fixed and what you introduced, commit by commit.
include:
- component: gitlab.com/guardia-ai/gitlab-component/scan@main
inputs:
guardia_api_key: $GUARDIA_API_KEY # optional — keeps the record
code_analysis: 'true'
fail_on_findings: 'none'
Findings resolve themselves. Fix the code — our patch or your own — and the next scan simply stops reporting it. Nothing to click.
To accept one instead, say so in the code:
# guardia: ignore GA-ART50-001 — notice is rendered by the chat UI shell
That never fails a build, and it arrives in your dashboard as a documented risk acceptance with the author from git blame, which is what an auditor wants to see.
A five-year-old repository will have findings nobody currently on the team caused. Freeze them once, and only new work has to be clean:
guardia-scan . --write-baseline .guardia/baseline.json
Commit that file. Baselined findings stay visible in the report and in your dashboard — they just never fail the check. Anything introduced afterwards does.
Each run can write a tamper-evident record — what was found, on which commit, under which version of the rule pack, and how much legal review each rule had at the time:
- uses: GharbiiAhmed/guardia-ai-action@v1
with:
evidence-file: guardia-evidence.json
evidence-signing-key: ${{ secrets.GUARDIA_EVIDENCE_KEY }} # optional
Records chain by hash, so altering a past one breaks every record after it. Without a signing key that proves internal consistency, not authenticity — the record says so itself rather than leaving you to assume.
Findings state what your code does and quote the obligation. They do not assert that you are in breach — whether an obligation applies depends on your system's purpose and deployment context, which no code scan can determine. The rules cite Regulation (EU) 2024/1689 verbatim so you can check the reasoning yourself.
Detection runs entirely offline. Your source never leaves the runner.