#134Case study3 min
Test your app’s security before publishing
Prepare a useful pre-release check, from target selection to a reviewed correction.
Hackaiz team · updated 21/11/2026

01Start with the version you intend to publish
A pre-release security check is most useful when you can identify exactly what was examined. Write down the repository, intended release and deployment address before launching. A scan of last week’s code is not evidence about a new authentication change. If the public launch has not happened yet, you can still examine an authorized GitHub repository; testing a deployed website requires a reachable environment and validated domain ownership proof.
Use a preview or staging environment you control when appropriate, and populate it with synthetic information rather than copies of customer records. Make a short list of release-sensitive functions: signing in, accessing private documents, administering accounts and calling any AI features. This list guides your review of coverage. It does not imply that one automated analysis will exercise every business workflow or replace your application tests.
02Prepare the analysis in the actual connection screen
Open “Scan your platform”. To inspect private source code, choose “Connect GitHub”, authorize access and select the intended repository. For a website or API, validate domain ownership first using the DNS TXT record _hackaiz, the HTTPS file /.well-known/hackaiz-verification.txt, or the instruction that asks your project’s AI to create that file. Merely knowing a preview address does not authorize an analysis.
Choose checks deliberately instead of treating launch day as a reason to accept every default without reading. The interface offers automatic selection of compatible check families and a manual selection mode. Analysis consumes credits; checks you untick are not run and are not billed. If you omit a family, record that decision with the release notes so the resulting report is not mistaken for evidence about that area.
- 1Confirm the repository and deployment belong to the same intended release.
- 2Complete the required proof for the website or API before starting its analysis.
- 3Review selected checks, then launch only the scope you intend to examine.
03Turn the report into release decisions
Review Danger findings first, then To fix and To watch, while reading the supporting evidence rather than relying on the label alone. A secret exposed in browser code, a route answering without a session and a known vulnerable dependency call for different corrective work. Assign a person to each relevant finding and decide whether the affected function should remain unavailable until the issue has been addressed.
Also read what was not checked. If an analysis is interrupted or a prerequisite is missing, a partial result cannot stand in for a completed review. A report with no finding means that the completed checks did not report an issue in their scope; it does not certify the application. Keep your functional tests, permission tests and review of sensitive product behavior alongside Hackaiz evidence.
04Review changes and check the deployed result
A proposed GitHub correction is a draft pull request. A human must review and approve the change, run the project’s tests and decide on merging; Hackaiz does not merge it automatically. Check that the proposal fixes the relevant behavior without breaking a legitimate workflow. For example, a stricter access rule should still let the correct account read its own synthetic document.
After the approved change reaches the intended environment, run a relevant follow-up analysis and compare the original evidence with the new result. Keep the report together with the release identifier and the decision to publish or postpone. Further analyses consume credits too. Optional monitoring is available only for connected systems; it is not a substitute for confirming the actual release you are about to make public.
Frequently asked questions
Take the next step
Was this article helpful?