Release checks

Release checks is a single orchestrated pre-release run: pick the categories you care about, and BugBrain conducts its engines to actually verify them, then returns an honest, coverage-aware launch verdict. This guide covers running one and reading the result. It is a metered add-on, so it must be enabled for your workspace first.

What it is#

A release check is one run that coordinates several engines against your app and scores the outcome per category:

1. Functionality — exploratory or Premium test runs exercise the app's flows.
2. Accessibility — WCAG checks at the level you select.
3. Compatibility — cross-environment behavior.
4. Security and Performance — drawn from their add-ons, when enabled.

It then scores each category honestly — unverified checks cap the score, never a fake 100 — and writes an AI Release Brief with a Ready / Review / Not-ready verdict.

Why use it#

• One verdict, many engines — instead of reading five separate reports, you get a single coverage-aware launch call.
• Honest by design — the score reflects what was actually verified, so it can't read green on thin coverage.
• Gate-aware — blocking problems (regressions, failed critical flows) drive the verdict first, the headline number second.

Before you start#

Release checks is a metered, feature-flag-gated add-on. Before it works for your workspace:

• A super-admin must turn on the release-checks feature flag and grant a monthly quota above zero. If the flag is off or the quota is 0, the project's Release Checks tab shows a "not enabled" notice with an upgrade or contact path.
• You need a project that BugBrain can reach and test.
• Security and performance categories require their respective add-ons to be enabled to be verified.

Run a release check#

1. 1

   Start a release check

   In the project's Release Checks tab, click Start release check.
2. 2

   Choose categories

   Select which categories to verify — functionality, accessibility, compatibility, and (if enabled) security and performance.
3. 3

   Set the WCAG level

   Pick your WCAG conformance level, and tick Security/Performance if those add-ons are enabled.
4. 4

   Read the verdict

   When the run finishes, read the per-category scores and the AI Release Brief's Ready / Review / Not-ready verdict.

How the score works#

The release-check score is coverage-aware. Each category is scored on what was actually verified, and items that weren't verified cap the score — so a category that ran little can never read 100. This is the opposite of a pass-rate over a handful of checks: it surfaces thin coverage instead of hiding it behind a green number.

How the verdict is decided#

The launch verdict is gate-first:

• Blocking gates come first — introduced regressions or failed critical flows push the verdict toward Not-ready regardless of the score.
• The coverage-aware score comes second, mapping the remaining picture to Ready, Review, or Not-ready.

A clean score with a failed critical flow is not "Ready"; the gate wins.

Release health across projects#

The workspace-level Release Health page rolls up the latest release-check verdicts and trends across every project, so you can see which apps are launch-ready at a glance and track how a project's verdict moves release over release.

Related#