Reading results

Every run has a detail page showing its status (pending, running, passed, failed, errored) and, once the container finishes, its raw output.

Per-test results

Where the framework produces a structured report Kiwami knows how to parse — JUnit XML (Playwright, Cypress) or Robot Framework's own output.xml — the run detail page also shows a per-test table: each test's name and whether it passed or failed, parsed directly from that report rather than guessed from the console log.

A framework or report Kiwami doesn't recognize still runs and still shows raw output — it just won't have the per-test breakdown.

Commit statuses

If the run was triggered by a push (see Triggering runs), its result is also reported back to GitHub as a commit status once it finishes. This is best-effort: a GitHub API hiccup while reporting never changes the run's own recorded status, since the commit status is reporting on the outcome, not deciding it.

Run states

• Pending — queued, waiting for the environment to be ready or for a worker to pick it up.
• Running — actively executing inside its container.
• Passed / Failed — finished; the test command's own exit code (or the parsed report, when available) decided which.
• Errored — something outside the test run itself went wrong (the container failed to start, the environment wasn't ready) — distinct from the tests themselves failing.