Skip to main content
The Tests tab shows the checks behind the stamps, laid out like an API client: the collection on the left grouped by story, the selected request in the middle, and its last run on the right.

​
Two lanes that never mix

  • Platform checks sit in a locked group — runnable and inspectable, but never editable, by you or by the agent. These are the checks the platform itself stands behind; if the agent could rewrite them, a stamp would only mean “the agent agreed with itself.”
  • Generated tests are the build’s own tests for your app’s behavior, and live in their own group below.
The two lanes arrive separately and there is no way to merge them — you can always tell which kind of check produced a verdict.

​
Running is real

Run executes the same verification path a story close uses, and the tab re-reads its results from what that run recorded. That means this tab can never show you a verdict the receipts haven’t recorded — a green here is the same green the Proof tab and the publish ledger will report.

​
What’s read-only (for now)

The request detail pane shows each request and its result read-only. Editing a request body and saving it back isn’t available yet — rather than render editors that would discard what you type, the pane shows the request as-is and says so. If a check doesn’t cover what you care about, ask for the behavior in chat; a new story brings its own checks.

​
Taking the tests with you

Your app’s API tests aren’t locked in. From the tab (and from the publish screen) you can download a Bruno collection of requests pointed at your app, plus the app’s OpenAPI contract — see Your API contract & test collection.