Results & Evidence

Read a run's verdict, the report behind it, and the screenshots, video, and traces that prove it

Every run ends with a verdict and the evidence for it. The point is that you shouldn't have to reproduce a failure to believe it.

The Run Detail Page

Tab	What's there
Report	The verdict, each test's result, the failures in detail, and the evidence attached to them
Agent activity	What the agent did, stage by stage, including the transcript of its reasoning
Knowledge graph	What this run used and what it learned about your application

The Report tab is available once the run completes. Agent activity is watchable while it's still going.

Run metadata sits above the tabs: what triggered it, when it started, how long it took, how many tests it covered, and how many artifacts it produced.

Reading the Verdict

Outcome	What it means
Passed	Every test in the run met its expected outcome
Failed	At least one test did not
App regression	A test failed in a way that points at the application, while the script itself is still valid
Error	The run couldn't complete — the environment, not the assertion
Canceled	Someone stopped it

The distinction between a failed test and an app regression is the useful one: the first may mean your script drifted out of date with the interface, the second says the script did exactly what it always does and the application answered differently. One is maintenance; the other is a bug.

Evidence

Failures carry their evidence inline, and you can open any of it full-screen and page through it.

Kind	What it's good for
Screenshot	The state of the page at the moment it went wrong
Video	The whole approach, when the failure is about how you got there
Playwright trace	Step-by-step replay with DOM snapshots, for when the screenshot isn't enough
Request / response	What the app actually sent and received
Console log	The errors the browser saw

To open a trace, download the ZIP and run it locally:

npx playwright show-trace trace.zip

Taking the Report Elsewhere

Action	Result
Copy report	The report as text, for a ticket or a chat message
Export PDF	The full report as a file, for people who don't have TestDrive open
Fix with Cursor	Hands the failure to your editor with the context already assembled

Next Steps

Goal	Documentation
Understand what the run learned	Application Knowledge
Repair or rebuild a script	Tests & Scripts
Run this automatically next time	Running Tests