Latest run
Buckets are exclusive: every test counted exactly once. Total maps to the pytest summary line.
The verdict card mirrors the same shape Verdikt produces for every real case decision: a paper artifact stamped with the outcome, meta below the line, no ambiguity about what the system decided.
By area of the codebase
Click a suite to see per-file counts and durations. Each suite is a coherent slice of the product: auth, pipeline, inbox, templates, boundaries, MCP, data, workers.
Recent runs
Rolling history of the last 50 pytest runs, newest first. Each row is one python reports/_gen_dashboard.py invocation; a run is uniquely keyed by commit and JUnit timestamp so re-generating from the same results.xml does not create a duplicate. Click a row to expand the per-suite breakdown for that run.
| When | Commit | Tests | Passed | Failed | Errored | Skipped | Duration |
|---|