← QA home · Configuration current frontend commit: … · auto-refreshes
The retest grid, kflow-admin style: commits across the top, test cases down the side. Every cell
is real recorded evidence — kflow test:* checkpoints in this seat's stream (the same convention the
backend's katrun robot uses). Tap any cell or row for the full run detail.
First cut (user 2026-08-17): the system as a building — player-facing UI at the top,
Hedera as the foundation, build scaffolding below ground. Each box is a module (data/modules.json);
each cell is a test case's latest verdict inside the current filters. A case tagged to several modules is
end-to-end and drawn as a riser spanning the layers it touches. Dashed = specified but never run in this
window — a drawn gap, not a hidden one. Tap a cell for the run; tap a dashed cell for the definition.
✓ pass · ～ meh · ✗ fail · ⚠ error · ○ started, no verdict · ×N = N runs (latest shown). Tap a cell for detail. The full configured case list lives in Configuration. Case/commit column stays pinned while scrolling; use ⛶ to review the full grid without the page's own margins cropping it.
ℹ️ The in-page run recorder ("Record a run / give
feedback") was retired 2026-08-17 — superseded by feedback sessions (long-press → Start session on the board; see
QA-UXGRID1 / FEEDBACK-COMMENTARY1). To run an indexed test by hand, open its definition
and follow "what it does" — observations go through feedback, which lands on this grid as 📝. The
test:start/note/verdict kflow vocabulary the recorder introduced remains the convention every harness
reports in.