agent-lab · evaluation platform

GAUNTLET

Raw UI agent versus full UI harness — rendered, judged, and replayable.

Make Omni frontend agents dependable enough that Chris can focus on product ideas while the harness handles UI standards, visible proof, and review pressure.

What we're answering The plan & progress Leaderboard

Phases

0/6

complete

Tasks

0/25

done

Open questions

0 of 5

Updated

2026-06-27

Build phases

p0done

Archive and reset the mission

Retire broad/raw-model runs from the active story without deleting evidence.

p1done

Goal and checkpoint spec

Make the autonomous endpoint explicit enough that Chris does not have to babysit each step.

p2done

Dashboard and live run surface

Make the UI dashboard show the active goal, checkpoints, archived status, and animated run progress.

p3done

Render proof pipeline

Prove the lab can see and judge what Chris sees before spending on comparisons.

p4active

Raw vs harness pilot

Produce the first reputable side-by-side Omni UI report.

p5active

Tune and confirm

Improve the UI harness from observed failures, then prove the lift holds out of sample.