Evals

Replay recordings or spar against synthetic personas to catch agent regressions.

Total runs
0
Pass
0
Fail
0
Partial / running
0

Recent runs

Personas (0)

No personas yet. The core ships five built-ins on startup — if you see this, the seed didn't run.

Recordings (0)

No recordings yet. From the Observability page, click any session row and use “Save as recording”.