Your first eval
An eval suite is a set of cases and the scorers that judge them. This page builds one working suite end to end, runs it, and reads the result.
Create a suite
Scaffold a suite, then add your first case:
•
omic init — creates omic.toml and a cases directory
•
cases — each case is an input and what you expect back
•
scorers — decide whether a case passed, and by how much
Run it
Run the suite and read the verdict:
•
omic run — runs every case and writes a .eno file
•
omic run --baseline main — compare against a previous run
What you get back
Every run produces the same three things:
•
A verdict — pass or fail per case, plus an aggregate score.
•
A diff — what changed against the baseline, case by case.
•
A .eno file — the full run, on disk, ready to compare later.