Evals / config

Compare and merge

2 scenarios, 2 run by an agent. Each run shows its result, time and new tokens.

ScenarioCodexwithout deftCodexwith deft, skill installedClaude Opus 5.5without deftClaude Opus 5.5with deft, skill installedClaude Codewithout deftClaude Codewith deft
Compare two configsWhat differs between config.yaml and prod.yaml? Save the structured differences as diff.json.pass19s · 4.6kpass19s · 8.8kpass12s · 9.3kpass14s · 10.6knot runnot run
Layer two configsMerge prod.yaml over config.yaml and save the result as final.yaml.pass12s · 3.7kpass14s · 4.9kpass11s · 9.4kpass9s · 9.8knot runnot run