v1.14.0

1.14.0 (2026-06-03)

Features

• support remote scenarios (44c44f1)

Bug Fixes

• adjust the evals for env and agents (2fee690)
• improve the judge system prompt to stay on track (2583d1d)
• reduce visual noise on cli output with judge clarifications (1ab6e5d)

v1.13.0

1.13.0 (2026-05-21)

Features

• add fallback support for agent auth using local sessions (b1b5fec)

Bug Fixes

• add favicon (d3b241c)

v1.12.0

1.12.0 (2026-05-20)

Features

• support scoring via different agents (2c76fec)

v1.11.0

1.11.0 (2026-05-16)

Features

• isolate the home dir from working dir (5aaf474)

Bug Fixes

• agent selection when you have models too (341faa2)
• report interaction expansions (78a8955)

v1.10.0

1.10.0 (2026-05-15)

Features

• add --failed support for targeting only tests that fail (2cdbc7a)
• allow canceling in-flight evals and retain data (9b79de6)

v1.9.0

1.9.0 (2026-05-15)

Features

• rubric -> judge (b15377c)

Bug Fixes

• make it clear that things are in teardown mode (a54de01)
• show scenario key on report (6d4a5bf)
• update default concurrency to 15 (38526fb)

v1.8.1

1.8.1 (2026-05-14)

Bug Fixes

• add execution status to scoring context (e9898ea)
• gemini token counting and acp error reporting (8f1d7c2)

v1.8.0

1.8.0 (2026-05-13)

Features

• support before/after all hooks (a3cab68)

v1.7.4

1.7.4 (2026-05-13)

Bug Fixes

• improve fuzzy detection of configs (5de722c)

v1.7.3

1.7.3 (2026-05-13)

Bug Fixes

• improve the exports for scenarioinput (f4742d7)