1 Commits

Author SHA1 Message Date
Matt Van Horn 0270c12352 feat(eval): research-quality regression harness with scored fixtures in CI (#799)
* feat(eval): research-quality regression harness with scored fixtures in CI

* fix: address self-review findings

* fix: harden eval replay (coherence floor, metadata-merge replay, offline star enrichment)

* fix: value-scrub module fixture recordings, per-fixture floors, pinned overlap predicate

---------

Co-authored-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>
2026-07-12 07:22:01 -07:00