mirror of
https://github.com/theclaymethod/unslop.git
synced 2026-09-19 01:45:09 +08:00
d81f519616
* Optimize and harden eval loop * Run behavioral acceptance on Luna * Define core product acceptance gate * Make behavioral acceptance honest and cheaper * Harden clean no-op and contextual scanner gating * Freeze blind Luna core benchmark v2 * Allow verified no-op benchmark results * Keep advisory benchmark signals non-blocking * Transform UNSLOP around the core rewrite contract * Fix release lint failures * Restore Python 3.8 compatibility --------- Co-authored-by: Clayton Kim <claytonkim@users.noreply.github.com>