Skip to content

v3.2.2

Choose a tag to compare

@masterleopold masterleopold released this 02 Oct 04:48
· 3 commits to main since this release
71ed942
  • Stage 1 computes the assumption volume with skills/brd/scripts/assumption-volume.sh instead of counting by hand. In the 2026-10-02 trials the lead reported 60.1% for a BRD that measures 53.1%, and the Stage 1 gate turns on this number. The same plan still gave 53.1% under one stack and 29.4% under another, because the denominator depends on how finely requirements are split; the script makes the number reproducible for a given BRD, not independent of granularity