From af002ba235a2b1960696d7021c6a56e9442d3882 Mon Sep 17 00:00:00 2001 From: Bastien Chanot Date: Wed, 26 Aug 2026 11:15:03 +0200 Subject: [PATCH] =?UTF-8?q?chore(darwin):=20T3=20baseline=20=E2=80=94=2054?= =?UTF-8?q?=20rows,=20mean=2083.4,=2013=20candidates=20<80?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .claude/tasks/TODO.md | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/.claude/tasks/TODO.md b/.claude/tasks/TODO.md index bbe2bce..2eaa116 100644 --- a/.claude/tasks/TODO.md +++ b/.claude/tasks/TODO.md @@ -13,8 +13,11 @@ paired same-judge majority, absolute scores triage-only. EXCLUDED — machine-owned ctx7 (BDR-053, gitignored) → 31 systems. - [x] T2 Phase 0.5 gate PASSED: reuse prompts as-is; dim8 full_test on candidates only (baseline dry_run); Phase 2 set = ALL units <80. -- [ ] T3 Phase 1 baseline: batched blind judges, 9-dim, 55 rows in - results.tsv; scorecard. +- [x] T3 Phase 1 baseline DONE: 7 blind judges, 54 rows (31 skills + 23 + agents), mean 83.4, 13 units <80, ~25 verified findings (hotfix + destructive restore, onboard/onboarder contract, init-project + allowed-tools, skills-perso 8/32 detection...). + - [ ] T4 Phase 1 gate: scorecard checkpoint, user picks optimization set. - [ ] T5 Phase 2: per-unit loops (1 dim/round, weighted-gap diagnosis, paired 3-judge majority keep/revert, HL-4 stop), human checkpoint