Deploy: Phase 1 surfaces + Phase 2 version stamps + filled legal/booking values

Deploy: dashboard naive-vs-rigorous reveal + sync serving to origin/main

calibrate(jury): 4A characterizes v1.1.1 residual as model-class-specific

calibrate(jury): v1.1+v1.1.1 — fix weighting bugs; recency-position paraphrase clause

rubric: clarify groundedness reference scope (snippets-only) for v1.1 gold (#20)

fix(calibration): per-corpus dispatch in generate-outputs (#19)

fix(judges,calibration,harness): three Codex adversarial-review findings

fix(judges,calibration): five review follow-ups (items 5, 6, 7, 9, 10)

feat(scripts): run_calibration.py orchestrator for Steps A/C/D

feat(calibration): 30-item stratified calibration_v1 sample

test(calibration): sklearn-parity fixtures + cross-check CI test

fix(ingest): exclude QUESTION_PLAN.md from corpus ingestion

feat: evaluate.py --corpus flag + CorpusConfig.golden_dataset

fix: deferred imports, match iteration budget, token cost tracking

feat: langchain evaluation CLI script and Makefile target

feat: add cross-encoder reranking with feature flag

feat: add grounded refusal gate based on retrieval score threshold

fix: retrieval metrics use ranked sources, LLM judge wired, report complete

feat: Day 7 — evaluation harness, metrics, report, expanded golden dataset

feat: add reproducible retrieval gate check with committed artifact

feat: Day 4 — corpus, ingest script, first 10 golden questions