driftproof run — skill "documentation-and-adrs" v0.0.0
  content_hash: b67a9f07ed105796…  suite: 1 cases (6cce54e7d9d0…)
  surface: claude-cli   models: claude-haiku-4-5-20251001   samples/case: 3   concurrency: 1
  registry: claude-haiku-4-5-20251001:registered   transcripts: retained-local
  judge: claude-opus-5   (--judge-model)
  isolation: same-user (--trusted-skill: self-authored skill)
  projected calls: 80/model × 1 model(s) = 80   per-model cap: 80
  projected cost: ~$0.81 (rough upper bound; budget $8.00, hard-stop $10.00)
  actual metered spend on claude-cli: $0.00 (subscription; the $ figure is the estimated-equivalent API cost, counted against the cap identically)

── model: claude-haiku-4-5-20251001 ──
  comment-intent-not-implementation / with_skill: pass (0.77 ± 0.06)
  comment-intent-not-implementation / baseline: fail (0.58 ± 0.05)
  → transcripts (retained-local): transcripts/3609ddc948814c96efd00ef98d89349bda96d29ca659265e784f82f478ccc44f/
  → ../runs/r014-20261007T203747Z/claude-haiku-4-5-20251001/documentation-and-adrs/receipts/documentation-and-adrs-claude-haiku-4-5-20251001-2026-10-07.json  (36 calls)
  → answered by: model claude-haiku-4-5-20251001 (attested by the surface; the same-user spawn (--trusted-skill)); verification_level TESTED
  → skill lift +0.193 ± n/a (1 case)  (with 0.772 ± n/a (1 case) vs base 0.579 ± n/a (1 case); band = sample stddev of the per-case means)

Done. 1 receipt(s) emitted to ../runs/r014-20261007T203747Z/claude-haiku-4-5-20251001/documentation-and-adrs/receipts/