This AET Lab Harness records raw local samples for the Quick acceptance budgets. It adds no Quick command and makes no model-service or cross-repository claim.
python eval/quick-performance/runner.py \
--root . \
--base v1.12.0 \
--intent aet.intent.json \
--repetitions 30 \
--output eval/quick-performance/results/v1.13.0.jsonThe report retains every time.perf_counter sample, uses nearest-rank P95, and
records Python, platform, architecture, tracked-file count, changed-file count,
and the exact base. Check and Scope use this repository; Fresh uses a minimal
temporary Git fixture with an exact proof.