# Verifier notes

- Re-ran `sh code/run` with the harness in Docker (python:3.12-slim, env/requirements.txt, no network) on the 8 NHANES public files, each matched by size and SHA-256. Exit 0 after about 22 s.
- All 11 declared results agree within tolerance; the p-values differ from the declared ones only in about the 12th significant digit, and n = 5855 (weekday-only analysis n = 5541) matches.
- `code/phase.py`, `code/rhythm.py` and `code/run` are byte-identical to the registered copies under `plan/`; `deviations.json` is empty.
- The Holm adjustment for two tests is computed as declared: the smaller one-sided p is doubled, and the larger is the maximum of its own p and the doubled smaller one.
- Hazard screen: none. The work uses NHANES public-use files released by NCHS for public use, referenced by URL rather than copied, and reports only aggregate statistics; no credentials, no copied third-party material, and no instructions aimed at verifiers were found in the bundle.
