Independent reproduction of Table III / §II F (arXiv:2604.11963)

Repository: https://github.com/VolMax-Studio/rotation-gap-fano-s1
Source work: arXiv:2604.11963

Following correspondence with Daniel Nüst, I would like to propose a community CODECHECK of an independent third-party reproduction.

I am not an author of the source work. The requested scope is deliberately narrow: independently execute the deposited workflow and determine whether the four preregistered subclaims W1–W4, corresponding to Table III / Section II F, can be reproduced from the public Google Willow detector data using the documented artefacts.

The workflow has one entry point, `reproduce.py`. It verifies the frozen preregistration and source-data manifest before computation and produces a deterministic machine-readable result.

**Expected canonical result artefact:**
`results/results.json`
SHA-256: `38fff894960d4c191b3c017c6d8d24ed35430a8eb19b610550a85f294a8281b8`

The source-data archive is approximately 5.72 GB. Compute after download is approximately 70 seconds with approximately 800 MB peak RSS.

For provenance only, a public GitHub Actions CI-as-Witness Reproduction execution, run 33951221253 on 5 September 2026, completed successfully and included execution of `reproduce.py`, assertion of the exact byte-level result hash above, and build-provenance attestation:
https://github.com/VolMax-Studio/rotation-gap-fano-s1/actions/runs/33951221253

This prior execution is disclosed only as existing provenance and is not offered as a substitute for an independent CODECHECK.

Please report any dependency issue, ambiguity, provenance mismatch, execution failure or numerical discrepancy rather than attempting to reconcile it with the expected result.

The source author was notified publicly on 5 September 2026 through:
https://github.com/SelinaAliens/The_Rotation_Gap_Is_Not_An_Error/issues/1

I have not identified a public email address for the source author in the reviewed work, so the notification was made through the source repository. Posting of that notification is confirmed; no claim is made that it was read.

I am happy for CODECHECK to involve the source author further in whatever way is appropriate.

No CODECHECK certification is claimed unless and until a certificate is formally issued.

Start with reproduce.py in the linked repository and verify its frozen preregistration and source-data manifest checks before running the workflow. Download the approximately 5.72 GB archive, execute the entry point, and compare results/results.json with the stated SHA-256 hash. Done means reporting dependency, provenance, execution, ambiguity, or numerical discrepancies and whether W1–W4 reproduce.

Written by the indexing model from the issue text.