Source verification of earlier GGPT predictions in Structural Cosmology, and a reproducibility check of the public v2.2 demonstration code

Earlier Structural Cosmology publications state deviations in the growth observable fσ₈ of approximately −5 to −7 % (v1.3 and v2.0) or O(5 %) (v2.2). This report examines the sources and stated status of these numerical claims, traces the calculation implemented in the public v2.2 demonstration code, and tests whether the reported suppression can be attributed to the structural-field contribution. PDFs, README files, and code from five public Zenodo records were retrieved and checked against their listed checksums. The v2.2 demonstration code was evaluated with its default parameters and compared with a zero-effect control in which mu_eff = 1. Separate diagnostic comparisons used the same numerical implementation, background, initial conditions, and integration settings, with either a common present-day amplitude (scheme A) or a common initial amplitude (scheme B). All 82 pages of the three principal PDFs were processed by OCR; visual inspection of figures covered selected pages, as documented in the report. For z ∈ [1, 1.5] and k ∈ [0.03, 1], the public code gives a negative ggpt_delta of approximately −4.6 % to −2.0 %. A similar deviation, approximately −4.6 % to −2.5 %, occurs in the zero-effect control. The growth rate f computed by the code is N times the logarithmic growth rate d ln D/d ln a, where N = D_raw(1) ≈ 0.779. However, N cancels in the product fσ₈, including when f and D are interpolated separately before multiplication. The product of the interpolated quantities is not generally identical to the interpolated derivative; that interpolation difference was not quantified. On the aligned diagnostic grid of 13 k values from 10⁻³ to 1 and 51 redshift values from 1 to 1.5, the relative difference in fσ₈ between the structural-field calculation and its zero-effect control ranges from −0.061 % to +0.098 % under scheme A and from approximately +1.75 × 10⁻⁶ % to +0.662 % under scheme B. Thus, a suppression of several percent attributable to the structural field is not demonstrated by the code, settings, and grid examined. The source documents describe the numerical claims using terms including “illustrative,” “phenomenological,” and “working hypothesis.” Within the scope examined, the calculation connecting the stated predictions to their numerical amplitudes could not be established. The baseline model and amplitude conditions underlying the prediction text could not be identified, and the original prediction conditions have not been reconstructed. O(5 %) is treated as an order-of-magnitude statement rather than a defined acceptance interval. This report is an investigation and recalculation by the author of the source publications, not an independent verification. Its findings do not establish or refute the theory as a whole or the earlier predictions. The original coefficient-selection process and identity with the files at their initial publication dates remain unconfirmed. The accompanying files comprise the report in PDF and Markdown, an evidence ZIP containing checking and diagnostic scripts, numerical outputs, reproduction instructions, source and license information, and checksum manifests. The original v2.2 demonstration code is not included; retrieval instructions and identifying hashes are supplied. The report, documentation, and numerical outputs are licensed under CC BY 4.0. The diagnostic code, including the adapted portions, is licensed under the MIT License.