Directional Anisotropy Gravity at Cosmic Dawn: Proto-Memory Halos and Falsifiable JWST/ALMA Signatures

Context. The James Webb Space Telescope has revealed luminous, compact, chemically evolved, and in some cases ultra-massive galaxies at very high redshift: MoM-z14 at z_spec = 14.44, JADES-GS-z14-0 at z = 14.1793, JADES-GS-z13-1 at z approximately 13.05, and the FRESCO red-monster population. These observations sharpen a specific problem: some baryonic systems appear dynamically and astrophysically efficient at cosmic dawn while the cosmic microwave background, the baryon acoustic scale, and large-scale linear structure remain close to standard LambdaCDM. Mechanism. Directional Anisotropy Gravity (DAG) interprets the apparent dark component around a galaxy not as a conserved particle halo but as an effective gravitational source generated by unrelaxed directional-anisotropy work stored in the substrate. This work is a Poisson-equivalent free energy of the anisotropic state, not a debit from baryonic rest mass. At cosmic dawn, compact nonlinear baryonic collapse writes this proto-memory rapidly; the forgetting rate is slow (Gamma_F approximately 0.013 Gyr^-1), so the local gravitational response can be large in rare compact systems without changing pre-recombination acoustic physics. Predictions. This paper presents a DAG-motivated predictive reduction for cosmic-dawn galaxies. The central prediction is a residual-correlation pattern: at fixed stellar mass and redshift, the inferred gravitational or efficiency excess should correlate with a dimensionless assembly-intensity index built from star-formation surface density, burst duration, compactness, burstiness, inflow, and merger activity, and should show hysteresis with respect to recent baryonic history. The SIS asymptote of the companion DAG paper provides a conservative benchmark normalization for the local response capacity. For MoM-z14 (R_e approximately 74 pc), this gives a conservative reference-floor range Xi_max^SIS in [0.07, 0.22] as M_b ranges over 10^9 to 10^8 solar masses. Larger values are possible if the burst-phase concentration factor g_burst > 1; its value must be calibrated observationally. Observational test. The prediction scheme is constructed with a collapse trigger that makes the cosmic-dawn response identically zero on linear pre-collapse modes, leaving CMB, BAO, and dark-acoustic-oscillation observables unchanged. The testable signatures are: (i) a positive Spearman correlation rho approximately 0.35 between high-z residuals and log assembly intensity, giving a first 2.5-3 sigma indication with N approximately 65 spectroscopically confirmed systems, or N approximately 80-100 for a robust controlled test; (ii) hysteresis in post-burst, post-merger, and recently quenched systems; and (iii) a slow secular decay of approximately 0.0055 dex/Gyr in fully quiescent systems. Absence of the predicted residual-correlation pattern in an adequately powered, spectroscopically confirmed, selection-controlled sample would rule out this cosmic-dawn prediction scheme as a useful organizing framework for the JWST anomalies. This paper is conditionally derived from "Directional Anisotropy Gravity: Theory, Verifications, and Predictions", DOI: 10.5281/zenodo.20071356.