Phase VIII of the Relativistic Coupling Ansatz (RCA) separates the two couplings beta (expansion geometry) and delta (gravitational slip) into independent physical channels. Delta is now the EFT-of-dark-energy gravitational-slip amplitude mu(z) = 1 + delta × gate(z), the same "mu" constrained by Planck 2018 and DESI 2024 modified-gravity analyses. This release archives three falsifiable predictions not present in Phase VII: [1] Growth-without-lensing: mu(z) > 1 while Sigma(z) ~ 1 at z=[0.16,0.40]. Confirmed in CLASS Boltzmann code Stage B: 6.32% Weyl-power separation, 0.00% growth spread (commit 49c209b). Survey target: DESI Y3 + Rubin LSST. [2] fsigma8 deficit at z=[0.10,0.40] if delta > 0. Current data miss this window entirely (chi2 monotone, Δchi2=-0.016). DESI Year 3 (2027-2029) will fill the gap with 3 points at z=0.15, 0.25, 0.35. Detection threshold: delta ~ 0.30 at TPR=50%. [3] fsigma8 sign-flip at z ~ 0.58. The deviation changes sign due to beta-delta competition — unique to two-channel models. No single-channel model produces this pattern. Phase XVIII MCMC (96 walkers, 15,000 steps, tau_max=123, converged): corr(beta,delta) = -0.052 (degeneracy broken from -1.000 in Phase VII) Omega_m = 0.294 ± 0.006 H0 = 69.30 +0.37/-0.22 km/s/Mpc delta 95% upper bound: < 0.716 (prior-dominated; not a detection)