Data and Code for "Boundary Amplification and Feedbac Criticality in a Nonreciprocal Langevin Chain"

This record contains the numerical data and software supporting the article “Boundary Amplification and Feedback Criticality in a Nonreciprocal Langevin Chain.” It provides the raw and processed data underlying Figs. 1–5, including seed-level open-chain covariance estimates, trajectory-level finite-time boundary-variance estimates, deterministic Lyapunov and entropy-production validation results, feedback-critical covariance and entropy-production calculations, complete simulation configurations, and scripts for regenerating the published figures. The stochastic campaign used master seed 20260714. The Fig. 1 campaign contains 32 independent seeds and 256 trajectories per seed for each simulated parameter point. The Fig. 2 campaign contains 32 seeds and 512 trajectories for each of 23 system sizes from N=8 through N=256, giving 376,832 trajectory estimates. The deterministic validation contains 10,000 independently sampled parameter cases. Human-readable CSV tables are organized by figure panel, while compressed NumPy archives preserve the lossless numerical arrays used by the plotting scripts. SHA-256 checksums are supplied for integrity verification. See README.md for the figure-to-data map, column descriptions, software requirements, and reproduction commands.