Homotopy Type Theory Engine for Reinforcement Learning R package PsiUEngineRL version 0.1.3 Core architecture for interpreting continuous data streams as homotopy types. It evaluates identity paths against the Gnomonic Ratio ('Lombardi', 2026) <doi:10.5281/zenodo.20385840> and processes them via a dynamic 'Tableau Refutation Tree'. The engine categorizes data into necessity (BOX), possibility (DIAMOND), or noise based on deviation thresholds from the invariant value. Includes adaptive auto-tuning and native high-contrast Cartesian graphics for structural entropy isolation. Software Depends: R (>= 3.5) Roberto Lombardi <lombardisedr@gmail.com> Comprehensive R Archive Network (CRAN) Roberto Lombardi [aut, cre] MIT + file LICENSE (https://CRAN.R-project.org/package=PsiUEngineRL/LICENSE) 2026-06-07 application/tgz https://CRAN.R-project.org/package=PsiUEngineRL doi:10.32614/CRAN.package.PsiUEngineRL