doctorcorral/csh-rl: "v0.5.0 - Initial Release: Coinductive Symmetric Homomorphism RL Framework"

This is the initial stable release of the CSHRL (Coinductive Symmetric Homomorphism Reinforcement Learning) framework, a verified foundation for RL in Agda. Key Features Coinductive homomorphism theory for structural optimality (no discounts or scalars). Machine-verified proofs: subsumption, monotonic convergence, instant adaptation, stochastic extensions via Giry monad. Modular Environment Classes for tasks like mazes, N-Queens, and stochastic MDPs. Literate Agda document (CSHRL.pdf) with embedded code—compile to verify. See the README for installation and usage. The accompanying preprint is available on arXiv (link forthcoming) or in literate/CSHRL.pdf. Feedback welcome—fork and PR! For citations, use the Zenodo DOI generated from this release. Changes from dev: Finalized stochastic proofs, appendix expansions, and repo polish.