Worked on the Card-Forge/forge repository to enhance the reliability and determinism of the game simulation copier, focusing on ensuring that copied game states remain identical to their originals. Addressed three critical state-fidelity issues by refining Java-based simulation logic to prevent hidden information leakage, maintain correct player effect card associations, and preserve original card IDs for deterministic behavior. Introduced targeted regression tests using unit testing frameworks, validating the fixes through an end-to-end test suite. These improvements strengthened the quality assurance process for copy-based workflows, supporting safer AI training and reproducible replays, and demonstrated a methodical approach to simulation reliability and correctness.