Switch dashboard to Light Mode Premium Aesthetic

Premium Dark Mode UI Overhaul and Train.py fixes

Skip DB init in TESTING; add reset_db

feat: finalize CodeLens. rebranding and production environment polish

feat: production-grade baseline agent and batch evaluation tool

feat: add SQLite persistence for episodes, results, and leaderboard via SQLModel

fix: resolve critical startup bugs, port mismatch, and leaderboard crash

Add scripts/validate.py

Add scripts/baseline.py