feat(tic-tac-toe): add complete game with optimal Minimax-based AI

[axelfrache]

Overview

This pull request introduces a complete Tic-Tac-Toe game featuring an unbeatable AI opponent powered by the Minimax algorithm.

The game runs entirely in the terminal with a clean text-based interface.
The AI evaluates every possible outcome to guarantee optimal play - it will never lose; the best a human can achieve is a draw.

---

Implementation details

The project is organized into modular components:

• board/ - manages the 3×3 game state stored as a 1D array (simplified indexing for positions 1–9)
• game.rs - implements the rules, turn logic, and win/draw detection
• ai/ - contains the Minimax algorithm with depth-based evaluation and scoring
• render/ - handles terminal rendering and user input

Each module is independent and testable in isolation.
The code emphasizes readability, separation of concerns, and idiomatic Rust patterns.

---

Documentation

Detailed architecture and rationale are provided in docs/architecture.md.
The document focuses on why each design decision was made - module boundaries, AI scoring strategy, and trade-offs for simplicity and clarity.

---

Testing

A comprehensive test suite (32 tests) covers:

• Board operations (bounds, invalid moves)
• All 8 winning configurations (rows, columns, diagonals)
• AI strategy validation (blocking, winning, optimal openings)

All tests pass (cargo test), no lints from clippy and the code is fully formatted with rustfmt.

---

Usage

cd topics/tic-tac-toe
cargo test   # Run all tests
cargo run    # Play against the AI