🎮 New Tic-Tac-Toe Game Features:
- Compact 3x3 board display using ASCII art
- Smart AI opponent with win/block/random strategy
- All messages under 200-character meshtastic limit (tested: 10-50 chars)
- Player vs Bot gameplay with X (player) vs O (bot)
- Win detection for rows, columns, and diagonals
- Tie game detection when board is full
- Game statistics tracking (games played, won)
🔧 Integration Features:
- Follows established game patterns from hangman/hamtest
- Added to restrictedCommands (DM-only like other games)
- Integrated with game tracker system for memory cleanup
- Added configuration option in config.template
- Automatic cleanup of stale game sessions
🎯 Game Mechanics:
- Players pick positions 1-9 corresponding to board layout
- Simple input parsing (extracts first digit from message)
- Graceful error handling for invalid moves
- 'end' command to quit game
- Automatic game cleanup on completion
📊 Message Examples:
- New game: 39 chars
- Game moves: 50 chars
- Win/lose: 40 chars
- Invalid move: 23 chars
- All well under 200-char limit
Tested: Complete game scenarios, AI behavior, message lengths
Follows: Existing game implementation patterns and memory management
this is a change as its looking to remove google lookups with the python module. if this changes is impacting please let me know in [general] add `rawLLMQuery = False` to reverse
LLM rewrite, the removal of RAG to keep things clean.
This default changes as well as puts input direct to the LLM further testing is needed, new LLM prompting is different.
The current behavior is that whenever ollama is enabled the LLM replies
to all non-command message. This setting allows limiting it the LLM to
run only in response to the "ask:" and "askai" commands.
Default is True to keep consistent with previous behavior.
